key: cord-130778-d6jtz3pm authors: hardy, peter; marcolino, leandro soriano; fontanari, jos'e f. title: the paradox of productivity during quarantine: an agent-based simulation date: 2020-08-21 journal: nan doi: nan sha: doc_id: 130778 cord_uid: d6jtz3pm economies across the globe were brought to their knees due to lockdowns and social restriction measures to contain the spread of the sars-cov-2, despite the quick switch to remote working. this downfall may be partially explained by the"water cooler effect", which holds that higher levels of social interaction lead to higher productivity due to a boost in people's mood. somewhat paradoxically, however, there are reports of increased productivity in the remote working scenario. here we address quantitatively this issue using an agent-based model to simulate a workplace with extrovert and introvert agent stereotypes that differ solely on their propensities to initiate a social interaction. we find that the effects of curtailing social interactions depend on the proportion of the stereotypes in the working group: while the social restriction measures always have a negative impact on the productivity of groups composed predominantly of introverts, they may actually improve the productivity of groups composed predominantly of extroverts, which offers then an explanation for the paradox of productivity during quarantine. introduction. -the interest on the trade-off between work productivity and social interaction is not a present-day fad, as attested by this verse of the 19th century writer maria edgeworth: "all work and no play makes jack a dull boy, all play and no work makes jack a mere toy" [1] . in fact, this is a staple subject of social psychology that addresses the influence of loneliness or, more generally, of mood on human cognitive function, often with discrepant findings (see, e.g., [2, 3] ). recent lockdowns and social restriction measures that isolated workers from their peers have placed these issues at the forefront of public attention, with news outlets reminding their customers that they do not need to be productive, and to expect a reduction in overall productivity during this time. but why is it that productivity would decrease during a period that one would think people have more time on their hands than ever before? typical answers to this question are in line with the so-called "water cooler effect" that has been highlighted (and satirized) in how effective placing an inanimate object for people to congregate around can stir up casual conversations, with many psychologists believing that this can also increase company productivity [4] . however, this is not an obvious conclusion since for a singular worker the extra down time communicating may actually detract from its productivity as a whole. here we address quantitatively the productivity and social interaction issue using an agent-based model to simulate a workplace scenario where the agents exhibit two social stereotypes, viz., extroverts and introverts, that differ solely on their propensities to initiate a conversation. social distancing is modeled by controlling the number of attempts an agent makes to find a conversation partner and the motivation to work (mood) is assumed to increase with the time spent talking and decrease with the time spent alone. moreover, the instantaneous productivity of a lone agent increases linearly with its motivation to work. we find that the effect of curtailing social interactions depends on the proportion of the different stereotypes in the group. for instance, while social restriction measures always have a negative impact on the productivity of groups composed predominantly of introverts, they may actually improve the productivity of groups composed predominantly of extroverts. in addition, within a same working group the productivity of the two stereotypes is affected differently by those measures. these results may explain the paradoxical findings that productivity is increased in some remote work experiments [5] . model. -we model the dynamics of a group of n agents that interact socially and work during a 480minute (eight-hour) workday. we first make a strong assumption that for all agents they are more productive when they are motivated, and their motivations increase when they participated in some form of social interaction [6] [7] [8] . in this framework, motivation has both an emotional dimension, in the sense that talking with peers improves the mood of the agent and helps it to focus on the task, and an informational dimension, in that talking with peers may result in the acquisition of valuable information that can help the agent to consummate its task. we assume that the motivation of agent k = 1, . . . , n is determined by the integer parameter l k = 1, 2, . . . , l max and that at the beginning of the day, the agents' motivations are the lowest, i.e., l k = 1. in addition, we assume that the instantaneous productivity of an agent at time t = 0, 1, . . . , t is p k (t) = l k − 1 if agent k is not talking and p k (t) = 0 if agent k is talking. since we measure time in minutes, we set t = 480. hence, at time t = 0 the instantaneous productivity is zero for all agents. the only way an agent can increase its motivation and, consequently, its productivity, is by engaging in social interaction. we introduce two different social stereotypes our agents can be, viz., extroverts and introverts, which determine their propensities to seek and engage in social interaction. in particular, if agent k has stereotype l, arxiv:2008.09461v1 [cs.ma] 21 aug 2020 where l = e for extroverts and l = i for introverts, the probability that it instigates a conversation is for l k ≤ t /τ l and p l k = 0, otherwise. here τ e and τ i are parameters measured in minutes that are necessary to make eq. (1) dimensionally correct. throughout this paper we set τ e = 1 without loss of generality, but set τ i > 1 in order to guarantee that p e k ≥ p i k , i.e., that the extroverts are more likely to engage in social interaction than the introverts. in addition, p e k = p i k = 1 for l k = 1 so that when the motivation of agent k is at the bottom line, it will try to engage in social interaction with certainty. the propensity to instigate a conversation decreases linearly with the motivation to work until it vanishes altogether when the motivation parameter reaches a threshold value that differs for the two social stereotypes. regardless of its stereotype, once a target agent decides to instigate a conversation, it selects a number m of contact attempts, where m = 0, 1, . . . is a random variable drawn from a poisson distribution of parameter q. in each contact attempt, a peer is selected at random among the n − 1 agents in the group and, in case the selected agent is not in a conversation at that time, a conversation is initiated and the target agent halts its search for a partner. the duration of the conversation d is given by a random integer selected uniformly in {1, ..., d}. we note that a conversation involves two agents only and the agent that is approached by the target agent is obliged to accept the interaction, regardless of its motivation and stereotype. to complete the model we need to give a prescription for changing the motivation parameter l k with k = 1, . . . , n , which is the central factor in the determination of the social behavior of the agents. we assume simply that the motivation of agent k increases by one unit for each minute it spends in a conversation and decreases by one unit for each minute it spends working alone. moreover, we set l k = 1 as the lower bound of the motivation parameter so that a lone agent k at the motivation bottom line l k = 1 cannot have its motivation reduced any further. we note that our prescription to change the motivation of the agents implies that the motivation parameter is bounded from above by l max = 480, so that the probability that an extroverted agent instigates a conversation, eq. (1), is never zero during the eight-hour workday. this produces a qualitative distinction between the two stereotypes since the extroverts are, in principle, always willing to initiate a social interaction (i.e., p e k > 0), whereas the introverts will not instigate a conversation if their motivation to work is above the threshold t /τ i . throughout this paper we set τ i = 5 but note that this choice is immaterial, provided that t /τ i is sufficiently distinct from the upper bound t /τ e to justify the existence of two stereotypes. in summary, we implement the synchronous or parallel update of the n agents as follows. at t = 0 we set l k = 1 and p k = 0 for all agents k = 1, . . . , n . the update procedure at time t begins with the selection of a random order of update of the n agents, so that at the end of the procedure all agents are updated and we can increment the time from t to t + 1. then we check the status of the agent to be updated -the target agent -to determine if it is participating in a conversation or not. in case it is, we increment its motivation parameter by one unit. in case it is not, we decrement its motivation parameter by one unit and test whether it initiates a conversation or not using the probabilities given in eq. (1) in accord with the stereotype of the target agent. the simulation ends at t = t . results. -we consider working groups composed of n e extroverts and n i = n − n e introverts and focus mainly on the mean cumulative productivity of the stereotypes at time t = 1, . . . , t , where and the superscript l = e, i specifies the social stereotypes of the agents, as before. here the sum over k is restricted to the subset s l of agents with stereotype l, whose cardinality is n l . we recall that p k (0) = 0 for all k. the mean cumulative productivity of the whole group is simply π w = n e π e + n i π i /n , where we have omitted the dependence on t for simplicity. it is also of interest to know the mean motivation of the stereotype subgroups, with l = e, i as before. then the mean motivation of the whole group is λ w = n e λ e + n i λ i /n . our main interest here is π l (t) and λ l (t) evaluated at the end of the day, i.e, at t = t . these quantities are averaged over 10 3 independent runs for each setting of the model parameters. in particular, in the following we vary the fraction of extroverts η = n e /n and the mean number of contact attempts q, but fix the group size to n = 100 and the maximum duration of a conversation to d = 20. as already pointed out, we set τ e = 1 and τ i = 5. to better appreciate the assumptions of the model, fig. 1 shows typical time dependences of the motivation and mean cumulative productivity of two agents with distinct stereotypes for a single run. the time intervals where the mean cumulative productivity decreases correspond to the periods when the agent is participating in a social interaction and are associated with the increase of its motivation. we note that, whereas the cumulative productivity is a non-decreasing function of t, the mean cumulative productivity, eq. (3), decreases with increasing t in the time intervals where the instantaneous productivity of the agent is zero. in addition, the motivation decreases in the periods where the productivity increases. most of the time in this run, the introvert agent exhibits a higher mean productivity than the extrovert agent, despite its lower motivation. in fact, the motivation parameter of the introvert stabilizes and fluctuates around the value l = t /τ i = 96, whereas the motivation parameter of the extrovert shows a tendency to increase linearly with t on average. it is easy to understand the behavior pattern of the introvert's motivation if we note that an introvert agent with l ≥ t /τ i will not attempt to instigate any conversation and so the tendency of its motivation is to decrease only, except for the events when another agent engages it in a conversation. figure 2 shows the influence of the fraction of extroverts η on the motivation and mean cumulative productivity of the agents at the end of the day. this figure reveals some interesting features. first and foremost, the insertion of extroverts in groups composed predominantly of introverts results in a significant increase of the mean productivity of the introverts, whereas the mean productivity of the extroverts is barely changed. too many extroverts, however, cause a decrease on the productivity of both stereotypes, although they considerably boost the motivation of the introverts. in addition, there is an optimum group composition (η ≈ 0.5 for the parameters used in fig. 2 ) that maximizes the total productivity of the group. in fig. 3 , we examine the influence of the mean number of contact attempts q on the optimum composition of the group. for instance, for small q, say q = 0.5, it is difficult to find conversation partners to increase the agents' motivations and so the highest total productivity is achieved by all-extroverts groups, since this stereotype seeks for partners more frequently than the introverts. as q increases, the odds of finding a conversation partner increases as well, and so it becomes advantageous to insert a few introverts in the group. for very large q, so that the target agent attempts to contact all the other agents in the group, the maximum total productivity occurs for η ≈ 0.47. for groups composed predominantly of introverts (η < 0.5), increase of q and η lead always to higher total productivity. however, the scenario is more complicated for groups composed predominantly of extroverts. in fact, for η > 0.5 our results indicate that there is an optimum value of the mean number of contact attempts that maximizes the total productivity. for instance, this optimum value is located between q = 1 and q = 2 for η ≈ 1. it is worthwhile to examine the effects of the parameter q in more detail. in the context of the 2020 coronavirus pandemic, this is the leading parameter of the model since, at least for those adhering to social distancing norms, the mean number of attempts to interact socially was considerably curtailed during the pandemic. accordingly, fig. 4 shows that a moderate decrease of q actually increases the productivity of the group for η > 0.5. for instance, for η > 0.8 the decrease from q = 6 to q = 1 results in a productivity increase of about 10%, which is in agreement with the experiments of ref. [5] . we note that in both our simulations and those experiments the boost in productivity is modest, though altogether unexpected. it is interesting to look at how the productivities of the two stereotypes within a same working group are affected by the social restrictions. figure 5 shows that, for η = 0.5, a moderate decrease of q reduces the motivation of both stereotypes as well as the mean productivity of the introverts, but increases the mean productivity of the extroverts. this finding may explain sporadic personal reports of heightening productivity in the quarantine. further decrease of q leading to a scenario where social interactions happen very rarely results in a sharp drop of the mean productivity of both stereotypes. a word is in order about the effects of the parameters n and d that determine the group size and the maxi-mum duration of a conversation, respectively. reducing the group size has no effect on our results and our choice of fairly large groups (n = 100) is so as to smooth out the variation of the fraction of extroverts η. however, increase of the parameter d has a strong effect on the productivity of the extroverts. for instance, for long conversations, say d = 60, all-introverts groups yield the highest total productivity regardless of the number of contact attempts. our choice d = 20, which corresponds to conversations of average durationd = 10, are meant to model a "water cooler talk" scenario. conclusion. -in the spirit of sociophysics [9, 10] and computational social science [11] , our agent-based model assumes only linear relationships between the quantities relevant to study the influence of social restriction measures on the productivity of working groups. in that sense, it offers a proof of concept that, somewhat paradoxically, quarantining and social distancing may boost the productivity of extroverted people. * * * perceived social isolation and cognition feeling low, thinking slow? associations between situational cues, mood and cognitive function how social networks network best does working from home work? evidence from a chinese experiment positive affect facilitates creative problem solving resource allocation model of the effects of depressed mood states on memory putting mood in context: using smartphones to examine how people feel in different locations statistical physics of human cooperation the social physics collective key: cord-008495-gjn8kh2t authors: nan title: cumulative chapter titles keyword index, vol. 1–41 date: 2007-01-26 journal: annu rep med chem doi: 10.1016/s0065-7743(06)41034-4 sha: doc_id: 8495 cord_uid: gjn8kh2t nan antihyperlipidemics, 15, 162; 18, 161; 24, 147 antihypertensives, 1, 59; 2, 48; 3, 53; 4, 47; 5, 49; 6, 52; 7, 59; 8, 52; 9, 57; 11, 61; 12, 60; 13, 71; 14, 61; 15, 79; 16, 73; 17, 61; 18, 69; 19, 61; 21, 63; 22, 63; 23, 59; 24, 51; 25, 51 antiinfective agents, 28, 119 antiinflammatory agents, 28, 109; 29, 103 anti-inflammatories, 37, 217 anti-inflammatories, non-steroidal, 1, 224; 2, 217; 3, 215; 4, 207; 5, 225; 6, 182; 7, 208; 8, 214; 9, 193; 10, 172; 13, 167; 16, 189; 23, 181 anti-ischemic agents, 17, 71 antimalarial inhibitors, 34, 159 antimetabolite cancer chemotherapies, 39, 125 antimetabolite concept, drug design, 11, 223 antimicrobial drugs-clinical problems and opportunities, 21, 119 antimicrobial potentiation, 33, 121 antimicrobial peptides, 27, 159 antimitotic agents, 34, 139 antimycobacterial agents, 31, 161 antineoplastics, 2, 166; 3, 150; 4, 154; 5, 144; 7, 129; 8, 128; 9, 139; 10, 131; 11, 110; 12, 120; 13, 120; 14, 132; 15, 130; 16, 137; 17, 163; 18, 129; 19, 137; 20, 163; 22, 137; 24, 121; 28, 167 anti-obesity agents, centrally acting, 41, 77 antiparasitics, 1, 136, 150; 2, 131, 147; 3, 126, 140; 4, 126; 5, 116; 7, 145; 8, 141; 9, 115; 10, 154; 11, 121; 12, 140; 13, 130; 14, 122; 15, 120; 16, 125; 17, 129; 19, 147; 26, 161 antiparkinsonism drugs, 6, 42; 9, 19 antiplatelet therapies, 35, 103 antipsychotics, 1, 1; 2, 1; 3, 1; 4, 1; 5, 1; 6, 1; 7, 6; 8, 1; 9, 1; 10, 2; 11, 3; 12, 1; 13, 11; 14, 12; 15, 12; 16, 11; 18, 21; 19, 21; 21, 1; 22, 1; 23, 1; 24, 1; 25, 1; 26, 53; 27, 49; 28, 39; 33, 1 antiradiation agents, 1, 324; 2, 330; 3, 327; 5, 346 anti-resorptive and anabolic bone agents, 39, 53 anti-retroviral chemotherapy, 25, 149 antiretroviral drug therapy, 32, 131 antiretroviral therapies, 35, 177; 36, 129 antirheumatic drugs, 18, 41, 183 antisense oligonucleotides, 23, 295; 33, 313 antisense technology, 29, 297 antithrombotics, 7, 78; 8, 73; 9, 75; 10, 99; 12, 80; 14, 71; 17, 79; 27, 99; 32, 71 antithrombotic agents, 29, 103 antitumor agents, 24, 121 antitussive therapy, 36, 31 antiviral agents, 1, 129; 2, 122; 3, 116; 4, 117; 5, 101; 6, 118; 7, 119; 8, 150; 9, 128; 10, 161; 11, 128; 13, 139; 15, 149; 16, 149; 18, 139; 19, 117; 22, 147; 23, 161; 24, 129; 26, 133; 28, 131; 29, 145; 30, 139; 32, 141; 33, 163; 37, 133; 39, 241 antitussive therapy, 35, 53 anxiolytics, 26, 1 apoptosis, 31, 249 aporphine chemistry, 4, 331 arachidonate lipoxygenase, 16, 213 arachidonic acid cascade, 12, 182; 14, 178 arachidonic acid metabolites, 17, 203; 23, 181; 24, 71 arthritis, 13, 167; 16, 189; 17, 175; 18, 171; 21, 201; 23, 171, 181; 33, 203 arthritis, immunotherapy, 23, 171 aspartyl proteases, 36, 247 asthma, 29, 73; 32, 91 asymmetric synthesis, 13, 282 atherosclerosis, 1, 178; 2, 187; 3, 172; 4, 178; 5, 180; 6, 150; 7, 169; 8, 183; 15, 162; 18, 161; 21, 189; 24, 147; 25, 169; 28, 217; 32, 101; 34, 101; 36, 57; 40, 71 atherosclerosis hdl raising therapies, 40, 71 atherothrombogenesis, 31, 101 atrial natriuretic factor, 21, 273; 23, 101 attention deficit hyperactivity disorder, 37, 11; 39, 1 autoimmune diseases, 34, 257; 37, 217 autoreceptors, 19, 51 bace inhibitors, 40, 35 bacterial adhesins, 26, 239 bacterial genomics, 32, 121 bacterial resistance, 13, 239; 17, 119; 32, 111 bacterial toxins, 12, 211 bacterial virulence, 30, 111 basophil degranulation, biochemistry, 18, 247 bcl2 family, 31, 249; 33, 253 behavior, serotonin, 7, 47 benzodiazepine receptors, 16, 39, 155 bioinformatics, 36, 201 bioisosteric groups, 38, 333 bioisosterism, 21, 283 biological factors, 10, 39; 11, 42 biological membranes, 11, 222 biological systems, 37, 279 biopharmaceutics, 1, 331; 2, 340; 3, 337; 4, 302; 5, 313; 6, 264; 7, 259; 8, 332 biosensor, 30, 275 biosimulation, 37, 279 biosynthesis, antibotics, 12, 130 biotechnology, drug discovery, 25, 289 biowarfare pathegens, 39, 20, 305; 40, 403 blood enzymes, 1, 233 bone, metabolic disease, 12, 223; 15, 228; 17, 261; 22, 169 bone metabolism, 26, 38, 111 bradykinin b2 antagonists, 39, 89 brain, decade of, 27, 1 c5a antagonists, 39, 109 calcium antagonists/modulators, 16, 257; 17, 71; 18, 79; 21, 85 calcium channels, 30, 51 calmodulin antagonists, sar, 18, 203 cancer, 27, 169; 31, 241; 34, 121; 35, 123; 35, 167 cancer chemosensitization, 37, 115 cancer chemotherapy, 29, 165; 37, 125 cancer cytotoxics, 33, 151 cancer, drug resistance, 23, 265 cancer therapy, 2, 166; 3, 150; 4, 154; 5, 144; 7, 129; 8, 128; 9, 139, 151; 10, 131; 11, 110; 12, 120; 13, 120; 14, 132; 15, 130; 16, 137; 17, 163; 18, 129; 21, 257; 23, 151; 37, 225; 39, 125 cannabinoid receptors, 9, 253; 34, 199 cannabinoid, receptors, cb1, 40, 103 carbohydrates, 27, 301 carboxylic acid, metalated, 12, 278 carcinogenicity, chemicals, 12, 234 cardiotonic agents, 13, 92; 16, 93; 19, 71 cardiovascular, 10, 61 caspases, 33, 273 catalysis, intramolecular, 7, 279 catalytic antibodies, 25, 299; 30, 255 cathepsin k, 39, 63 ccr1 antagonists, 39, 117 ccr3 antagonists, 38 1, 99; 2, 91; 4, 56; 6, 68; 8, 93; 10, 90; 12, 91; 16, 83; 17, 89; 18, 89; 20, 117; 23, 201, 38 non-steroidal, 1, 191; 3, 184 hormones, peptide, 5, 210; 7, 194; 8, 204; 10, 202; 11, 158; 16, 199 hormones, steroid, 1, 213; 2, 208; 3, 207; 4, 199 host modulation, infection, 8, 160; 14, 146; 18, 149 hsp90 inhibitors, 40, 263 5-ht2c receptor modulator, 37, 21 human gene therapy, 26, 315; 28, 267 human retrovirus regulatory proteins, 26, 171 11 b-hydroxysteroid dehydrogenase type 1 inhibitors, 41, 2, 273; 7, 47; 21, 41 hypercholesterolemia, 24, 147 hypersensitivity, delayed, 8, 284 hypersensitivity, immediate, 7, 238; 8, 273 hypertension, 28, 69 hypertension, etiology, 9, 50 hypnotics, 1, 30; 2, 24; 3, 28; 4, 28; 7, 39; 8, 29; 10, 30; 11, 13; 12, 10; 13, 21; 14, 22; 15, 22, 16; 31; 17, 11; 18, 11; 19, 11; 22, 11 ice gene family, 31, 249 ige, 18, 247 immune cell signaling, 38, 275 immune mediated idiosyncratic drug hypersensitivity, 26, 181 immune system, 35, 281 immunity, cellular mediated, 17, 191; 18, 265 immunoassay, enzyme, 18, 285 immunomodulatory proteins, 35, 281 immunophilins, 28, 207 immunostimulants, arthritis, 11, 138; 14, 146 immunosuppressants, 26, 211; 29, 175 immunosuppressive drug action, 28, 207 immunosuppressives, arthritis, 11, 138 immunotherapy, cancer, 9, 151 17, 291; 19, 241; 24, 71 lhrh, 20, 203; 23, 211 lipid metabolism, 9, 172; 10, 182; 11, 180; 12, 191; 13, 184; 14, 198; 15, 162 , 19, 313; 20, 315; 21, 323; 22, 315; 23, 325; 24, 295; 25, 309; 26, 297; 27, 321; 28, 325; 29, 331; 30, 295; 31, 337; 32, 305; 33, 327 mass spectrometry, 31, 319; 34, 307 mass spectrometry, of peptides, 24, 253 mass spectrometry, tandem, 21, 213; 21, 313 mast cell degranulation, biochemistry, 18, 247 matrix metalloproteinase, 37, 209 matrix metalloproteinase inhibitors, 35, 167 mechanism based, anticancer agents, 25, 129 mechanism, drug allergy, 3, 240 mechanisms of antibiotic resistance, 7, 217; 13, 239; 17, 119 medicinal chemistry, 28, 343; 30, 329; 33, 385; 34, 40, 38, 31 melatonin, 32, 31 melatonin agonists, 39, 21 membrane function, 10, 317 membrane regulators, 11, 210 membranes, active transport, 11, 222 memory, 3, 279; 12, 30; 16, 51 metabolism, cell, 1, 267; 2, 286 metabolism, drug, 3, 227; 4, 259; 5, 246; 6, 205; 8, 234; 9, 290; 11, 190; 12, 201; 13, 196, 304; 14, 188; 23, 265, 315 metabolism, lipid, 9, 172; 10, 182; 11, 180; 12, 191; 14, 198 metabolism, mineral, 12, 223 metabonomics, 40, 387 metabotropic glutamate receptor, 35, 1, 38, 21 metal carbonyls, 8, 322 metalloproteinases, 31, 231; 33, 131 metals, disease, 14, 321 metastasis, 28, 151 microbial genomics, 37, 95 microbial products screening, 21, 149 microtubule stabilizing agents, 37, 37, 247 migraine, 22, 41; 32, 1 mitogenic factors, 21, 237 mitotic kinesin inhibitors, 39, 135 modified serum lipoproteins, 25, 169 molecular diversity, 26, 259, 271; 28, 315; 34, 287 molecular modeling, 22, 269; 23, 285 monoclonal antibodies, 16, 243; 27, 179; 29, 317 monoclonal antibody cancer therapies, 28, 237 monoxygenases, 9, basis of, 41, 337 multivalent ligand design, 35, 321 muscarinic agonists/antagonists, 23, 81; 24, 31; 29, 23 muscle relaxants, 1, 30; 2, 24; 3, 28; 4, 28; 8, 37 muscular disorders, 12, 260 mutagenicity, mutagens, 12, 234 mutagenesis, sar of proteins, 18, 237 myocardial ischemia, acute, 25, 71 narcotic antagonists, 7, 31; 8, 20; 9, 11; 10, 12; 11, 23; 13, 41 natriuretic agents, 19, 253 natural products, 6, 274; 15, 255; 17, 301; 26, 259; 32, 285 natural killer cells, 18, 265 neoplasia, 8, 160; 10, 142 neurodegeneration, 30, 31 neurodegenerative disease, 28, 11 neurokinin antagonists, 26, 43; 31, 111; 32, 51; 33, 71; 34, 51 neurological disorders, 31, 11 neuronal calcium channels, 26, 33 neuronal cell death, 29, 13 neuropathic pain, 38, 1 neuropeptides, 21, 51; 22, 51 neuropeptide y, 31, 1; 32, 21; 34, 31 neuropeptide y receptor modulators, 38, 61 neuropeptide receptor antagonists, 38, 11 neuroprotection, 29, 13 neuroprotective agents, 41, 39 neurotensin, 17, 31 neurotransmitters, 3, 264; 4, 270; 12, 249; 14, 42; 19, 303 neutrophic factors, 25, 245; 28, 11 neutrophil chemotaxis, 24, 233 nicotinic acetylcholine receptor, 22, 281; 35, 41 nicotinic acetylcholine receptor modulators, 40, 3 nitric oxide synthase, 29, 83; 31, 221 nmr, 27, 271 nmr in biological systems, 20, 267 nmr imaging, 20, 277; 24, 265 nmr methods, 31, 299 nmr, protein structure determination, 23, 14, 36, 119, 38, sweeteners, 17, 32, 37, 1, 224; 2, 217; 3, 215; 4, 207; 5, 225; 6, 182; 7, 208; 8, 214; 9, 193; 10, 172; 13, 167; 16, 189 novel analgesics, 35, 21 nsaids, 37, 197 nuclear orphan receptors, 32, 13, 316 nucleic acid, sequencing, 16, 299 nucleic acid, synthesis, 16, 299 nucleoside conformation, 5, 272 nucleosides, 1, 299; 2, 304; 3, 297; 5, 333; 39, 241 nucleotide metabolism, 21, 247 nucleotides, 1, 299; 2, 304; 3, 297; 5, 333; 39, 241 nucleotides, cyclic, 9, 203; 10, 192; 15, 182 obesity, 1, 51; 2, 44; 3, 47; 5, 40; 8, 42; 11, 200; 15, 172; 19, 157; 23, 191; 31, 201; 32, 21 obesity therapeutics, 38, 239 obesity treatment, 37, 1 oligomerisation, 35, 271 oligonucleotides, inhibitors, 23, 295 oncogenes, 18, 225; 21, 159, 237 opioid receptor, 11, 33; 12, 20; 13, 41; 14, 31; 15, 32; 16, 41; 17, 21; 18, 51; 20, 21; 21, 21 opioids, 12, 20; 16, 41; 17, 21; 18, 51; 20, 21; 21, 21 opportunistic infections, 29, 155 oral pharmacokinetics, 35, 299 organocopper reagents, 10, 327 osteoarthritis, 22, 179 osteoporosis, 22, 169; 26, 201; 29, 275; 31, 211 oxazolidinone antibacterials, 35, 135 oxytocin antagonists and agonists, 41, 409 p38a map kinase, 37, multidrug transporter, 25, 253 parallel synthesis, 34, 267 parasite biochemistry, 16, 269 parasitic infection, 36, 99 patents in medicinal chemistry, 22, 331 pathophysiology, plasma membrane, 10, 213 pde iv inhibitors, 31, 71 pde7 inhibitors, 40, 227 penicillin binding proteins, 18, 119 peptic ulcer, 1, 99; 2, 91; 4, 56; 6, 68; 8, 93; 10, 90; 12, 91; 16, 83; 17, 89; 18, 89; 19, 81; 20, 93; 22, 191; 25, 34, 189 peptide conformation, 13, 227; 23, 285 peptide hormones, 5, 210; 7, 194; 8, 204; 10, 202; 11, 158, 19, 303 peptide hypothalamus, 7, 194; 8, 204; 10, 202; 16, 199 peptide libraries, 26, 271 peptide receptors, 25, 281; 32, 277 peptide, sar, 5, 266 peptide stability, 28, 285 peptide synthesis, 5, 307; 7, 289; 16, 309 peptide synthetic, 1, 289; 2, 296 peptide thyrotropin, 17, 31 peptidomimetics, 24, 243 periodontal disease, 10, 38, 71 pet, 24, 277 pet imaging agents, 40, 49 pet ligands, 36, 267 pharmaceutics, 1, 331; 2, 340; 3, 337; 4, 302; 5, 313; 6, 254, 264; 7, 259; 8, 332 pharmaceutical innovation, 40, 431 pharmaceutical productivity, 38, 383 pharmaceutical proteins, 34, 237 pharmacogenetics, 35, 261; 40, 417 pharmacogenomics, 34, 339 pharmacokinetics, 3, 227, 337; 4, 259, 302; 5, 246, 313; 6, 205; 8, 234; 9, 290; 11, 190; 12, 201; 13, 196, 304; 14, 188, 309; 16, , 8, 37 sleep, 27, 11; 34, 15, 69; 16, 213; 17, 203, 291 snps, 38, 249 sodium/calcium exchange, 20, 215 sodium channel blockers, 41, 59 sodium channels, 33, 51 solid-phase synthesis, 31, 309 solid state organic chemistry, 20, 287 solute active transport, 11, 222 somatostatin, 14, 209; 18, 199; 34, 209 spider toxins, 24, 287 srs, 15, 69; 16, 213; 17, 203, 291 statins, 37, 197; 39, 187 statins, pleiotropic effects of, 39, 187 stats, 31, 269 stereochemistry, 25, 323 steroid hormones, 1, 213; 2, 208; 3, 207; 4, 199 stroidogenesis, adrenal, 2, 263 steroids, 2, 312; 3, 307; 4, 281; 5, 192, 296; 6, 162; 7, 182; 8, 194; 11, 192 stimulants, 1, 12; 2, 11; 3, 14; 4, 13; 5, 13; 6, 15; 7, 18; 8, 11 stroke, pharmacological approaches, 21, 108 stromelysin, biochemistry, 25, 177 structural genomics, 40, 27, 271; 30, 265; 34, 297 substance p, 17, 271; 18, 31 substituent constants, 2, 347 suicide enzyme inhibitors, 16, 289 superoxide dismutases, 10, 257 superoxide radical, 10, 257 sweeteners, non-nutritive, 17, 323 synthesis, asymmetric, 13, 282 synthesis, computer-assisted, 12, 288; 16, 281; 21, 203 synthesis, enzymic, 23, 305 t-cells, 27, 189; 30, 199; 34, 219 tachykinins, 28, 99 target identification, 41, 331 taxol, 28, 305 technology, providers and integrators, 33, 365 tetracyclines, 37, 105 thalidomide, 30, 319 therapeutic antibodies, 36, 237 thrombin, 30, 71, 31, 51; 34, 81 thrombolytic agents, 29, 93 thrombosis, 5, 237; 26, 93; 33, 81 thromboxane receptor antagonists, 25, 99 synthase inhibitors receptors, adenosine, 28, 295; 33, 111 receptors, adrenergic, 15, 217 receptors, 14, 81 receptors, benzodiazepine, 16, 21 receptors, cell surface, 12, 211 receptors, drug, 1, 236; 2, 227; 8, 262 receptors, 23, 221, 27, 291, receptors, 28, 29 receptors, histamine, 14, 91 receptors, muscarinic, 24, 31 receptors, neuropeptide, 28, 59 receptors, neuronal bzd, 28, 19 receptors, neurotransmitters, 3, 264; 12, 249 receptors, neuroleptic, 12, 249 receptors, opioid, 11, 33; 12, 20; 13, 41; 14, 31; 15, 32; 16, 41; 17 11, 138; 14, 219; 18, 171; 21, 201; 23, 171, 181 ribozymes, 30, 285 rnai, 38, 261 sar, quantitative, 6, 245; 8, 313; 11, 301; 13, 292; 17, 291 same brain, new decade, 36, 1 schizophrenia, treatment of, 41, 3 secretase inhibitors, 35, 31; 38, 7, 39; 8, 29; 11, 13; 12, 10; 13, 21; 14, 22; 15, 22; 16, 31; 17, 11; 18, 11; 19, 11; 22, 11 sedatives, 1, 30; 2, 24; 3, 28; 4, 28; 7, 39; 8, 29; 10, 30; 11, 13; 12, 10; 13, 21; 14, 22; 15; 22; 16, 31; 17, 11; 18, 11; 20, 1; 21, 26, 287 serine proteases, 32, 71 serms, 36, 149 serotonergics, central, 25, 41; 27, 21 serotonergics, selective, 40, 17 serotonin, 2, 273; 7, 47; 26, 103; 30, 1; 33, 21 serotonin receptor, 35, 11 serum lipoproteins, regulation, 13, 14, 114 sh2 domains, 30, 227 key: cord-264350-4zxp3uae authors: kelley, james l. title: chapter 12. antiviral agents date: 1984-12-31 journal: annual reports in medicinal chemistry doi: 10.1016/s0065-7743(08)60688-0 sha: doc_id: 264350 cord_uid: 4zxp3uae publisher summary this chapter discusses the agents with activity primarily against rna viruses. the communicable diseases of the respiratory tract are probably the most common cause of symptomatic human infections. the viruses that are causative agents for human respiratory disease comprise the five taxonomically distinct families: orthomyxoviridae, paramyxoviridae, picornaviridae, coronaviridae, and adenoviridae. the influenza viruses, which consist of types a, b, and c, belong to the family orthomyxoviridae. types a and b have been associated with significant increases in mortality during epidemics. the disease may be asymptomatic or cause symptoms ranging from the common cold to fatal pneumonia. immunization against influenza has been recommended for high-risk groups and antiviral chemotherapy (amantadine) is available for the treatment and prophylaxis of all influenza a infections. there is both a great need for and interest in developing a chemotherapeutic agent for the treatment of these two viral, respiratory tract pathogens. the family picornaviridae contains the genus rhinovirus that is composed of over a hundred distinct serotypes. amantadine and rimantadine are specifically active against influenza a virus infections. the amantadine recipients reported a higher incidence of side effects largely attributed to the central nervous system (cns) symptoms. this difference in side effects may be a pharmacokinetic phenomenon that results in higher plasma concentrations of amantadine. significant progress continues to be made in the clinical use and development of agents active against dna viruses. acyclovir (9-(2-h droxyethoxymethyl)guanine) has been the subject of several reviews and of a syrnposium. considerable progress has been made in evaluating the clinical promise of acyclovir; however, there remains much to be learned concerning the best use of this drug in clinical practice. significant strides have been made in the development of clinically useful antiviral agents, especially against the dna viruses of the herpes family. most of these agents are directed against viral nucleic acid synthesis and require activation by a virus-induced thymidine kinase. researchers have begun to focus on other strategies that may produce broader spectrum anti-viral agents with different mechanisms of action. the previous review of antiviral agents in annual reports in medicinal chemistry emphasized compounds with activity against dna viruses.' primarily against rna viruses. in anti-dna virus agents is provided. antiviral agents with activity against rna viruses were previously reviewed in the 1981 volume.2 neither interferons nor interferon inducers are included in this review as they were covered in the 1981 and 1982 volumes, respectively. several reviews have been published that give an overview of the most promising clinical and experimental antiviral agents3-5 or provide background on viral diseases. 6-8 with specific drugs are cited in the sections where these compounds are discussed. the focus of this year's chapter is on agents with activity a brief update of this year's advances more comprehensive reviews dealing viral respiratory disease rna viruses are the major causative factors of the various forms of acute respiratory disease .8 respiratory tract are probably the most common cause of symptomatic human infections. among both children and adults, acute respiratory tract illness results in significant morbidity, lost time from work, physician visits, and death. it has been calculated that throughout the world over two million deaths occur annually from acute respiratory disease.9 disease comprise the five taxonomically distinct families orthomyxoviridae, paramyxoviridae, picornaviridae, coronaviridae and adenoviridae. 9 the influenza viruses, which are comprised of types a, b and c, belong to the family orthomyxoviridae. types a and b have been associated with significant increases in mortality during epidemics. the disease may be asymptomatic or cause symptoms ranging from the common cold to fatal pneumonia. immunization against influenza has been recommended for high-risk groups, and antiviral chemotherapy (amantadine) is available for the treatment and prophylaxis of all influenza a infections. the family paramyxoviridae includes respiratory syncytial virus (rsv) and parainfluenza virus which are a major cause of lower respiratory tract infections. rsv is a factor in severe bronchiolitis and pneumonia in infants and young children. efforts to develop a vaccine for rsv have been ineffectual, but recent clinical trials with ribavirin as a small particle aerosol have been promising. the parainfluenza viruses, of which there are four human serotypes, are second only to rsv as a cause of lower respiratory tract illness. there is both a great need for and interest in developing a chemotherapeutic agent for treatment of these two viral, respiratory tract pathogens. the family picornaviridae contains the genus rhinovirus, which is composed of over a hundred distinct serotypes. viruses are recognized as the most important causative agents of the upper respiratory tract illness although technical difficulties appear to have hindered studies on vaccines and antiviral agents. the adenoviruses (family adenoviridae) are a ubiquitous group of doublestranded dna viruses which are responsible for a wide variety of respiratory illnesses. these infections are most common among children, although acute respiratory disease and pneumonia are also common among military recruits. two comprehensive reviews of viral respiratory diseases and measures for their control and treatment have been recently published. 8 p 9 the development of a vaccine has been precluded due to 'coronaviruse (family coronaviridae) also has an appreciable role in agents active primarily against rna viruses amantadine (l-adamantanamine) and rimantadine (a-methvl-l-adamantanemethvlaminel -amantadine (1) and rimantadine (2) are specifically active against influenza a virus infections. in 1966 amantadine was licensed for general use against in a large-scale trial, both drugs were highly effective with no significant differences between the rates of illness or infection in the two drug-treated groups. l3 the amantadine recipients reported a higher incidence of side effects largely attributed to central nervous system (cns) symptoms. l39 l4 effects may be a pharmacokinetic phen menon which results in higher plasma concentrations of amantadine. lg a controlled study of healthy, adult volunteers found 1 and 2 to have minor side effects comparable to those of a common antihistamine. l7 hospital employees receiving amantadine showed a high incidence of cns symptoms. l8 guidelines for the use of amantadine in patients with impaired renal function have been formulated from the results of pharmacokinetic studies on amantadine in patients with normal and impaired renal fun~tion.19,~~ since approximately 90% of an oral dose of amantadine is excreted unchanged in the urine, patients with renal insufficiency can accumulate the drug, resulting in toxic manifestations. the mechanism by which 1 and 2 inhibit influenza a virus replication had previously been narrowed to a virus-specific event after virus penetration but prior to primary transcription. with radioactive precursors seems to show that uncoating is a multistep process.21 rimantadine prevents the second step of uncoatin the release of matrix (m) protein from ribonucleoproteins (rnp) .$$ this blocks the penetration of rnp into the nucleus and prevents the nuclear phase of virus reproduction. amantadine produces the same effect on uncoating as rimantadine.21 these adamantyl amines may also have utility in the treatment of other types of viral infections. been shown to be an effective inhibitor of dengue virus replication & -vitr0.~3 the amelioration of post-herpetic neuralgia due to recurrent herpes simplex sciatica24 and from acute herpes zoster25 has been reported with early administration of amantadine. ribavirin (l-~-~-ribofuranosyl-1,2,~-triazole-3-carboxamide) -this nucleoside has activity against a broad range of dna and rna viruses in an tissue culture and in animal model systems.2 analysis of the status of ribavirin 3) as an is still unresolved but may involve guanosine nucleotides and inhibition of inosine monophosphate dehydrogenase.z6 in a clinical trial against influenza a, oral ribavirin failed to alter clinical signs and symptoms of the disease.2 however, it or or has recently been reported to have a therapeutic effect against both influenza a and influenza b virus infections when administered to patients by inhalation of small-particle aerosol through a face m a s k . 2 7~~~ find utility in those patients at high risk such as the elderly and the chronically ill. ribavirin has also been shown to inhibit respiratory syncytial virus (rsv) infection in an animal model when administered i.p. or by aerosol treat~nent.~g this in vivo activity has now been substantiated in two controlled, double-blind clinical t r i a l~. 3~, 3~ administration of a continuous aerosol of ribavirin to infants hospitalized with lower respiratory tract disease from rsv resulted in significant clinical improvement.3 several antiviral agents against colorado tick fever virus (ctfv), ribavirin inhibited ctfv in vitr0.3~ ribavirin did not protect mice inoculated intracerebrally with ctfv. however, the 2',3',5'-triacetate derivative 4 significantly increased the number of survivors when administered i.p., which suggests that this lipophilic prodrug of ribavirin is able to crass the blood-brain barrier .32 administration of 2 was also found to protect mice inoculated intracerebrally with dengue virus, under conditions where ribavirin had no significant protective effect.33 in another study, ribavirin reduced the growth of four types of dengue virus in vitro, but it had no effect on virus replication in human peripheral blood leukocytes (pbl) which have been implicated in the pathogenesis of dengue virus in viv0.3~ however, a combination of ribavirin with 6-mercapto-9-(tetrahydro-2furyl)purine, an inhibitor of hypoxanthine-guanine phosphoribosyl transferase, resulted in a marked suppression of dengue virus replication in human pbl.34 significant progress continues to be made in the clinical use and development of agents active against dna viruses. (9-(2-h droxyethoxymethy1)guanine) has been the subject of several r e v i e~s~3 -~~ and of a syrnposium69 during the past year. clinical trial results have been published that attest to the efficacy of oral acyclovir in the treatment of primary7o and re~urrent7~ genital herpes simplex virus (hsv) infections and in the protection of bone marrow transplant patients from herpes infe~tions.7~ was highly effective in suppressing recurrent herpes sim lex genitalis in a group of patients with unusually frequent episodes.y3 topical acyclovir cream was effective for treatment of recurrent herpes labiali~.7~ considerable progress has been made in evaluating the clinical promise of acyclovir; however, there remains much to be learned concerning the best use of this drug in clinical practice.75 some studies on candidate prodrug forms of acyclovir have recently been published. several esters of acyclovir were reported to have improved acyclovir chronic, oral acyclovir treatment water ~olubility.7~ hydroxyethoxymethyl)purine), which is metabolically deaminated to acyclovir by adenosine deaminase,77 has been reported to be better absorbed from the gut, resulting in higher plasma levels of acycl0vir.7~ several new reports have further documented the potent antiherpetic activity of 9-(2-hydroxy-l-(hydroxymethyl)ethoxymethyl)guanine. this compound has been reported to be highly effective in reducing the severity of both primary and recurrent lesions of hsv-2 in animal m0dels.79,~~ it also appears to hold substantial promise in the treatment of human cytome alovirus (hcmv) based on its specific anti-hcmv activity & i v i t ro . 8o 9 in an animal model , 9-( 2-hyd roxy-1 -( hydroxyme thy1 ) -ethoxymethy1)guanine was shown to be a potent orally active, chemotherapeutic agent against equid herpesvirus .82 l-6-d-arabinofuranosylcytosine) has been reported to stabilize cutaneous lesions in immunosuppressed patients suffering from acute herpesvirus infection.84 a minor metabolite of fiac, 2'-fluoro-5-methyl-l-f3-~arabinofuranosyluracil (fmau), is also antiherpetic, but, in addition, it has i.p. and p.0. activity against murine leukemias resistant to 1-6-garabinofuranosylcytosine. pharmacokinetic studies on fiac and fmau have been p~blished,~5 and the synthesis and & itro anti-hsv antiviral activity of bvdu ((e)-5-(2-bromovinyl)-2'-deoxyuridine) and other 5-substituted pyrimidine nucleoside analogs has been published ,87 several 2',3'-diester and 5'-monoester prodrug forms of ara-a (9-(l-b-q-arabinofuranosyl)adenine) have been evaluated with favorable results in hsv-2 induced genital infections in female guinea pigs. 88 the 5'-monophosphate of ara-a was found to be as effective as ara-a in the treatment of immunosuppressed patients suffering from varicella-zoster .89 a carbocyclic analog of ara-a, cyclaradine, and its 5'-methoxyacetate ester have been reported to have activity in an hsv-1 encephalitis animal model that was comparable to the activity demonstrated by ara-a.go these two compounds may possess some clinical advantage over ara-a due to their low systemic toxicity. acid (pfa, foscarnet sodium), which is presently in phase 111 clinical trials in eur0pe,~3 have been discussed in a recent re~iew.9~ synthesis and structural requirements for antiherpes activity of a series of pfa esters have also been rep0rted.9~ a diaminopurine analog a134u (2,6-diamino-9-(2-fiac (2'-fluoro-5-iodoactivities of a series of analogs have been reported. k a review of the the antiviral effects of phosphonoformic the approaches to the design of antiviral agents significant strides have been made in the development of clinically useful antiviral agents, especially against the dna viruses of the herpes family. most of these agents are directed against viral nucleic acid synthesis and require activation by a virus-induced thymidine kinase. researchers are beginning to focus on other strategies which may produce broader spectrum antiviral agents with different mechanisms of action. inhibition of polyamine biosynthesis may serve as a suitable target for antiviral drug design. bis(guany1hydrazone) and u-difluoromethylornithine, an inhibitor of polyamine biosynthesis have recently been reported to inhibit replication of human cmv in vitro.43 methylation reactions. for some viruses, efficient replication is dependent on viral mrna that has been methylated at its 5i-terminal guanosine residue. effective inhibitor of this cappin mrna( g~anine-'7-)methyltransferase.~~ another permutation of this theme is preservation of cellular s-adenosylhomocysteine (sah) by inhibition of sah hydrolase. sah is an endogenous inhibitor of transmethylation the polyamine antimetabolite methylglyoxal another target is the inhibition of essential ribavirin 5'-triphosphate has been reported to be an reaction which is catalyzed by a reactions in which s-adenosylmethionine is the methyl donor .95 system i effe~t.9~997 there has been substantial interest in exploiting this system as an approach to antiviral drug development. primarily been aimed at the synthesis of 2-5a core analogs that are permeable to cells and stable to degradative enzymes. during the process of infection some viruses change the permeability of infected cells. this allows cellular penetration by compounds that are normally excluded. the selective antiviral activity of hygromycin b supports the suggestion that screening for agents which are selectively permeable to virus infected cells could result in broad-spectrum antiviral agents .g8 a review of the concept of selective delivery of antiviral agents by con jugation with protein carriers has been published. 99 this activity appears to be mediated by an obligatory lysosomal step in the uncoating of enveloped viruses. amines such as chloroquine may prevent uncoating by increasing the lysosomal ph above a value required for release of the virus nucleocapsid into the cytoplasm. a better understanding of this process could lead to the development of broad-spectrum antiviral agents. the 2-5a one of the mechanisms by which interferon exerts its antiviral these efforts have advances in virus research abstracts, 185th american chemical society national meeting abstracts, north american medicinal chemistry symposium drugs of the future program and abstracts, 23rd icaac program and abstracts, 23rd icaac sect. 111 -chemotherapeutic agents sciavolino proc. natl. acad. sci. usa proc. natl. acad. sci. usa key: cord-016912-vnx74hft authors: kornguth, s. title: strategic actionable net-centric biological defense system date: 2005 journal: defense against bioterror doi: 10.1007/1-4020-3384-2_1 sha: doc_id: 16912 cord_uid: vnx74hft technologies required for strategic actionable net-centric biological defense systems consist of : 1) multiplexed multi-array sensors for threat agents and for signatures of the host response to infection; 2) novel vaccines and restricted access antivirals/bacterials to reduce emergence of drug resistant strains preand post-event; 3) telemedicine capabilities to deliver post-event care to 20,000 victims of a biological strike; and 4) communication systems with intelligent software for resource allocation and redundant pathways that survive catastrophic attack. the integrated system must detect all threat agents with minimal false positive/negative events, a seamless integrated broad-band communications capability that enables conversion of data to actionable information, and novel preand post-event treatments. the development of multiplexed multi-array sensors, appropriate vaccines and antibiotics, and integrated communication capabilities are critical to sustaining normal health, commerce, and international activities. the overarching objectives in developing effective countermeasures to biological threats are to protect the defense community and citizenry from such threats, and to develop agile responses to unanticipated events considering that successful terrorists do the unexpected. the need for protection against and responses to biological threats has been strikingly demonstrated by the use of anthrax contaminated letters that were sent through the u.s. mail in october 2001. that attack resulted in human illness, the loss of life, and discontinuity of government operations because of contamination of federal office buildings in washington, dc. a recent report prepared by the center for strategic and international studies (csis) and supported by the defense threat reduction agency (dtra) of the department of defense (dod) came to the conclusion that the u.s. is at present not well prepared for a similar attack using anthrax1. the major problems include a lack of: 1) a clear chain of command, and 2) tools to provide the public with information that permits appropriate responses. the incidence of congo-crimean hemorrhagic fever in afghanistan, an area where coalition forces are being deployed, increases this need. the potential threat posed by emergent disease (e.g., severe acute respiratory syndrome [sars] and west nile fever virus) or from a major release of a contagious biological agent such as smallpox, has been a growing concern at all levels of the international community. this article outlines and discusses a new strategy that is needed if we are to be fully capable of sensing, preventing, and managing biological threats. the current paradigm addresses biological and chemical terrorist threats in a vertical (stove-piped) response. in the arena of developing sensors for the detection of biological agents, the paradigm has been to develop separate detectors for each agent or to develop a platform for detecting 12-24 threat agents using a single probe for each agent. there is a lack of an interactive networked communication system that is capable of managing a devastating emergent disease. to establish a highly networked system that is rapid enough to permit effective protection, it is necessary to evolve from the stove-piped, compartmentalized model currently in use to an integrated, fused, probabilistic, and frequently updated information model. multiplexed multi-array sensor systems, capable of recognizing all bacterial or viral genomic materials related to pathogenicity or of recognizing antigenic domains that are specific indicators of pathogens are one component of a network needed for rapid detection and identification of biological threats. with respect to therapeutics, modern technologies for vaccine and antibiotic production provide decided advantages over older methods. the traditional vaccines require extensive development times before they become available for human use and undesired side effects commonly result from vaccines produced by these protocols. the cost associated with developing and testing vaccines, using traditional technology, approximates 50-100 million dollars per vaccine. the dissemination of antibiotics and antivirals through the world markets has resulted in the appearance of pathogenic bacteria and viruses that are resistant to these drugs. one approach to reduce the development of antibiotic resistance is to restrict the distribution of newly developed antibiotics. such an approach presents ethical and social dilemmas. the consideration of options available for reduction of drug resistance, prior to a threat event, may permit development of a rational policy. a major problem facing our nations in the event of a biological attack or emergent disease is the large numbers of patients that can be anticipated to require medical treatment. although improvements in emergency medical care and hospital equipment have been achieved during the past two decades, the ability of any community to manage an outbreak of infectious disease affecting >10,000 people is lacking. rapid progress has been achieved whereby medical care can be provided to patients at sites that are distant from primary caregivers using telecommunication systems (e.g., the armed services in theater, large scale hmo providers such as kaiser permanente, or the correctional institutions in the u.s.). the funds needed to acquire telecommunication equipment for such distributed medical care delivery are estimated to be less than 100 million dollars for the entire u.s. at the present time such a distributed care system is not readily available. the new paradigm couples a network centric integrated sensor alert system that can detect all threat agents simultaneously, with a seamlessly integrated communication software capability that converts large scale data to actionable information. for this to be effective, the sensor system must yield minimal false positive and false negative results. the new paradigm incorporates large-scale databases on normative values of threat agents in many regions of the world so that significant changes in real time can be detected. the paradigm also includes the development and implementation of novel pre-and post-event treatment capabilities. attention must be paid to the ability of high level decision makers and operators to recognize that a new state of threat has emerged, based upon output of the sensors, data fusion system, and iconographic display. ambiguity of data, lack of an autonomous processing system, and high stress on the operator (e.g., sleep deprivation, lack of training) may all compromise the utility of a highly networked system of systems. what is needed for this new paradigm to succeed? the needs include multiplexed multi-array sensors for biological agents that infect people, livestock, and edible crops. the agents of concern include many on the militarily critical technologies list prepared for the office of the secretary of defense. we need multiplexed multi-array sensor systems with high specificity and selectivity for the rapid detection of host responses to infection. we need a new generation of effective therapeutics, including vaccines, antibiotics, and antivirals. a network centric intelligent communication system that can provide accurate comprehensible information to decision makers (from command officers to unit operators) is required. to minimize morbidity and mortality and optimize containment of disease, a biosurveillance system based on archival health databases, statistical models, and data mining strategies that can provide an early alert to a disease outbreak is required. in many cases the operator may be required to understand the meaning of acquired data in very short time periods (seconds to minutes) if the response is anticipated to change the outcome of an engagement. the tactical decision making under stress (tadmus) program is one example of such a situation. in the biological threat arena, the detection and identification of toxins require rapid analysis and operator comprehension. the large increase in numbers of sensors (for high explosives [hx], biological and chemical agents, meteorological conditions) together with the rapid changes in op tempo required to manage emergence of clinical disease would suggest a need for the development of systems capable of autonomous generation of an alert when threat conditions arise. in the sensors area, the genomes of most biological threat agents have been sequenced and the signatures of toxins described. novel multiplexed multi-array sensor-platform systems utilize the genomic datasets to detect the appearance of threat levels of these agents. in the therapeutics area, researchers are working towards identifying critical antigenic epitopes of these agents. new therapeutics can emerge that have an antigen binding capacity significantly greater than antigen-cell receptor binding, resulting in the potential for agent neutralization. technologies have been developed over the past decade for the development of new drugs and dna based vaccines. restricted access antivirals/antibacterials will need to be developed to reduce the emergence of drug resistant strains pre-and post-event. a significant development in our program at the university of texas at austin (ut-austin)2 has been the novel design and production of an antibody that binds the anthrax pa antigen 1000 times stronger (kd<10-11) than any antibody to date. the antibodies were produced using phage display technology for selection of the antibodies. in tests with experimental rodents in a controlled facility, administering the bacillus anthracis pa antigen to the animals resulted in 100 percent fatalities, whereas the co-administration of the newly developed antibody against the pa antigen resulted in 100 percent survival2. research is also being conducted to determine unique nucleic acid sequences in the genome of pathogenic bacteria and viruses that contribute to the pathogenic properties of the organisms. this information is being used to develop multiplexed assay systems that can detect selected agents simultaneously. by quickly screening for multiple pathogenicity island sequences or pathogenic factors, end-users will have the capability to detect the first signs of emergent disease without requiring screening for each particular organism. in the communications area, researchers are developing 'belief maintenance' software to provide decision makers some estimate of the validity of incoming data. an estimate of information credibility is critical to effective decision-making in crisis situations when one must rely on an 80 percent solution (i.e., 80 percent of needed information is available). waiting for a 100 percent solution could have a catastrophic impact on response effectiveness. in the area of telecommunications, researchers are developing the means to provide effective medical triage to victims in a contaminated hot zone. the hot zone in this case refers to a region experiencing an outbreak of a highly contagious disease that causes death in a significant percentage of infected individuals. advanced telecommunications technology can permit physicians and other medical experts at remote locations to provide medical information and support care delivery to personnel in the hot zone. prior training of these personnel (from physicians to local citizens) is required. an extensive use of local persons will be necessary if it is deemed inadvisable to introduce health care workers and communications experts into the hot zone; the external support teams are required to provide 'back-fill' support to the overburdened local community. biosurveillance systems are being developed to serve as an early warning of emergent disease. a variety of databases are being developed that are health related. examples of these databases include school absenteeism, over-the counter drug sales, hospital emergency clinic visits, and archival records on the incidence of diseases in different geographic regions, conus and oconus. each database must be statistically characterized regarding parameters such as variance, confidence intervals, seasonality, etcetera, and be integrated into validated predictive models. once the reference databases are in place and suitably modeled, statistically significant departures from baseline values can be detected and transmitted in real time to decision makers through intelligent communication systems. a number of critical technology gaps exist that must be addressed if we are to recognize, prevent, and minimize the effect of biological agents. these gaps include: deficiencies in the availability of multiplexed multi-array agent sensors and platforms; critical reagents; capability for large-scale production of effective vaccines, antibiotics and antivirals; ability to treat a large number (10,000) of infected people 24/7 for several weeks in a biological hot zone; archival biosurveillance databases and intelligent and secure communications networks. with the new capabilities and devices anticipated during the next decade, approaches that address these gaps include the use of autonomous (e.g., cell phone-based) microelectronic detectors for the transmission of data on agent exposure, development of novel antibodies, antibiotics and antivirals to manage disease outbreaks, and establishment of global surveillance systems for emergent diseases (e.g., sars, west nile fever, congo-crimean hemorrhagic fever). because of the broad scope of needs for technology to prevent and minimize biological threats, a number of research areas have been identified as critical. these include: the scientific validation that a biological incident has occurred (requisite tools/capabilities include situation awareness systems, sensors and signatures); the availability of medical countermeasures (vaccines, pharmaceuticals, and medical transport); and a highly effective communications network for the secure transmission of data and the conversion of such data to comprehensible information so that decision makers can take appropriate actions. for effective sensors, a variety of materials are being developed that include effective high-affinity binders of biological threat agents. the high affinity binders include antibodies, cdna gene probes, polynucleotide aptamers, and combinatorial chemicals. using phage display methods, antibody fragments can be selected that have a high affinity for agents such as anthrax toxin and brucella. another binding system that has been examined uses polynucleotide aptamers about 31 nucleotides long that have good binding affinity to ricin toxin. these sensor materials require opto/electronic transduction platforms. sensor platform research currently is being focused on micro-electro-mechanical systems (mems) devices, microelectronics technology, microfluidics (laboratory-on-a-chip), dna/protein microarrays, and transduction devices. efforts are also being directed toward the development of multiplexed multi-array systems that detect approximately 100 biological threat agents of concern. for military application, it is essential that sensor systems can detect and identify agents present in samples rapidly, using platforms that are small and have low power requirements. with current approaches, the development-to-market of new vaccines, antibiotics, and antivirals is in the order of 5-10 years. a paradigm shift to newer culture and dna-based technology is needed if we are to have an effective response to a major biological or chemical event. current estimates regarding the time required for developing and fielding new vaccines and antibiotics/antivirals to specific threat agents, using new technology and expedited approval, is in the order of three years. while current computer/informatics research includes the development of telecommunications assets, a critical need in the communications area is the development of seamless integrated communication networks. these network centric systems enable the conversion of data to actionable information. research is being conducted to provide intelligent agent software designs for such communication systems. this will enable an enhanced accuracy in critical decision-making and resource allocations. the integrated system must have redundant pathways that can survive a catastrophic attack. the communication system must be capable of integrating data on an emerging threat in a timely manner, and provide useful information for public safety coordination and perimeter management. telemedicine capabilities can aid in the delivery of post-event care to 10,000-20,000 victims of a biological strike in a densely populated area for 24 hours a day, seven days a week, for several months. in the event of smallpox attack in which 10,000 people develop clinical symptoms of infection within 7-10 days following exposure, local hospitals and medical response capabilities would be overwhelmed if current treatment protocols were used. telemedicine allows physicians at a remote location from the hot zone to provide medical support via telemedicine capabilities (visual, audio) to aid local physicians in treating patients. a treatment level of 50 patients per day per physician would require 200 physicians to provide telemedicine care for 10,000 patients. each physician would require telemedicine devices; hence 200 telemedicine devices would be required at the remote location, and a similar number in the hot zone. a national telemedicine system could include the establishment of approximately eight telemedicine response centers nationally, interconnected via satellite to telepresence and telemedicine/robotic systems. the remote care capability reduces the likelihood of the dissemination of disease to physicians and communities in which the physicians reside. a telemedicine system would also retain health care delivery in communities providing health care back-fill. the end goal is achievement of a rapid appropriate action in response to detection of threat (within minutes of threat identification). in order to meet this time constraint in a real world scenario, it is probable that the man-inthe-loop may have a denial capability rather than an approval function. the tools required for autonomous weighting of data and subsequent reduction of data elements for particular missions remain to be developed and agreed upon. the technologies required for data acquisition, communication, and autonomous processing fundamentally differ from that required for comprehension by an operator. the three technologies are systems that must be developed prior to deployment and will have known probabilities of accuracy and reliability. the fourth element involves the training and state of vigilance/alertness of the operator as well as the development of software (e.g., icons, data mining, and compression) used to display threat conditions. the point of the fourth element is to permit an operator to have rapid comprehension of the state of threat in a rapidly changing environment. because there is a time factor involved in the comprehension of threat conditions by an operator, and in the translation of the information into action, the fourth element must include temporal qualities. since the time dependence (seconds to minutes) of an intense threat situation will differ from that involved in a peacekeeping or supply distribution situation (many minutes to hours on the average), the autonomous weighting factors and data reduction factors can be expected to vary widely. this variability complicates programming of autonomous systems. command teams on the modern information-age medical delivery front face an increasing variety of cognitive stressors: information ambiguity, information overload, fatigue, and fear of contagion and quarantine. there is a requirement for a useful, predictive model of the effect of these stressors on individual and collective cognition within the medical delivery team. a model to quantify stress experienced by the caregiver and to identify countermeasures and mitigators of stress, develop organizational strategies for optimum performance under stress is needed. psychological assessments that can predict individual and team cognitive functioning and physiological markers that can determine quantitatively and objectively the effect of stress experienced on an operators vigilance have been identified ( table 1 ). the identification of specific physiological markers that are predictive of stressinduced changes in complex cognitive functioning will aid in the construction of autonomous weighting systems. lessons from the anthrax attacks: implications for u.s. bioterrorism preparedness. csis funded by the defense threat reduction agency of the department of defense. david heyman investigator protection against anthrax toxin by recombinant antibody fragments correlates with antigen affinity the research reported in this document was performed in connection with contract number daad13-02-c-0079 with the u.s. edgewood chemical biological command. 1. perception/comprehension 2. primary issue of concern 3. full situation awareness is contingent on at least four elements including: 4. large scale acquisition of data (i.e., sigint, masint, humint, sensors) 5. high fidelity communication of data sets to autonomous processing centers 6. data fusion involving weighting of data and marked reduction in data volume to yield information that provides users a common operational picture 7. rapid comprehension of time dependent information by operators facilitated by new iconographic displays, training, measures of vigilance key: cord-349066-546ozkly authors: walker, d.h. title: principles of diagnosis of infectious diseases date: 2014-08-21 journal: pathobiology of human disease doi: 10.1016/b978-0-12-386456-7.01713-5 sha: doc_id: 349066 cord_uid: 546ozkly infectious diseases are diagnosed by detection of a bacterium, virus, fungus, protozoan, or helminth in a patient with a compatible clinical illness. the methods of detection include cultivation of bacteria and fungi on growth medium, isolation of viruses in cell culture, and identification of the agent biochemically, antigenically, or genetically. infectious diseases can also be identified by detection of a specific immune response, usually antibodies, that develop during the course of illness. visualization of an agent in infected tissue can provide a diagnosis based on specific morphological characteristics or identify the category of organism, for example, gram-positive or gram-negative bacterium or virus (e.g., eosinophilic cytoplasmic inclusion bodies in neurons in rabies virus infection). methods that detect and allow visualization of antigens (immunohistochemistry) or nucleic acid sequences (in situ hybridization) provide more specific diagnoses. detection of specific nucleic acid sequences amplified by polymerase chain reaction is a powerful molecular diagnostic tool. since the elucidation of the etiology of the first infectious disease, anthrax, more than 135 years ago, koch's postulates have been applied and modified as novel technologies and agents have emerged. during the last 45 years, more than 70 previously unknown agents of infections have been identified in emerging infectious diseases, a phenomenon that is likely to continue. dh walker, university of texas medical branch, galveston, tx, usa 2014 elsevier inc. all rights reserved. contagion the ability of a disease to spread by exposure of a person to an infected patient. cytopathic effect the microscopic observation of damage to a cell in culture after infection by a pathogenic agent, usually a virus. differential diagnoses a list of potential diagnoses in a patient who develops a particular set of clinical manifestations. immunohistochemistry a method by which antigens such as those of a particular agent are detected in a tissue section by microscopy by reaction with an antibody and subsequent colorimetric or fluorescent signal-generating methods. in situ hybridization a method for the detection of a genetic sequence in tissue by annealing with a complementary nucleic acid probe coupled with colorimetric, fluorescent, or radiographic detection. opportunistic infection infection with a normally nonpathogenic organism that occurs in an immunocompromised host. prior to the work of koch and pasteur demonstrating in 1876 that anthrax is caused by the microscopic bacterium bacillus anthracis, the concept of infectious diseases did not exist. indeed, at that point, there was only one curiosity that could be called an infectious disease, namely, anthrax. robert koch's teacher, jacob henle, had published in 1840 a theoretical treatise, 'on miasmata and contagia,' in which he put forth the criteria that one would have to meet to establish that a microorganism caused disease. these criteria now known as koch's postulates are considered the fundamental steps by which an agent is proven to cause a particular infectious disease (table 1) . three centuries earlier, hieronymus fracastoro had described clearly the principles of contagion, that is, that exposure of a person to a specific disease such as measles could result in the individual's development of measles, not smallpox, and exposure to smallpox resulted in cases of smallpox, not measles. the scientific identification of viruses, bacteria, fungi, algae, protozoa, and helminths as etiologic pathogens of specific diseases then exploded after elucidation of the etiology of anthrax from the late nineteenth century to today. progress employed available tools such as microscopy, experimental animals, and microbial cultivation techniques and advanced more rapidly with the knowledge of the nature of viruses, immunologic methods, electron microscopy, and molecular methods. when a patient seeks medical attention because of illness, the physician searches for evidence to establish a diagnosis that leads to treatment that will cure or ameliorate the illness. prior to the medical encounter, the diagnostic possibilities are enormous ranging from medical and surgical diseases to the field of psychiatry. infectious diseases in theory offer the greatest opportunities for specific nonsurgical treatment of illness. the key to a favorable outcome is an astute clinician who seeks appropriate clinical, occupational, social, and travel history; considers wisely the potential epidemiological exposures; and gathers incisive data from the physical examination, radiological studies, and laboratory testing. the intermediate result is a prioritized list of differential diagnoses with focus on identifying diseases with the highest clinical consequences and those that are amenable to treatment. specific diagnoses require isolation of the agent in culture, microscopic visualization of the pathogen in tissue lesions, and/or detection of a specific host immune response to the organism. these tasks are accomplished by clinical microbiology and immunology laboratories and pathologists. bacteria comprise the group of agents of the greatest volume and therapeutic consequence that are cultivated in the microbiology laboratory. the culture results must be interpreted with knowledge of the pathogenic potential of the bacterial species isolated and the normal microbial flora at the anatomical site that was the source of the cultured sample. the recognition of contaminants and opportunistic pathogens requires specific knowledge and judgment. escherichia coli could be the cause of septic shock or enterocolitis or merely normal flora or a contaminant. on the other hand, mycobacterium tuberculosis is always a pathologically significant isolate. there are numerous approaches to the isolation of bacteria that are designed to recover particular pathogens that involve the appropriate media containing particular nutrients, salts, inhibitors of other organisms, and growth conditions (e.g., temperature, ph, and anaerobic atmosphere). selection of culture media (nonselective, selective, and differential) for a particular specimen is mostly based on the clinical history and the specimen source (blood, feces, csf, and sputum). automation of blood culture systems has greatly expedited and improved the isolation of infectious agents from blood. it is necessary to specify whether the culture requested is routine, anaerobic, or other specific microbial groups (e.g., mycobacteria) or a particular agent (e.g., brucella). some bacteria grow slowly requiring that cultures be requested to be held longer if certain organisms are suspected. some bacteria have yet to be cultivated (e.g., treponema pallidum). although a few fungi grow on some standard bacterial media, most fungi are recovered in the mycology laboratory on media designed especially for their propagation. fungi that are widely disseminated as spores in the environment and may contaminate cultures or normal yeast flora of some body sites require judgment to initiate or withhold antifungal treatment. identification of fungi has been accelerated greatly in microbiology laboratories by performing either hybridization tests or polymerase chain reaction (pcr) on media growing a fungus that is not identifiable by conventional morphological techniques such as blood culture bottles that contain yeast growth. identification of mycobacteria has also been greatly improved by the availability of dna-based tests on specimens inoculated into liquid medium. once mycobacterial growth is flagged by the automated system, molecular identification is possible, shortening the time until appropriate treatment is initiated. viral infections can be diagnosed by propagation of viruses in cell culture. the presence of viral replication in monolayers of cells in the virology laboratory is classically recognized by injury to the cells caused by the viral infection, known as cytopathic effect. many viruses do not cause microscopically detectable injury to the cell monolayer, and there is no cell type that is universally susceptible to all families of viruses. moreover, there are viruses that have yet to be recovered in cell culture. it is wise to provide the name of the patient's condition (e.g., meningoencephalitis or pneumonia) to the virology laboratory and to indicate the particular viruses that are suspected in order that appropriate cells and procedures will be employed. viruses that are recovered in cell culture can be identified by immunologic and molecular methods. direct detection by such methods has replaced a large portion of viral isolation efforts. some viruses (e.g., the herpes viruses including herpes simplex virus types 1 and 2, cytomegalovirus, varicella-zoster virus, and epstein-barr virus) cause infections that persist long after the acute illness. recovery of any of these viruses other than in the classic acute infection may reflect asymptomatic viral shedding or an opportunistic reactivation causing clinical illness. the viral quantities are greater in clinically significant conditions. the immune system, which evolved to control infectious diseases, also provides an approach to determine the diagnosis of an infectious disease. the adaptive immune response recognizes antigens of the infecting agent by production of antibodies and generation of antigen-specific t lymphocytemediated cellular immunity. the antigens that are recognized range from epitopes that are present on only one species, subspecies, or strain to antigens that are shared by agents in a genus, family, or larger group. the patient's antibodies to particular antigens that are detected and measured in the serology laboratory frequently identify the agent of the infection. various serological tests employ the methods of immunofluorescence, enzyme immunoassay, agglutination, western immunoblot, or others. the drawbacks and pitfalls of the serological diagnostic approach include the frequent absence of antibodies in serum at the time of the patient's presentation for medical diagnosis early in the course of illness prior to their production, the occurrence of cross-reactive antibodies stimulated at an earlier time by other organisms that may be misinterpreted as diagnostic of the current illness, the occurrence of persistent specific antibodies stimulated by a previously resolved infection, and misinterpretation of the reported results owing to lack of knowledge of the levels of antibodies that are generally considered diagnostic. although it is sometimes possible to observe infectious organisms by direct examination of a clinical specimen macroscopically (e.g., worms passed in feces) or microscopically (e.g., t. pallidum by dark-field microscopy of a genital lesion of suspected syphilis), diagnostic evaluation usually begins with a stained smear of sputum, cerebrospinal fluid, urine, wound exudate, or other specimen. the pathologist examines microscopic sections stained routinely with hematoxylin and eosin (h&e). large organisms such as helminths are readily observed and can be identified by their specific morphological characteristics by one with particular expertise. routinely stained sections reveal brown-pigmented fungi, which stand out from the pink and blue of eosin and hematoxylin. table 1 henle-koch postulates for the demonstration that a microorganism causes a disease 1. the organism occurs in every case of the disease in question and under circumstances that can account for the pathological changes and clinical course of the disease 2. it occurs in no other disease as a fortuitous and nonpathogenic parasite 3. after being fully isolated from the body and repeatedly grown in pure culture, it can induce the disease anew some eukaryotic protozoa and fungi can be detected and identified by the presence of pigment (e.g., some malarial parasites), subcellular structures (e.g., kinetoplasts of leishmania), or particular size, reproductive process, and cell wall structure (e.g., broad-based budding and refractile cell wall of blastomyces dermatitidis and large sporangiophores containing endospores of coccidioides immitis). some viruses cause pathological changes in the infected cells such as the formation of syncytial multinucleated giant cells and accumulations of proteins often of viral origin that are stained with eosin or hematoxylin (viral inclusions) in the nucleus or cytoplasm. recognition of the cytopathic effects and eosinophilic or basophilic cytoplasmic or intranuclear inclusions can lead to a specific diagnosis such as rabies or cytomegalovirus infection or point to a group of viruses that cause similar pathological changes such as herpes simplex viruses 1 and 2 and varicella-zoster virus. more often, infectious organisms are detected in smears of clinical samples or histological sections that have been stained by methods such as the gram stain that color gram-positive bacteria dark blue and gram-negative bacteria pink. the morphology of the bacteria as cocci or bacilli further characterizes the agent. the acid-fast stain is particularly useful for detecting mycobacteria and diagnosing tuberculosis. bacteria that stain poorly or not at all by the gram method can be visualized after silver impregnation staining (e.g., the warthin-starry method), which can detect treponema, legionella, leptospira, and indeed nearly all bacteria. a separate silver-staining technique (e.g., gomori's method) is very useful for detecting fungal yeast and hyphae. the only pathogenic fungus with a mucin capsule, cryptococcus, is identified by a stain such as mucicarmine. the most powerful tool for specific microscopic diagnosis of infectious diseases is immunohistochemistry. antibodies to antigens of particular bacteria, viruses, fungi, and protozoa have been produced, standardized, and made commercially available to react with the agents after appropriate processing of the tissue sections. after incubation of the antibody with the appropriately prepared microscopic section, the unbound antibody is washed away. only antibody bound to the infectious agent's antigens remains bound to the tissue section. the bound antibody is reacted with a series of reagents that deposit a colored precipitate on the antigen, allowing the detection, anatomical localization, and morphological analysis that has the ability to identify the agent and establish the diagnosis. the location of the agent is often very particular, and the expected size and shape of an organism add weight to confidence in the diagnosis. each infectious species contains a unique genome, dna for bacteria, fungi, protozoa, and some viruses and rna for the other viruses. transcription of the genetic dna into messenger rna results in many copies of some rna sequences. in situ hybridization is the method of detecting specific dna or rna sequences by annealing complementary dna or rna containing a label that is used to generate a visible product at the location of the target nucleic acid sequence in infected tissues. those powerful techniques have not been exploited for the diagnosis of infectious diseases as extensively as immunohistochemistry. however, they are very effective and can be used to distinguish among even closely related organisms. detection of dna or rna with nucleic acid sequence of a specific infectious agent powerful methods for the amplification of the nucleic acid sequences of few genomes of a bacterium, virus, or other agent to produce and detect millions of copies of the targeted sequence have provided a sensitive approach to the diagnosis of infections. the key to the technique is polymerase chain reaction (pcr), a heat-stable dna polymerase from organisms that reside in thermal springs. thirty or more thermal cycles of dna synthesis initiated by primers with specific dna sequences yield logarithmic increase in the particular dna fragment. for rna genomes such as those of rna viruses, a step of enzymatic conversion of rna to dna is performed using reverse transcriptase. identification of a novel, causative agent of a previously unknown infectious disease is a special challenge. in 1967, the us surgeon general stated, "it's time to close the book on infectious diseases. the war on infectious diseases is over, and we have won." it was believed for the next quarter of a century that vaccines and antibiotics had controlled infectious diseases. in reality, approximately 70 previously unknown agents of infectious diseases have been identified beginning in 1967 with marburg virus, the original agent of an african viral hemorrhagic fever. because it is clear that the emergence of infectious diseases will be a continuing occurrence, medical personnel should understand how these mysteries are solved. there are four steps in the discovery of emerging pathogens: (1) awareness of the presence of an unknown disease, (2) detection of an infectious agent in association with the disease, (3) determination that the agent causes the disease (koch's postulates revisited), and (4) determination that the etiologic agent is novel. awareness of an unknown disease may occur because of the abrupt onset of a cluster of cases of a severe illness or recognition of distinctive gross or microscopic pathological lesions. severe acute respiratory syndrome caused by a novel coronavirus is a recent example. in contrast, a syndrome of unknown etiology may have been well defined clinically for many years prior to discovery of an infectious etiology. gastric and duodenal ulcers caused by helicobacter pylori and cat scratch disease caused by bartonella henselae are examples of such diseases for which a microbial agent was unsuspected or long elusive, respectively. the methods for the initial detection of a previously unknown agent have included microscopy, bacterial culture, cell culture, animal inoculation, electron microscopy, cross-reaction of antigens in serological tests, immunohistochemistry, and pcr amplification followed by dna sequencing of the products. these are the same methods used to diagnose infection experimentally and characterize the agents. the methods used to support the etiologic role of the pathogen have included experimental animal infection, self-experimentation, laboratory accident, and demonstration of the development of a specific immune response to the agent. demonstration that the pathogen is novel includes classical biochemical phenotyping, antigenic analyses, dna or rna gene sequencing, unique ultrastructural morphology, and identification of a novel toxin. the field of emerging infectious diseases is where diagnostic infectious diseases and basic science meet to solve current biomedical infectious challenges. the roles of the astute clinician, pathologist, and laboratory technologist have been important. knowledge of microbiology, virology, immunology, molecular biology, and both cutting-edge and classic technologies has been critical to effective application to the discovery of novel pathogens. often, identification of the genus of the agent allows the selection of effective antimicrobial treatment that would have been previously unknown. the payoff of knowledge is noteworthy. pathology of infectious diseases infectious diseases: atlas, cases, text. the mcgraw-hill companies emerging pathogens: challenges and success of molecular diagnostics immunohistochemistry of infectious diseases diagnosing emerging and re-emerging infectious diseases: the pivotal role of the pathologist koneman's color atlas and textbook of diagnostic microbiology infectious disease pathology: clinical cases detection of infection or infectious agents by use of cytologic and histologic stains key: cord-020568-c5425959 authors: blatny, janet martha title: detecting and responding to bioterrorism date: 2007 journal: risk assessment and risk communication strategies in bioterrorism preparedness doi: 10.1007/978-1-4020-5808-0_7 sha: doc_id: 20568 cord_uid: c5425959 nan preparedness, 77-92. janet martha blatny incapacitate or kill man, or to destroy livestock, crops or food. toxins may to bioterrorism their production is by "dual-use" equipment. for civilian purposes such equipment is used for production of beer, yoghurt, vaccines, and antibiotics. there are several barriers in obtaining an effective bw, and two of the major challenges are (i) the development of a sufficiently virulent and infectious strain for the seed stock and (ii) the selection of the most appropriate dissemination method of the biological threat agent. the centers for disease control and prevention (cdc) has established a list of biological agents and toxins that may pose a severe threat to public health and safety (http://www.cdc.gov/od/sap). the requirements for including the agent or toxin to this "select agent list" are based on the effect on human health of exposure, the degree of contagiousness, the availability and effectiveness of medical treatment, and the vulnerability of various populations. since october 2005, the reconstructed 1918 pandemic influenza virus has been added to this list. table 1 provides examples of potential biological threat agents. the australia group has provided guidelines and control lists for national export of equipment, technology, and biological material that could contribute to bw activities (http://www.australiagroup.net/). [9] . there is a strong focus on the use of anthrax as a bw. this is due to the great stability of the anthrax spores ( 80 years), > the effectiveness as an infectious agent by inhalation, and the easy dissemination of the spores. many experts believe that biological threat agents may be more useful for obtaining panic and anxiety causing serious psychological impact instead of resulting in high preserved tissue samples [10, 11] . the 1918 flu virus killed approximately 40 million people and might be regarded as one of the most effective bioweapon known. newly emerging (e.g., sars, hendra, nipah, and avian flu) and re-emerging (e.g., west nile, human monkeypox, multidrug-resistant mycobacterium tuberculosis) pathogenic microorganisms are of global concern urging the needs for national preparedness plans, and the development and production of vaccines, antivirals, and other therapeutics. these infectious agents could potentially be used in a deliberate biological attack. three accidental escapes of the 2003 and 2004 [12] . gene sequences may be purchased from various bars and supermarkets in oregon, usa, in 1984, causing an outbreak of salmonellosis where 751 people fell ill [6] . the dissemination of rity of incidents between 1960 and 1999 using biological material in order to kill, incapacitate, or threaten, included frequent use of ricin, hiv-infected blood, and food contaminants (e.g., salmonella spp. and shigella spp.). 80 into account; the capability (technology and skills) and intention of the in order to assess a bioterrorism threat several factors need to be taken to minimize the effects of a biological attack, public health authorities need to be aware of the threat biological agents may have in biological warfare and bioterrorism. physicians need to be alerted and well-trained, have a high suspicion for these agents, and must recognize the clinical symptoms derived from such an infection. symptoms of those exposed to such agents may be nonspecific and resemble common flu-like diseases. many biological threat agents are zoonotic. animals may show the first symptoms of a clinical infectious disease after a deliberate release of a biological agent. in such cases, veterinarians may be the first to encounter the disease caused by a zoonotic threat agent. planning for necessary actions, national and global coordination, responsibility, enhanced law enforcement, medical countermeasures, and implementing efficient syndromic surveillance systems are all essential parts of bioterrorism preparedness. in addition, designing efficient detection systems for early warning, and rapid and reliable diagnostic systems contributes to improve the response efforts. the avian flu outbreak in several asian countries killing approximately 50 million chickens has revealed the need for establishing rapid molecular diagnostics for mass screening of the biological threat agents may be difficult to detect and identify quickly and reliable both from a civilian (public health) and a military point of view. there is a distinction between the terms "detection" and "identification". the establishment of the presence or absence of a biological agent is termed genes synthesis firms by e-mail requests. few companies check and compare the ordered sequences against sequences from biological threat agents and there are no national regulations requiring these firms to do so. thus, there is a concern that terrorists may order specific virulence genes and perform genetic engineering to create new or altered pathogenic microorganisms [13] . to a release of biological agents may decrease the infectious rate and the people exposed ( figure 1 ). by the time the clinical symptoms have emerged, it might be too late for treatment. in some cases, antibiotics may be effective as postexposure prophylaxis, but this treatment needs to start before the onset of symptoms. flu carriers to improve public health responses [14] . early detection 81 detecting and responding to bioterrorism responding to bioterrorism 5. detection, while identification is the determination of the precise nature of the biological agent. many systems can only detect, and not identify the biological agent. the identification system is usually dependent on specific signatures (dna, protein) of the microorganism. identification of s. typhimurium as the causing strain for the deliberate outbreak of salmonellosis by the rajneeshee cult took 4 days, but it took more than a year to identify and confirm that only a single strain of s. typhimurium had been used (in addition to the confession by one of the cult members about the deliberate release). this illustrates that, in some cases, identification may be time-consuming. many bacterial threat agents occur naturally, and some may be closely related to other bacteria found in the environment. thus, it is necessary to distinguish between terrorist events, naturally occurring outbreaks, and background levels. false positives (i.e., alarm, but no agent) may arise when the biological detector device responds to detect and identify an interfering substance in the sample (contamination), which is not the actual biological agent. if a biological agent exists, but below an instrument's threshold value for detection and identification, a false negative may occur. thus, the detection and identification schemes need to be carefully designed. a biological point detector for environmental (air) monitoring contains several components (figure 2) . a trigger may determine in real time any change in the biological background in air and discriminate between a biological threat agent and other particles in air, i.e., nonspecific detection. particle sizers may be used as a trigger, exemplified by the model 3321 aerodynamic particle sizer from tsi and the fluorescence aerodynamic particle sizer flaps2 from defence r&d canada. the flaps2 measures the intrinsic fluorescence produced by living microorganisms. using an ultraviolet (uv) laser, the wavelength 266 nm excites fluorescence from the amino acids tryptophan and tyrosine, while 355 nm excites fluorescence ranging (lidar) may also be used as a trigger and for detecting potential biological threat clouds. lidar is regarded as a detect-to-warn system false alarms. so far, lidar is not sufficiently operative during full daylight and needs good environmental conditions to function efficiently. a collector is used for concentrating the biological particles in air usually into a liquid. spores, bacteria, and viruses are usually together or detecting acpla is to collect large enough air samples through a collector. the impinger spincon® air sampler from sceptor industries collects particles at a flow rate of 450 l/min in the range of 0.2-10 µm into a liquid. omni 3000, based on the spincon® technology, and the biotrace intelligent cyclone air sampler (icas) collects air with a flow rate of 300 and 750 l/min, respectively. the biocapture 650 from mesosystems is a portable handheld air collector, suitable for first responders, sampling at a flow rate of 200 l/min. the efficiency of air collection is also dependent on the size of the particles. ffi is using both the spincon® and omni 3000 air collector for outdoor and indoor sampling of air (figure 3) . these air samples are spiked with biological threat agents for polymerase chain characterize to a certain extent the biological origin of the aerosols. even [15]. short-range lidars can detect at a radius of approximately 5 km from the instrument. most lidars use uv radiation at 266 nm or 355 nm such that biological material will fluoresce, but uv excitation may also fluoresce fuel oils, diesel, and agrochemicals causing efficient and reliable biodetection depends on the selectivity and sensitivity of the assay and system, as well as the collection and handling of the sample. from the cofactor nadh. a stand-off detector, such as light detection and attached to dust and other particles in air. thus, the term "agent-containing particles per liter of air" (acpla) has been adopted. the first step in reaction (pcr) analyses and determination of the detection limit (unpublished results). in many biodetection devices the trigger and detector have overlapping or the same functions. a detector is used to determine and 83 detecting and responding to bioterrorism usually consists of a trigger, collector, detector, and identifier. reliable and efficient though biological agents are detected, further identification of the agent is usually needed. an identifier performs specific identification of the nonbacterial atp (yeast, somatic, or free atp), and to detect spores. spores are deficient in atp and a germination step is required before there are several methods available for identifying biological threat viable cells, inspection of colony morphology, and determination of antibiotic sensitivity. such classical microbiology may identify the bacterial agent at the genus level and to a certain extent at the species level. however, these methods do not identify toxins, they are time-consuming, and not suitable for first responders. immunoassays include the use of specific antibodies targeted against a toxin or a particular antigen at the surface of a bacterial cell or spore. immunological methods usually provide quick results and are suitable for fast screening of a large number of samples. however, the method is less specific and sensitive, and the detection limit may be a 100-1,000-fold devices are the bioveris detection system, meso scale discovery sector pr, and luminex 100. assay combined with specific phage-associated lytic enzymes may be used for further identification of the bacteria. performing the bioluminescence assay [16, 17] . the bioluminescence higher than the infective dose [20, 22] . in general, immunoassays are good for presumptive detection but confirmatory analysis is needed, usually by nucleic acid-based detection methods. antibody specificity and affinity are often the limiting factors of immunoassays. tetracore's biological threat agents may be present as vegetative cells, spores, or in a dormant state (viable but nonculturable state; vbnc) in environmental detection of viable bacterial cells (bioluminescence assays). some of these assays have been further improved to separate bacterial atp from samples (such as water, air, and soil). atp is frequently used for nonspecific biothreat alert test strips are reagent strips using a lateral flow immunochromatography technique allowing biological threat agents to be identified within 15 min. examples of commercially available immunological agents, but there is no single approach for identification. definitive identification requires several methods; conventional culture-based methods as well as clinical diagnosis of those exposed to such agents. the cultivation of bacteria in selective growth medium allows detection of 85 immunoassays, and nucleic acid-based methods (reviewed in [18] [19] [20] [21] ), detecting and responding to bioterrorism 8. real-time pcr is the most commonly used nucleic acid-based method for specific and sensitive identification of biological threat agents. pcr may detect as low as 10-100 cells, but this method usually requires a clean sample. disruption of bacterial cells and spores is often needed in order to make the dna available for amplification in the pcr assay. an effective sample preparation may also reduce the presence of false negatives since impurities from the sample may inhibit the pcr assay. specific identification by pcr is obtained by using specific primers and probe combinations. each probe (e.g., taqman probes, fluorescence resonance energy transfer [fret]-prober, and molecular beacons) has a different method of separating the fluorophore from the quencher when reporting the amplification process. different reporter dyes (fluorophores) may be attached to the probes allowing simultaneous identification of several biological threat agents (multiplex identification). the pcr assay should include internal controls to avoid false positive signals. internal controls may consist of either a plasmid or a dna fragment in which the amplified dna sequence is several real-time pcr assays have been outlined for a number of biological threat agents, and commercial kits containing the specific reagents are available. the target genes and regions for pcr identification are specifically chosen for each microorganism. for bacillus anthracis, several genes located on the virulence pxo1 and pxo2 plasmids, respectively. however, plasmid-free b. anthracis cells will only be identified by pcr when using specific chromosomal targets. the closest relative of b. anthracis is the opportunistic human pathogen b. cereus (soil bacterium) and the insect pathogen b. thuringiensis. these are functionally different, mainly distinguish by plasmid-encoded genes. the genomes of b. anthracis, regions that may be suitable as specific targets for pcr identification unique in the assay [23, 24] . genetic targets located on the chromosome are used for identification [25] [26] [27] in addition to the well-known target lef, cya, pga toxin, and cap capsule b. cereus, and b. thuringiensis strains show a close similarity [28] complicating the search for unique chromosomal targets for thier differentiation and identification. however, genome sequencing of many biological threat agents has provided significant data about unique [29] . ffi has identified one unique chromosomal target for specific identification of b. anthracis (unpublished results). several real-time pcr smartcycler from cepheid and the lightcycler from roche applied assays for identification of various biological threat agents listed in table 1 have been established at ffi (exemplified in figure 4 ) using either the 86 8.2. nucleic acid-based methods science ( figure 5 ). idaho technologies and smiths detection (bio-seeq) have developed handheld pcr devices suitable for military use and first responders, such as ruggedized advanced pathogen identification device (rapid), and razor, and bio-seeq, respectively. bruker daltonik gmbh has constructed a microarray system based on pcr for bioidentification. for review of various nucleic acid detection assays and systems see [18, 19, 21] . tm (lightcycler ) using specific primers and probes (19) . pcr can detect dna from both viable and dead cells, and thus, culturebased methods are needed for the assessment of bacterial viability. nucleic acid sequence-based amplification (nasba) is a method in which rna biodetection equipment for use in a battlefield is different from the use in a civilian community. if deployed in the right place and at the right time, valuable information of a bioterrorism event would be provided. the joint biological point detection system (jbpds) is an automatic point detector for real-time identification of biological threat agents within 15 min and suitable for military use. it contains a trigger, collector, detector, and identifier, and will be used by the us air force and marine corps. integrated systems for identification of various biological threat agents in large areas (e.g., arenas, airports) and postal service systems have been constructed, such as the autonomous pathogen detection system (apds) from lawrence livermore national laboratory and the biohazard detection system (bds) from northrop grumman, respectively. other biodetection techniques in advance are light scattering surface plasmon resonance (lsspr), surface-enhanced raman spectroscopy (sers), and matrix-assisted laser desorption/ionization time-of-flight mass spectroscopy (maldi-tof-ms). there have been many attempts to develop biosensors based on electrochemics, micro-fluidics, nanocrystals (quantum dots), and optics, combined with immuno-and nucleic acid-based assays, classification of bacterial strains is often based on the identification of dna polymorphisms. when the genetic diversity within a bacterial species is high, it is often adequate to sequence only a few number of dna instead of dna is amplified (reviewed in [30] [31] [32] ). nasba can be used to detect viable cells since mrna is specifically detected and amplified. the design of specific primers and molecular beacons is crucial for the nasba assay. nasba has been widely used for virus diagnostics, and only few reports describe the use of this technique for bacterial detection. this technique has been used together with liposomal-based biosensors to identify b. anthracis, escherichia coli, and cryptosporidium parvum [33] [34] [35] . ffi has used nasba for identification of vibrio cholerae in water samples [36] . but only few are commercially available (reviewed in [37, 21] ). integrated and advanced detection methods 9. fragments in order to classify the strain. in contrast, strains belonging to more homogenous species, in which little sequence divergence has occurred, it is necessary to sequence long dna regions or to analyze several loci with high mutation rates. variable number of tandem repeats (vntr) is a linear arrangement of multiple copies of short repeated dna sequences biological threat agents for the use in biological warfare or bioterrorism are infectious microorganisms or toxins with the intent to incapacitate or kill man, or to destroy livestock, crops, or food. an essential part of bioterrorism preparedness and response includes the design of efficient and reliable systems for detection and identification of biological threat agents. various biodetection systems for environmental monitoring are available. many of these systems have been primarily constructed for military use. there is no single approach for identification of biological threat agents, and several methods are needed for verification. classical microbiology, immunoassays, and nucleic acid-based methods, including molecular forensics, are laboratory approaches for detecting, identifying, and verifying various biological threat agents. to identify the b. anthracis ames strain used in the anthrax attacks to the origin of the bacterial agent [39] . vntr analysis was used from sample collection to molecular typing. mlva techniques have already been established at ffi and are used for genetic fingerprinting of b. cereus and v. cholerae strains (unpublished results). the size of the dna fragments containing vntrs is measured by pcr. most bacterial genomes contain several vntrs, and multi-locus vntr (mlva) analysis is now a suitable tool for strain typing and for tracing back in the united states in bioterrorism: from threat to reality the abcs of bioterrorism for veterinarians, focusing on category a agents sensitive detection of bacteria and spores using a portable bioluminescence atp measurement assay system distinguishing from white powder materials usamriid's medical management of biological casualties handbook smittsomme sykdommer fra mat. høyskoleforlaget large community outbreak of salmonellosis caused by intentional contamination of restaurant salad bars the national anthrax epidemiologic investigation team. investigation of bioterrorism-related anthrax bioterrorism and threat assessment. the weapons of mass destruction commission characterization of the 1918 influenza virus polymerase genes characterization of the reconstructed 1918 spanish influenza pandemic virus the 1918 flu virus is resurrected the bioweapon is in the post the necessity of molecular diagnostics for avian flu set lasers to…detect. nbc international rapid atp method for a review of molecular recognition technologies for detection of biological threat agents rapid diagnostic assays in the genomic biology era: detection and identification of infectious disease and biological weapon peruski ah, peruski jr. lf. immunological methods for detection and identification of infectious disease and biological warfare agents current and developing technologies for monitoring agents of bioterrorism and biowarfare multi-pathogens sequence containing plasmids as positive controls for universal detection of potential agents of bioterrorism preparation of a positive control dna for molecular diagnosis of bacillus anthracis identification of anthrax-specific signature sequence from bacillus anthracis use of 16s rrna, 23s rrna, and gyrb gene sequences analysis to determine phylogenetic relationships of bacillus cereus group microorganisms real-time pcr assay for a unique chromosomal sequence of bacillus anthracis bacillus anthracis, bacillus cereus, and bacillus thuringiensisone species on the basis of genetic evidence a genomics-based approach to biodefence preparedness nasba and other transcription-based amplification methods for research and diagnostic microbiology characteristics and applications of nucleic acid sequence-based amplification (nasba) the use of nasba for the detection of microbial pathogens in food and environmental samples biosensor for the detection of a single viable b. anthracis spore nucleic acid approaches for detection and identification of biological warfare and infectious disease agents hort-sequence dna repeats in prokaryotic genomes multiple-locus variable number tandem repeats analysis for genetic fingerprinting of pathogenic bacteria multiple-locus variable-number tandem repeat analysis reveals genetic relationships within bacillus anthracis the genome sequence of bacillus anthracis ames and comparison to closely related bacteria toward a system of microbial forensics: from sample collection to interpretation of evidence biosensors for the detection of bacteria detection of viable vibrio cholerae cells by nasba universal nucleic acid sequence biosensor with nanomolar detection limits a rapid biosensor for viable b. anthracis spores key: cord-354130-mi7saerx authors: compton, susan r.; homberger, felix r.; clark, judy macarthur title: microbiological monitoring in individually ventilated cage systems date: 2004 journal: lab anim (ny) doi: 10.1038/laban1104-36 sha: doc_id: 354130 cord_uid: mi7saerx housing rodents in ivc racks has many advantages over conventional cages but also presents unique challenges related to health monitoring. the authors review the issues to consider in design of a sentinel program using ivc systems. there exist many methods and strategies for monitoring rodent colonies for microbial agents. several general principles are applicable when developing and assessing the efficacy of a microbiological monitoring program, irrespective of the type of caging used for rodent housing. first and foremost, the monitoring program must specifically target the infectious agents of concern. in particular, there should be consideration of the route and duration of transmission of the infectious agents, the probability that the agents will be found, and the likelihood that the agents will cause disease or affect research results. second, the monitoring program should make the most efficient use of personnel and resources. finally, the outcome of a successful monitoring program should aid research, not hinder it, and should result in animals that are sufficiently pathogen free for research and should facilitate exchange of animals between institutions. rodents in contemporary animal colonies live in several types of caging. the three types of caging systems commonly used are open-top cages, static isolator cages, and individually ventilated cages (ivcs). the type of housing used affects the ease with which infectious agents can be spread between rodents housed in separate cages. rodents housed in open cages and serviced in the open have the highest risk of transmitting infectious agents from cage to cage via aerosols and fomites. this is one reason why this type of housing is being eliminated in many facilities. rodents housed in static isolator cages have a lower risk of infectious-agent transmission than do rodents housed in opentop caging, because the filter lid serves as a physical barrier to infectious-agent transmission, although transmission can occur if filter lids are not properly installed 1 or during cage changes. it has also been shown that the exhaust gases do not actu-ally pass through the filter material in the lid but escape through the gap between the cage and the cage lid 2 . this is the reason that many facilities have instituted the use of ventilated cage change stations for husbandry and experimental manipulations of rodents housed in static isolator cages. ivcs differ from static isolator cages in several ways. each cage receives a supply of hepa filtered air at a rate of between 30 and 120 air changes per hour (ach). this results in lower temperature and humidity as well as lower concentrations of ammonia and carbon dioxide than in static isolator cages 3 . some ivc systems balance the air supply with the air exhaust system, thus permitting control of differential pressure at the cage level. thus each cage, theoretically at least, is its own biocontainment zone. an alternative approach in some ivc systems is to seal the cages to achieve biocontainment (instead of balancing differential pressure). this approach has, however, the obvious disadvantage of being potentially lethal in the event of cage ventilation failure 4, 5 . rodents housed in ivcs have the lowest risk of cage-to-cage spread of infection [6] [7] [8] [9] [10] [11] (j. schmidt, personal communication). transmission between animals housed in ivc systems generally occurs as a result of dissemination during husbandry procedures or experimental manipulations by investigators, thus justifying the use of ventilated cage change stations for such activities. nevertheless, it is essential to exercise great care to ensure against transmission by fomites on hands or equipment, such as watering valves, particularly when active infections are present 11, 12 . monitoring methods, such as placement of room sentinels adjacent to the hvac exhaust duct, as used traditionally with mice housed in open-top caging, require reassessment with regard to their efficacy in static isolator cages and in ivc systems. tion of the testing methods (e.g., serology on immunodeficient animals). the use of sentinel mice to monitor a colony decreases the number of mice needed to detect an infection within a colony, because each sentinel can serve in monitoring of many colony mice. there are three types of sentinel animals. first, contact sentinels (i.e., mice housed in the same cage as the colony mice to be monitored) are highly reliable at detecting infectious agents transmitted by all routes (air, feces, urine, wounds, contact, etc.). second, soiled-bedding sentinels (i.e., mice housed on soiled bedding that has been removed from several cages of colony mice) are most effective at detecting agents that are transmitted in feces or urine. third, exhaust air sentinels (i.e., mice that are exposed to rack exhaust air) are effective at detecting agents that are transmitted in respiratory excretions. lipman and homberger 13 provide an overview of the advantages and disadvantages of the three types of sentinels. finally, direct detection of infectious agents present within the environment of a colony is feasible by swabbing such surfaces as cage and cage racks, by taking appropriate samples from rack exhaust air, and so on. several methods exist for testing samples collected from mice or from their environment. traditionally, most viral infections are detected using serological assays. the advantages of serological assays are that they are inexpensive and can provide a historical picture of which agents are present in the colony (i.e., at any time after a mouse becomes seropositive, antibodies to the rel-evant agent can be detected). the primary disadvantage of serological assays is that mice infected with an agent may not become seropositive until 2 or more weeks after the infection's initiation. therefore, in the midst of an outbreak, many mice may become infected and may potentially be transmitting virus to other mice, yet may still be seronegative. additionally, immunocompromised mice, especially those with b-cell deficiencies, may produce few or no antibodies in response to an infection. recently, molecular methods have augmented our arsenal of testing methods. molecular methods, in particular pcr and rt-pcr, detect the nucleic acids present within a bacteria or virus. in general, these methods are highly sensitive and rapid, and can be very effective means to determine which animals are infected with a particular agent. molecular methods are particularly useful during outbreaks, when it is essential to determine rapidly (within a few hours) the location of infected mice within a rack, so as to develop a strategy for containing the spread of the agent. the limitations of molecular methods are that they are relatively expensive, they can yield both false negative and false positive results, their high sensitivity makes them prone to cross-contamination, and many substances found in blood, feces, and other animal tissues can function methods such as exposing sentinels to soiled bedding, as used traditionally with mice housed in static isolator cages, also require reassessment for their applicability in ivc systems. moreover, we should add to our monitoring approach new methods specifically designed to exploit ivc system characteristics by the sampling of air exhausted from the ivc system. there are two ways of monitoring the exhausted air, either from individual cages or from the whole ivc rack. small gauze filters placed on the cage exhaust opening, on the exhaust opening from a zone of the rack (e.g., a row or column of cages), or on the exhaust manifold of the entire rack can permit periodic monitoring of the effluent air. after removal from the rack, molecular pcr-based methods can detect infectious agents on the filters. additionally, there can be continuous monitoring of exhaust air from the ivc rack (or a selected zone of the rack) that is achieved by testing of sentinels housed in a specially designed cage (bioscreen system; biozone inc., fort mill, sc, and biozone ltd., margate, uk) that receives as its air supply a portion of the mixed exhaust air from several cages on the rack, before hepa filtration (fig. 1) . the effectiveness of such a system is clearly dependent on the uniformity of ventilation of all cages on the rack so as to be confident that a low incidence of infection will be represented in the exhaust air without regard to the location on the rack of the infected animals 12 . there are many ways of detecting an infectious agent that has been transmitted to one or more rodents housed in ivcs. the most direct method of surveillance is to monitor the colony mice themselves for evidence of an infection. this method seems highly reliable on first consideration, but it presents significant drawbacks. if the animals are properly screened, considering that each cage is essentially a self-contained unit, then it is by far the least cost-effective method, because it requires the largest investment of mice, staff, and testing reagents. moreover, the age, genotype, and experimental use of the animals may interfere with the applicaas inhibitors of the enzymes used in these assays, causing false negative results. finally, the nucleic acids detected may be present in a noninfectious form of the agent, such as in inactivated material present on the surface of cages or change stations. detection of bacteria usually follows culture of fecal, cecal, nasopharyngeal, or other tissue samples on specialized agar plates. specific identification of pathogenic bacteria after culture usually involves a combination of enzymatic and fermentation tests. microscopy is the usual method for visual detection in diagnostic specimens of endoparasites and ectoparasites. there are several infectious agent-specific variables that affect the ease with which an agent can be detected in mice housed in an ivc system. the first influence is the route(s) by which the agent is shed from the animal ( table 1 ). agents that are transmitted only through animal-animal contact, such as mites, are very difficult to detect in mice housed in an ivc system. it is possible to detect these agents directly in colony animals or using contact sentinels. one can effectively detect agents that are transmitted via the fecal-oral route, such as mouse parvovirus (mpv), using contact sentinels, sentinels exposed to soiled bedding, or molecular testing of feces from colony animals, but exhaust air sentinels are not effective at detecting these 12 . in contrast, the use of contact sentinels, exhaust air sentinels, or molecular testing of cage surfaces is effective in detecting agents that are shed in respiratory secretions, such as sendai virus, but sentinels exposed to soiled bedding are ineffective at detecting these 12 . the methods used to sample the cage or rack surfaces for air-borne infectious agents (cage wipe material or protocols, filter material, or placement location) may influence the effectiveness of detecting the agent of interest. for example, our lab detected sendai virus rna on the cages of experimentally infected mice for 2 weeks using calcium alginate swabs wet with saline, whereas another group was able to detect sendai virus rna on the cages of experimentally infected mice for only 3 days using alcohol wipes 12, 14 . a second major influence that affects the ease with which an infectious agent can be detected in mice housed in an ivc system is the infectious-agent load to which the sentinel mice are exposed. infectiousagent load depends on the duration of shedding, on the concentration of agent shed, and on the stability of the agent after it is shed ( table 1) . detection of agents such as murine rotavirus (epizootic diarrhea of infant mice, or edim), which causes an acute infection in adult immuno erally shed at high concentrations, are much easier to detect using all monitoring methods, but the reliability of detection is still a problem [10] [11] [12] . the stability of the infectious agent in the environment also influences the difficulty of its detection ( table 1) . for agents that persist as infectious particles in the environment for only short periods of time, such as most enveloped viruses, detection is more difficult than for agents that are stable as infectious particles for long periods in the environment, such as parvoviruses or pinworm eggs. labile agents can be missed if the timing of exposure by contact, exhaust air, or soiled bedding does not coincide with the period of shedding. even though many enveloped viruses, such as coronaviruses, lose their infectivity upon drying and rupture of the envelope, the nucleic acids inside the viral particles remain detectable 15 . therefore, molecular assays can detect the recent presence of agents even after they lose their infectivity and transmissibility to mice within the colony. for monitoring methods that detect infectious agents in the exhaust air, the size of an agent should theoretically affect its detectability. exhaust air and particulates should carry smaller agents such as viruses more effectively than larger agents such as parasites ( table 1) . for example, mpv and helicobacter hepaticus both cause chronic intestinal infections and are shed in feces, yet mpv, a 20-nm virus, was detectable on filters placed in the rack exhaust air, whereas h. hepaticus, a 5-to 10-µm bacterium, was not 12 . there are several system-specific conditions that may influence the efficacy of microbiological monitoring of mice housed in an ivc system. the air change rate that is achievable in an ivc rack is much higher (30-120 ach) than that achieved in static isolator cages (>5 ach) 2 . a recent study showed that the higher air change rate in an ivc system compared with static isolator cages resulted in lower relative humidity 3 . furthermore, the mean bedding weight gain per mouse was 50% greater in static isolator cages than in the ivc system. the less humid cage environment present in the ivc system could result in the dehydration and inactivation of many infectious agents. this effect, together with the dilution of organisms in the exhaust air by the higher ventilation rate, could lead to decreased transmission of agents to colony and sentinel mice. the decreased intercage spread of an infectious agent means that infections are sporadic and confined to just a few cages at a time, and it is therefore essential to use an adequate sample size when monitoring mice housed in ivc systems. several studies have shown that the high airflow achieved in ivc systems results in lower ammonia levels than those seen in static isolator cages at 5-7 days after cage change 3,16-18 . the lower ammonia levels measured in ivcs are probably the direct result of lower relative humidity, because high humidity has been linked to proliferation of urease-positive bacteria that can convert urea to ammonia 19 . furthermore, the type of bedding used in ivc systems can affect the rate of ammonia accumulation, because some bedding types may contain endogenous ureases 20 . the frequency of cage change generally depends on the time required for intracage ammonia concentrations to rise to a level considered irritating to the mucous membranes of husbandry personnel (generally >25 p.p.m.) 21 . because ammonia levels rise more quickly in static isolator cages, it is generally necessary to change static isolator cages at least once a week, whereas mice housed in ivc systems may need changing only once every 2 weeks. less frequent cage changing decreases demands on personnel time, decreases the quantity of bedding used, increases cage longevity, and may decrease pup mortality 22 , but it can also decrease the efficacy of soiled-bedding monitoring. if husbandry personnel add soiled bedding to sentinel cages only during routine biweekly cage changes, then this reduces the number of times that soiled bedding will be added to the sentinel cage in the ivc system as compared with static isolators, and agents pre-competent mice, with shedding of virus for ∼5 days, can be very difficult. direct tests for the agent in the animal or the environment must occur during the short window of shedding and therefore are quite unreliable at detecting a low-level infection within a colony. detection of edim using serology and sentinels exposed to soiled bedding has the limitation that the timing of the soiled-bedding transfer must coincide with the short period of shedding to be effective. in ivc systems, a biweekly cage-changing schedule, with biweekly addition of soiled bedding to sentinel cages, has the disadvantage that the bedding transferred may or may not contain infectious virus. for agents that cause acute infections, methods that monitor the mice continuously over an extended period, such as seroconversion of colony mice, contact sentinels, exhaust air sentinels, or long-term placement of gauze filters in the exhaust air stream, are most effective. frequent monitoring, using sentinels exposed to soiled bedding or shortterm placement of gauze filters in the exhaust air stream, can also be effective. periodic monitoring methods, such as sentinels exposed to soiled bedding, and continuous monitoring methods, such as exhaust air sentinels, are both likely to be effective at detecting agents that cause chronic or persistent infections. the concentration of an excreted infectious agent influences the difficulty of its detection. agents, such as mpv, that cause a chronic infection with low-level fecal shedding, can be difficult to detect. detection of these agents necessitates testing an adequate sample size. a sample size that is too small or a monitoring protocol that samples a nonrepresentative group of cages can result in sporadic detection of the agent within a colony. for example, in a recent study, mice that received a single dose of soiled bedding from a group of cages in which 20% of the cages housed mpvinfected mice seroconverted, whereas those that received several doses of soiled bedding from a group of cages in which only 5% of the cages housed mpv-infected mice did not seroconvert 12 . agents, such as mouse hepatitis virus (mhv), that are gen-volume 33, no. 10 lab animal november 2004 sent in the soiled bedding have a longer time in which to become noninfectious before being exposed to susceptible sentinel animals. it will be necessary to optimize soiled-bedding sentinel monitoring protocols used in ivc systems so they can accurately detect agents if they are present. this means following a strict protocol to avoid bias by regularly sampling all cages on the rack, and one should also consider supplemental sampling of soiled bedding between cage changes. both of these have implications for labor and may have the negative effect of increasing the probability of cageto-cage transmission during soiled-bedding collection. the rate at which a particulate accumulates on a filter may also affect the sensitivity of filter-based tests. agents such as mpv are present at low concentrations in exhaust air. recently, we reported that mpv dna was not detectable on filters placed on the rack exhaust filter for 24 h, but mpv dna was detectable on filters placed on the rack exhaust filter for 2 weeks 12 . the positioning of filters within the ivc rack can also influence the chance that an agent will be detected. for example, mpv, mhv, and sendai virus were all detectable on filters placed on the lids of cages housing infected animals for 24 h at 6 days after inoculation, but only mhv and sendai virus were detectable on filters placed on the rack exhaust filter for 24 h at 6 days after inoculation 12 . this indicates that the closer to the source of the infection, where the infectious-agent concentration in the exhaust air is highest, the greater the likelihood of detecting it. it will be necessary to determine for each ivc system the optimal positioning of test filters and the time required to accumulate adequate particulates, which serve as fomites to transport infectious agents through the air, and should consider the ventilation pattern of the system. theoretically, any location within the exhaust system of the ivc rack that a small (1 cm 2 ) filter can be temporarily placed can serve as a test site. the type of bedding used in ivc cages could affect the rate of particulate generation and accumulation on filters and therefore could affect air-testing efficacy. aspen chip bedding has been found to be substantially less dusty than wood fiber or straw pelleted bedding 23 . theoretically, the greater the amount of dust generated from bedding, on which agents can be carried through the air, the greater the chance of detecting the agent on a filter. the accuracy of monitoring methods involving exhaust air sampling is dependent on the uniformity of the airflow achieved in the ivc rack and thus on the ivc rack design. when infections are present in only a small number of cages on an ivc rack, as is often the case, the reliability of exhaust air monitoring, whether using gauze filters placed on the rack prefilter or using exhaust air sentinels, is likely to be highly dependent on the uniformity of exhaust airflow within the rack. nonuniform airflow resulting in large variations in exhaust airflow from individual cages may be a cause of inaccurate or unreliable sampling of cages housing infected mice seen in some studies (b. tiemann, personal communication). furthermore, some ivc systems have a single manifold that is used for supply air only, and exhaust air simply escapes from the cage into the room. in such ivc systems, it may be impossible to appropriately sample exhaust air either by using gauze filters or with exhaust air sentinels. one can operate ivc racks using either a positive or negative air pressure differential. under a positive air pressure differential (bioexclusion mode), the supply fan for the rack pushes more air into the rack than is pulled out by the exhaust fan, resulting in a small amount of air leaking out of the cages into the animal room. under a negative air pressure differential (biocontainment mode), the supply fan pushes less air into the rack than is pulled out by the exhaust fan, resulting in a small amount of air being sucked into the cages from the animal room. theoretically, different air pressure differentials might affect the efficacy of air monitoring methods, but in our experience the efficacy of microbiological monitoring was equivalent under positive and negative air pressure differentials 12 . it is important to remember that in an ivc system, the primary means of infectious-disease spread within a rack occurs during husbandry procedures or experimental manipulations. during an outbreak, monitoring exhausted air from individual cages using cage-top gauze filters, or monitoring exhausted air from the entire ivc system using filters placed on the rack exhaust manifold, can be highly efficient in determining the extent of an infection within a rack or within a facility, and in confirming elimination of an agent after such an outbreak. both of these filter-based methods do not require handling of potentially infected mice and can produce results within a few days. this can be very valuable when handling animals for sentinel bleeding or for collection of soiled bedding could lead to spread of the infection. rodent microbiological monitoring continues to be a highly dynamic, nonstandardized, and continually evolving process. monitoring strategies and approaches to diagnostic sample collection are confounded regularly by the convergence of innovative husbandry refinements, new developments in a wide array of caging systems, evolving pathogens, and novel rodent genotypes. in this environment of constant change, and often in the face of relatively meager financial resources for assessment of these mitigating influences, the validation of monitoring approaches has proved to be a challenge. consequently, we recommend a multifaceted approach to monitoring for infectious agents in ivc systems. each monitoring program should consider both infectious agent-related and equipment-related variables. the routes of transmission for the infectious agents to be monitored, the duration and infectiousload shed, and the stability of the agent in the environment should all inform the decision as to which approach to monitoring is appropriate. likewise, the frequency of testing should specifically target the battery of agents to be tested and the type of caging used. it is also essential to understand and consider airflow rates and uni-formity of airflow in each type of ivc system. until techniques and technology, still to be developed, simplify the monitoring process, a fully integrated program of microbiological monitoring should make use of a sensible combination of soiledbedding sentinels, contact sentinels, and, where possible, exhaust air monitoring. even a broad sentinel-based monitoring program, however, is unlikely to be fully reliable and completely sound. consequently, the most comprehensive monitoring program should also include health evaluations of colony rodents that present as clinical cases. received 6/11/04; accepted 8/11/04. the international joint meeting twelfth iclas general assembly & conference, seventh felasa symposium assessment of static isolator cages with automatic watering when used with conventional husbandry techniques as a factor in the transmission of mouse hepatitis virus efficacy of three microbiological monitoring methods in a ventilated cage rack rodent quality assurance testing: use of sentinel animal systems detection of sendai virus and pneumonia virus of mice by use of fluorogenic nuclease reverse transcriptase polymerase chain reaction analysis reverse-transcriptase polymerase chain reaction-based diagnosis and molecular characterization of a new coronavirus strain isolator rodent caging systems (state of the art): a critical view the effects of intracage ventilation on microenvironmental conditions in filter-top cages micoenvironment in ventilated animal cages with differing ventilation rates, mice populations and frequency of bedding changes ammonia build-up in animal boxes and its effect on fat tracheal epithelium characterization and qualification of microenvironmental contaminants in isolator cages with a variety of contact beddings american conference of governmental industrial hygienists. threshold limit values for chemical substances and physical agents and biological exposure indices the impact of reduced frequency of cage changes on the health of mice housed in ventilated cages the bedding of laboratory animals as a source of airborne contaminants the influence of filter top caging on the transmission of pinworm infections in mice an evaluation of intra-cage ventilation in three animal caging systems comparison of environment and mice in static and mechanically ventilated cages with different air velocities and ventilation designs failed air supply to individually ventilated caging system causes acute hypoxia and mortality in rats carbon dioxide concentrations in unventilated ivc cages evaluation of isolator caging systems for protection of mice against challenge with mouse hepatitis virus effectiveness of pressurized individually ventilated (piv) cages in reducing transmission of pneumonia virus microbiological validation of the m.i.c.e. caging system performance evaluation of a positive/negative ventilated murine rack system following direct and indirect aerosol exposure to bacillus subtilis spf status of mice can be maintained in spfmice in ventilated cage systems key: cord-315617-mhm9wh9q authors: gottschalk, rené; preiser, wolfgang title: bioterrorism: is it a real threat? date: 2004-09-02 journal: med microbiol immunol doi: 10.1007/s00430-004-0228-z sha: doc_id: 315617 cord_uid: mhm9wh9q the geneva protocol of 1925 commits the signatory nations to refraining from the use of biological weapons. however, the terrorist assaults of september 2001 and, subsequently, the anthrax-containing letters are cause for great concerns: new threats to the security of nations are expected, as terrorist organizations seem to increasingly explore novel ways of spreading terror. in this context, naturally emerging diseases such as sars, monkeypox or west nile fever assume new importance because it is difficult to distinguish between natural epidemics and possible bioweapon assaults. great efforts on the part of governments and public health authorities are necessary to counteract these threats. the geneva protocol of 1925 relating to the protection of civilian persons in time of war and additional protocols commit the signatory nations to refraining from the use of biological weapons since these not only have disastrous effects on the armed opponents involved in the conflict but also on the civilian population [26] . nevertheless, efforts to develop such weapons continued throughout the cold war and even after its end; not least because compared to nuclear weapons, biological ones are relatively cheap to develop and produce, earning them the attribute ''a poor man's atomic bomb''. however, it is the developments over the past years that are causing the greatest concern: new threats to the security of nations are emerging in the form of terrorist organizations that seem to increasingly explore novel ways of spreading terror [1] . smallpox was declared eradicated in 1980, following a global eradication campaign led by the world health organization (who). the only officially remaining smallpox virus stocks are maintained at the centers for disease control and prevention (cdc) in atlanta and at vector, the russian viral research institute at koltsovo in siberia. these stocks were supposed to be destroyed in 1999, making the smallpox virus officially extinct. but the us president at the time, bill clinton, persuaded who members to postpone destroying them until 2002, so more research could be done on new vaccines and drugs and on smallpox genetics. the reason for the delay was a growing fear of smallpox as a bioweapon in a world no longer vaccinated against the disease, for routine smallpox vaccinations had been stopped after eradication had been achieved. therefore, who members agreed to a smallpox research plan [35] . on october 4, 2001 , cdc and state and local public health authorities reported a case of inhalational anthrax in florida. additional cases of anthrax were subsequently reported from florida and new york city [5, 8] . this was the first known ''successful'' attack after the ineffective aerosolization of a bacillus anthracis suspension in july 1993 by the aum shinrikyo sect in tokyo (also responsible for the sarin nerve gas attack on the tokyo subway system in 1995) [32] . epidemiological findings indicated that these cases of inhalational anthrax most likely occurred through aerosols generated when opening or processing letters containing b. anthracis powder. there is some evidence that these attacks may have been carried out by a us scientist, although their exact background and circumstances remain officially unresolved until now. they resulted in 22 people developing anthrax, 11 of whom suffered the pulmonary form and of whom 5 died [17] . following the anthrax attacks in the us, the us department of defense decided that smallpox stocks should not be destroyed before two anti-smallpox drugs and a new, safer vaccine were licensed, along with new diagnostic methods. not surprisingly, russia consequently also refused to destroy its stocks. following this, the government of the federal republic of germany-like those of several other states-decided to acquire and to stockpile sufficient smallpox vaccine for an emergency vaccination program for the whole population of germany (around 82 million people) in case of a smallpox attack [4] . preparations were made for such a program, such as training of medical staff to conduct the vaccinations, etc. compared to other vaccines currently in routine use, smallpox vaccination using vaccinia virus entails a considerable rate of side effects. based on previous experiences, 1 in 1,000 vaccinees will develop a serious illness requiring symptomatic medical treatment, and there will be about 30 cases of permanent damage and 1-2 deaths per million vaccinees [10, 15] . in addition, it is likely that during a mass vaccination program, some individuals, who have unrecognized contraindications, will receive the vaccination, increasing the rate of complications further. transmission of the vaccinia virus to contacts with or without contraindications is another risk. in the case of biological warfare, the aim of which would be to kill or to incapacitate the largest possible number of enemy soldiers, the aggressor will consider measures to protect his own army and civilian population. in contrast, the aim of bioterrorism is to cause maximum disruption and to seed terror. this may be achieved by actual acts or by simply causing panic. while this does not necessarily imply actually harming more than a few people, many modern terrorist groups on the other hand do not seem to pay much attention to the safety, well being or even survival of their own followers. terrorists will know that using highly infectious agents such as the smallpox virus for biological attacks might well mean their spread also to their own followers because they do not have smallpox vaccine or other preventative measures available. how likely is a bioweapon attack and, more tightly focused, how real is the chance of such an attack by a terrorist group? it has to be taken into account that sophisticated microbiological and biochemical techniques are required to cultivate highly pathogenic biological agents and to make them suitable for use as bioweapons. due to the high infectiousness of many potential bioweapon agents, all handling and manipulation would have to be done under biosafety level 3 (bsl 3) or bsl 4 conditions to protect the handlers [36] . the cdc have developed a classification system for potential biological agents [9, 29] ; table 1 lists their definitions and gives some prominent examples for each of the three categories. particularly categories b and c may be of interest to bioterrorists, due to the fact that some of these agents are not as infectious as those in category a and that dealing with them is, therefore, much easier. for example, multidrug-resistant mycobacterium tuberculosis is relatively easy to cultivate at a relatively low biohazard level, and with its incubation period between weeks and years it would be very difficult for public health systems to realize that a large number of patients were the result of a bioweapon attack. however, the insidious course might compromise its terrorising effect. on the other hand, marburg or ebola viruses are, despite their fearsome reputation, rather unsuitable, not only because those involved in their propagation are likely to succumb to them before achieving their goals, but also because of their low ability to survive outside the human body. efforts are underway to develop refined methods for assessing the suitability of biological agents as bioweapons [6] . however, even such ''impractical'' agents might still be utilized, for it might be much more efficient to paralyze public health organs and healthcare systems through threats and alarms rather than doing much direct damage. hoaxes were common in many countries table 1 classification system for potential biological agents [9] category a diseases/agents (e.g., variola virus, bacillus anthracis) high-priority agents include organisms that pose a risk to national security because they: -can be easily disseminated or transmitted from person to person -result in high mortality rates and have the potential for major public health impact -might cause public panic and social disruption -require special action for public health preparedness category b diseases/agents (e.g., brucella species, ricin toxin) second highest priority agents include those that: -are moderately easy to disseminate -result in moderate morbidity rates and low mortality rates -require specific enhancements of diagnostic capacity and enhanced disease surveillance category c diseases/agents (e.g. drug resistant m. tuberculosis, nipah virus) third highest priority agents include emerging pathogens that could be engineered for mass dissemination in the future because of: -availability -ease of production and dissemination -potential for high morbidity and mortality rates and major health impact in the aftermath of the 2001 anthrax attacks, with numerous letters allegedly containing b. anthracis spores causing considerable disruption and costs. cleverly planned arrangements pretending biological attacks can indeed wreak havoc and can cause great expense, particularly because of a lack of systems for the rapid and accurate detection of real bioterrorism incidents and their reliable distinction from hoaxes [3] . natural threats: lessons to be learned west nile virus (wnv) emerged in the new world for the first time in late summer 1999 when an outbreak of human encephalitis occurred in new york, concurrent with extensive mortality in free-living crows as well as deaths of several exotic birds at a zoological park in the same area [7, 24] . the strain of wnv found in new york in 1999 was indistinguishable from one isolated 1 year earlier in israel. how this virus covered such a vast distance is unknown; possibilities include importation of infected birds, infected mosquitoes, or viremic human beings [18, 20] . although there is no evidence that the virus was introduced deliberately, it nonetheless would represent an extremely effective biological weapon. despite intensive control efforts, wnv has now firmly established itself in the americas, spreading across the continent and reaching the west coast, and so far has caused almost ten thousand of cases of clinical illness and hundreds of deaths in humans in the united states alone [11] . in addition, surveillance and control measures such as the implementation of blood-donor testing for wnv by polymerase chain reaction have caused enormous expenditure. severe acute respiratory syndrome it can be difficult to distinguish between natural epidemics and possible bioweapon assaults. in the first half year of 2003, a novel infectious disease termed severe acute respiratory syndrome (sars) impacted significantly on the medical and scientific world. with approximately 800 deaths in total, the sars outbreak did not reach the number of annual influenza victims by far, but it nevertheless became the nightmare of public health systems all over the world. in addition, sars had grave social and economic consequences for several countries in east and southeast asia as well as in the western world. in china, unemployment especially among the migrant workforce nearly doubled to over eight million; air traffic to hong kong fell by 80% in may 2003; tourism to southeast asia virtually came to a standstill; and in canada, the loss of tourism and airport revenues amounted to $950 million, $570 million in toronto alone [22] . through an unprecedented level of international cooperation led by who, the agent responsible for this first epidemic of the new century was soon identified as a previously unknown coronavirus and the outbreak was brought under control within a few months [2] . although the source of the new virus is still not known for sure, it is likely to be linked to the so-called ''wet markets'' of southern china and not to have been caused by bioterrorist activity. instead, sars was another example of how infections may be triggered by agents originating in animals when they adapt to new hosts in whom they may cause dangerous diseases [21] . another virus so far limited to the old world recently entered the americas: in summer 2003, 72 humans were infected with monkeypox in the united states [27] . these cases were quickly linked to contact with pet animals, mostly prairie dogs, obtained from pet shops. while being held prior to sale to the public, these animals appear to have been infected through contact with gambian giant rats and other animals originating in ghana that were carrying the monkeypox virus [16] . thus, the obviously badly conducted and poorly regulated international trade in wild animals was responsible for the introduction of a potentially very harmful pathogen from west africa into the united states-at a time when mass vaccinations against smallpox were undertaken in military personnel and were (albeit with little success due to wide-spread refusal) envisaged for large numbers of civilians! again, there is no evidence that anything but human ignorance and greed were at play, but this event again shows the potential for a bioweapon attack [12] . to date, most discussions regarding the creation of a national biodefense strategy have focused largely on addressing existing threats posed by naturally occurring pathogens and toxins. with the advent of recombinant dna technology, however, researchers have techniques at their disposal for altering an organism's genetic makeup and thus biological properties. this might allow enhancing the usability of ''traditional'' biological warfare agents [31] . therefore, genetically modified bioweapon agents have been classified as a separate category (advanced biological warfare agents, abw) [25] . in addition to increasing a pathogen's virulence [23] , its ability to survive under different environmental conditions such as high temperature, ultraviolet radiation and desiccation could potentially be improved (table 2) ; by enhancing dissemination, this might make some hitherto rather unsuitable candidate agents such as filoviruses more likely to be utilized. not only will advances in biotechnology facilitate novel agents engineered to attack specific human biological systems at the molecular level, but they will permit modification of existing agricultural pathogens and the development of new anti-agricultural and even anti-material agents. likewise, technology targeted towards development of transgenic plants and insects that produce a desired protein could also be diverted toward nefarious ends. the more sophisticated the biowarfare agent to be developed, the more it requires in terms of highly skilled scientific staff and other personnel. but will such individuals be available? a study concerning the willingness of physicians to participate in the death penalty [14] is interesting in this context. it showed that one in five american physicians would be prepared to give the deadly injection to prisoners sentenced to death, 36% would establish the death of the offender, and 41% agreed to participate in the procedure in a subordinate way. it should therefore not come as a surprise if there were enough scientists to develop deadly and poisonous weapons, not only in the former ussr. following the recent developments outlined above, the center of competence for highly contagious diseases in hesse, germany, developed a catalogue listing fac-tors to facilitate the detection of a biological attack. these include: unusual, unexpected clusters of cases (large numbers of patients with similar symptoms, a large number of unclear illnesses, unexplained increase in incidence of an endemic disease), an unusual distribution (appearance of the same agent from different geographical and temporal sources, clusters occurring in geographically distinct areas, etc.), unusual modes of spread (absence of typical vectors or reservoirs, unusual spread of an agent through water, air, food, or a vector), atypical clinical courses (unusually high morbidity and mortality for a given illness, failure of normally adequate therapy, etc.), unknown or atypical infectious agents (genetically modified, atypical, or currently non-endemic strain, etc.), indirect evidence for increases in disease incidence (e.g., increase in requests for certain laboratory tests or in prescriptions for certain antibiotics, etc.), and non-medical criteria (threats, intelligence information, etc.) [34] . there are indeed possibilities to defend society against attacks with suspected biological agents [37] . besides assuring an heightened awareness among health care workers to increase the likelihood of early detection, letters can be humidified, for instance, to decrease the risk of aerosolization of the content, or they can be irradiated as done in some us post offices. in case of a real or doubtful contamination, there are pharmaceuticals, vaccines and therapeutic regimens available to treat exposed individuals [33] . it should not be forgotten either that efficient public health systems are able to fight epidemics successfully even in the absence of specific preventative measures such as vaccines, just by applying ''good basic public health measures'' [13] , as was recently shown with sars. attacks with biological weapons are indeed a real threat and the responsible government agencies need to be aware of this. however, the actual likelihood of such attacks currently seems to be low compared to that of naturally emerging agents. an effective implementation of a national biosecurity strategy will require a variety of independent efforts across federal and public health organizations as well as bioscience research. there has to be a discussion whether the current cost-intensive smallpox vaccination programs are efficient. while the above-mentioned unwanted side effects were acceptable when there was still the danger of a smallpox epidemic, this is no longer true after its eradication. neither does the immunity induced against the closely related cowpox or monkeypox viruses justify widespread vaccination with vaccinia virus, so that it should only be considered in case of an acute smallpox threat. it is certainly beneficial to allocate grants for investigations into the ways in which microbiological agents change hosts and into virulence factors. on the other hand we have to ensure that scientists comply with regulations and accept the ethical constraints of their activities, as well as apply the necessary safety precautions in their work [19] . increased funding of bioweapon research programs will lead to more individuals with training and skills in this area and might thus have paradox effects [28] . therefore, new initiatives to deal with broader threats that may result from misuse of technology need to be pursued, in parallel with existing and planned programs. ultimately, whenever novel infectious agents appear suddenly, it is probably inevitable that there will be speculations and rumors: that they were designed in biowarfare laboratories or emerged from biomedical research (as was argued for hiv), that they came from space (as suggested for sars), or that they represent a bioterrorist attack. however, reality is more banal but nevertheless painful: nature itself is the best bioreactor for apocalyptic biological agents and, through evolution, has at the same time developed the best defense strategies against them [30] . biological warfare and bioterrorism severe acute respiratory syndrome (sars)-paradigm of an emerging viral infection evaluating detection and diagnostic decision support systems for bioterrorism response index case of fatal inhalational anthrax due to bioterrorism in the united states the weapon potential of a microbe outbreak of west nile-like viral encephalitis update: investigation of anthrax associated with intentional exposure and interim public health guidelines bioterrorism agents/diseases; last modification smallpox vaccination and adverse reactions: guidance for clinicians west nile virus activity in the united states (reported as of human monkeypox: an emerging zoonosis epidemiology: modeling the sars epidemic physicians' willingness to participate in the process of lethal injection for capital punishment developments in vaccination and control between 1900 and 1966 zaki sr, the veterinary monkeypox virus working group (2004) monkeypox transmission and pathogenesis in prairie dogs investigation of bioterrorism-related anthrax, united states, 2001: epidemiologic findings genetic analysis of west nile new york 1999 encephalitis virus mixing bugs and bombs origin of the west nile virus responsible for an outbreak of encephalitis in the northeastern united states viral zoonoses-a threat under control? sars down but still a threat. national intelligence council intelligence community assessment creation of killer poxvirus could have been predicted the outbreak of west nile virus infection in the new york city area in 1999 opened for signature: 17 the detection of monkeypox in humans in the western hemisphere buying biosafety-is the price right? public health assessment of potential biological terrorism agents the best defence against bioweapons has already been invented by evolution preparedness and response to bioterrorism bacillus anthracis incident emea/cpmp guidance document on use of medicinal products for treatment and prophylaxis of biological agents that might be used as weapons of bioterrorism. cpmp/ 4048/01 management of an attack with biological agents. reflections on the necessary infrastructure for states and towns smallpox eradication: destruction of variola virus stocks public health response to biological and chemical weapons: who guidance, 2nd edn. world health organization key: cord-209269-7ojtwe78 authors: parisi, daniel r.; patterson, germ'an a.; pagni, lucio; osimani, agustina; bacigalupo, tomas; godfrid, juan; bergagna, federico m.; brizi, manuel rodriguez; momesso, pedro; gomez, fermin l.; lozano, jimena; baader, juan martin; ribas, ignacio; meyer, facundo p. astiz; luca, miguel di; barrera, nicol'as e.; 'alvarez, ezequiel m. keimel; oyhanarte, maite m. herran; pingarilho, pedro r.; zuberbuhler, ximena; gorostiaga, felipe title: social distance characterization by means of pedestrian simulation date: 2020-09-08 journal: nan doi: nan sha: doc_id: 209269 cord_uid: 7ojtwe78 in the present work, we study how the number of simulated clients (occupancy) affects the social distance in an ideal supermarket. for this, we account for realistic typical dimensions and process time (picking products and checkout). from the simulated trajectories, we measure events of social distance less than 2 m and its duration. between other observables, we define a social distance coefficient that informs how many events (of a given duration) suffer each agent in the system. these kinds of outputs could be useful for building procedures and protocols in the context of a pandemic allowing to keep low health risks while setting a maximum operating capacity. recent works [2, 3] propose to combine microscopic agent simulation with general disease-transmission mechanisms. however, because of the complexity and uncertainties in the actual knowledge for quantifying these transmission processes, we will not consider in this work any particular contagion mechanism, instead, we will focus on studying the distance between people in an everyday pedestrian facility as an isolated aspect to be integrated in the future by experts considering all mechanism for any particular disease propagation. an antecedent of this kind of analysis was recently reported considering field data from a train station [4] . one of the key questions we try to answer is how to describe the realized social distance for a given occupation of an establishment. for solving this problem, it is necessary to consider the pedestrian displacements and trajectories that they do while performing certain tasks, and thus, a natural tool is to use pedestrian simulation. the time evolution of positions from simulated agents can provide not only the relative distance between agents, but also the duration of events in which the recommended social distance is not kept. many industries and shops were closed in different phases of the covid-19 pandemic. however, grocery shops had to be kept open, in particular supermarkets. to avoid crowding and to keep some physical distance between clients, the authorities lowered the allowed capacity. different countries regulation adopt social distance requirements between 1 and 2 m [4] . in the present study we will consider a distance of 2 m as the social distance threshold, because this distance is an upper bound for the saliva droplets from a human cough which cannot travel more than 2 m in space at approximately zero wind speed [5] . we propose to investigate how the value of the allowed capacity affects the social distance in an ideal supermarket of 448 m 2 . the results should not be directly extrapolated to other supermarkets or facilities, nevertheless, the methodology and indicators can be useful for applying them to other source of data, being from simulations or field data of other pedestrian systems. in order to simulate the complex environment and the agent's behavior, the proposed model involves three levels of complexity: operational, tactical, and strategic [6] . the more general level of the model is in charge of providing a master plan for the agent when it is created. in practical terms for the present system, it gives a list of n p products for agents to acquire (a shopping list). each one of the n p items is randomly chosen between a total of m p products available. also, they are identified with a unique target location (x pn ) in the supermarket. once the agent is initialized with its shopping list, the strategic level shows the first item in the list to the agent. the agent will move toward it using the lower levels of the model. when the agent reaches the position of the product, it will spend a picking time (t p ) choosing and taking the product, after which the strategic level will provide to the agent the next item on the list. after the list of products is completed, the agent must proceed to the less busy supermarket checkout line. it will adopt a queuing behavior until it gets the checkout desk and spends a time t co processing its purchase. the function of the tactical level is to provide successive visible targets to the agent that will guide it to the location of the desired product (x pn ) or checkout line. as inputs, the tactical module takes the current agent position (x i (t)) and the position of the current product (x pn ) in the list. the output is a temporal target (x v (t)) visible from the current position of the agent. the definition of visibility is that if we take a virtual segment between (x i (t)) and (x v (t)), this segment does not intersect any of the walls or obstacles (shelves). the information delivered by the tactical module is obtained by implementing a squared network connecting all the accessible areas of the simulated layout (see fig.2 ). given any two points in the walkable domain, the corresponding nearest points on the network are found and then the shorter path between these points is computed by using the a* algorithm [7] . once the path on the network is defined, the temporary target x v (t) is chosen as the farthest visible point on that path seen from the current agent position. clearly, x v (t) will change with time, as the position of the agent changes. when the product target is visible from the agent, this is set as the visible target and the network path is no longer considered until a new product should be found. as the lower level describing the short-range movements of agents we propose an extended version of the "contractile particle model" [8] . this extension will provide an efficient navigation allowing to avoid potential collisions with other agents and obstacles. the basic model is a first-order model in which particles have continuous variable radii, positions and velocities that change following certain rules. specifically, the position is updated as where v i is the desired velocity and x i (t) the position at time t. the radius of the i th particle (r i ) is dynamically adjusted between r i min and r i max . when this radius has large values, it represents a personal distance necessary for taking steps, but when it has low values, it represents a hard incompressible nucleus which limits maximum densities. when particles are not in contact, the desired velocity v i points toward the visible target with a magnitude proportional to its radius, where the direction e i t and the magnitude v are defined by the following equations where v d is the desired speed. while the radius has not reached the maximum r max , it increases in each time step following being τ a characteristic time in which the agent reaches its desired speed as if it was free, and ∆t is the simulation time step of eq. 1. when two particles enter into contact (d ij = |x i − x j | − (r i + r j ) < 0) both radius instantaneously collapse to the minimum values while an escape velocity appears moving the particles in the directions that will separate the overlap: the escape velocity has the magnitude of the free speed and, thus, it can be written as v i e = v d e ij . this velocity is only applied during one simulation step because, as the radii are simultaneously collapsing, the agents are no longer overlapping. up to here, we described the basic cpm as it appears in ref. [8] . this model satisfactorily described experimental data of specific flow rates and fundamental diagrams of pedestrian dynamics. however, particles do not anticipate any collisions, and this capacity is a fundamental requirement for simulating the ideal supermarket (displaying low and medium densities, and agents circulating in different directions). for this, we propose to extend the calculation of the agent velocity (eq. 2) by considering a simple avoidance mechanism. the general idea is that the self-propelled particle will produce any action only through changing its desired velocity v i (t) as stated in ref. [9] . in this case, the new mechanism will only change the direction of desired velocity v depending on the neighbor particles and obstacles. first, the collision vector (n c i ) is calculated where j indicates the nearest visible neighbor, k the nearest point of the nearest visible wall or obstacle, andη is a noise term for breaking possible symmetric situations. visibility field of agent "i" then the avoidance direction is obtained from and finally the velocity of the particle to be used in eq in fig.1 the vectors associated with the original and modified model can be seen in detail. because the agents must perform different tasks, more complex than just going from one point to another, it was necessary to define five behavioral states. this was achieved by setting different model parameters and movement patterns. more concisely, the five behavioral states of agents were: -going: this is the normal walking behavior when going from one arbitrary point to another with the standard velocity and model parameters. only in this state, the agent uses the modified cpm velocity (eq. 9) for avoiding potential collisions. the rest of the behavioral states only use the basic cpm (eq. 1 to 6). -approximating: when the agent is closer than 2 m of the current product, it reduces its desired speed and, because of how parameters are set, it will not be forced to reach it if there is another agent buying a product in the same target x pn . -picking: once the agent reaches the product (closer than 0.1 m) a timer starts and it will remain in the same position (eq. 1 does not update its position) until the picking time (t p ) is up. -leaving: after spending the time t p , the agent leaves the current location going to the next product on the list. while abandoning this position, it could find other waiting agents (in approximating behavioral state) and, thus, its parameters must be such that it can make its way through. once the agent is farther than 2 m from the past product, it changes to the "going" behavioral state. -queuing: finally, when the agent completes the shopping list, it proceeds to the checkout desks by choosing the one with the smaller line. it waits at a distance of 1.5 m from the previous queuing agent, and when it reaches the checkout position, it remains there for a t co time. by considering these behavioral states in the agent model, the conflicts and deadlock situations are minimized. in this way, this model improvement allows us to simulate higher densities than with the basic operational models. finally, as a closure for this general section (2) about the model, we point out that for the sake of comparison, we also implement other operational models: the "social force model" [10, 11] , and the "predictive collision avoidance model" (pca) [12] . the results for all models will be compared in selected observables, while the more deep study is performed using the modified cpm described above. the plant of 448 m 2 of the ideal supermarket to be simulated is shown in fig. 2 . the dimension of shelves (1 m x 10 m) and corridor width (2 m) are taken from typical real systems. also, the different process time and other data considered were provided by a supermarket chain from argentina. we define n as the allowed capacity or the occupation of the supermarket as the total number of agents buying simultaneously inside the system. this is the more important input to be varied in our study and it ranges from n = 2 to n = 92. the agent generator produces an inflow of 1 agent every 5 seconds until it reaches the n value for the simulation. from that moment on, the agent generator monitors the occupation, and it generates a new agent every time an existing agent completes its tasks and is removed from the simulation. by doing this, the value of n is maintained constant over the whole simulation. every agent created by the generator is equipped with a shopping list of exactly n p = 15 items, which are chosen randomly from a total of 228 items available (shown in fig.2 b) . the corresponding product locations (x pn ) are separated by one meter between adjacent locations. agents visiting the products on its lists, spend a picking time with a uniform distribution (t p ∈ [60s, 90s]). after completing the lists, agents choose the shortest queue to one of the eight checkout points shown in fig. (2 b) . the ideal supermarket has a maximum of four queues, and each of them leads to two checkout desks. the distance between the agents that form the queue is 1.5 m. the first positions on these queues are at a distance of 3 m (at y = 4 m, in fig.2 ) from the checkout points. once an agent reaches the cashier (at y = 1 m, in fig.2) it spends a checkout time t co uniformly distribute between t co ∈ [120s, 240s]. for each value of n , we simulated two hours (7200 s) and recorded the state of the system every ∆t2 = 0.5 s, thus producing 14400 data files with positions, velocity, and behavioral state of agents. the rest of the model parameters depends on the behavioral state of the agent. for the case of "going", the parameters of the avoidance mechanism described in eq. (7) we first show general results of the simulated supermarket by displaying typical trajectories (fig. 3) and density fields (fig. 4) . figure 3 plots ten random chosen trajectories in the second hour of simulations for the selected n values. qualitatively it can be seen more intricate trajectory patterns as the occupancy increases. however, in all cases, it can be observed that the available area is uniformly visited by simulated agents while visiting their product list. complementary information is shown in fig.4 , where the density is averaged over all the simulation time (2 hours). as expected, greater occupancy presents greater mean density values. besides, these density fields present higher values at the spots that agents stay longer, revealing picking points of products and predefined places at queuing. also, as a macroscopic observable of the system, we study the number of agents that could be processed (i. e. fulfill the shopping list and exit the supermarket between the two hours simulated) and the mean residence time for those agents. the results are presented in fig. 5 . as can be observed, both quantities increase monotonically with the allowed occupancy for the studied range of values and the supermarket set up, considering eighth checkout desks. also, it can be seen that different operational models display similar observables. the sfm [10, 11] and pca [12, 10] models are force-based models which present more limitations in terms of the maximum density they can simulate before forces get balanced (generating dead-locks) for the complex scenarios and behavior considered. this is the reason why the maximum occupancy studied with these models is lower than that simulated with the cpm described in sec.2. in this subsection, we characterize distance between agents during the simulations with the modified contractile particle model (cpm) for different allowed capacities. an interesting outcome is the distance to the first neighbor for each agent shown in fig. 6 . the pdf of first-neighbor distances (d f n ) shows that for lower occupations of the simulated supermarket, the probability of having the first neighbors further away than d f n ∼ 5 m is greater. on the other hand, higher occupancy values display higher probabilities of having a distance shorter than 5 m. in particular, all distributions show a maximum probable value around d f n ∼ 4 m. moreover, the height of these probability peaks decreases for lower occupancy values. now we take the social distance threshold of 2 m, as discussed in sec.1, and calculate related probabilities of agents below this critical social distance. the first observable we calculate is the probability of having the first neighbor closer than 2 m (p f n<2m ). in other words, this is the probability of having at least one neighboring agent within 2 m. it is determined by averaging over recorded data every ∆t2 = 0.5 s, from minute 20 to 120 as shown in eq.(10) where n ti = 12000 = 14400 − 2400 is the data at recorded times after 20 minutes, n is the occupancy and n f n2m is the number of particles having a first neighbor at less than 2 m. note that if two particles i and j are the only particles at less than 2 m, n f n2m = 2. also, if j is the first neighbor of i, not necessarily, i will be the first neighbor of j. the above probability (p f n<2m ) only considers if the first neighbor is closer than 2 m, but it does not take into account if there are many occurrences of neighbors at less than 2 m. for this reason, we now take into account the number of pairs having distance less than 2 m, and the corresponding probability (p pair<2m ) where n p2m is the number of pairs of particles at distance closer than 2 m and [n (n − 1)]/2 is the total number of possible pairs, having n particles in the system. in this case, if only particles i and j are closer than 2 m, n p2m = 1 because one pair is counted. in fig. 7 both probabilities ( p f n<2m and p pair<2m ) are displayed for the modified contractile particle model (cpm) and also by comparison with the social force model (sfm) and predictive collision avoidance model (pca). it can be seen that the probability of having the nearest neighbor at less than 2 m increases monotonically with the allowed capacity. however, the pair probability quickly increases for low occupancy, and after n ∼ 15, it remains nearly constant indicating that the number of pairs n p2m scaled with n as the number of total possible pairs (∼ n 2 ). furthermore, fig. 7 indicates that different operational models display similar macroscopic behavior regarding social distances, at least for values bellow or above 2 m. in the above analysis, the occurrence of certain distances between simulated agents was studied, but the duration of these events was not considered explicitly. this will be done in the following subsection. here we study the time that lasts the events when pairs of agents are found at less than 2 m (see sec.1). the main sources of these events are when agents are picking products at neighbor product locations or when queuing for checking out from the supermarket. if two particles i and j encounter at a given time, then they separate for more than 2 m, and the same particles re-encounter at a future time, it is considered as two separate events. considering that: (a) the parameter we choose to maintain constant during each simulation is the allowed capacity n , and this capacity is reached at the beginning of each simulation in a very short time comparing to other processes, and (b) all agents have the same number of items in their list, and thus the required time to complete it is similar in average. the first group of n agents will go to the checkout points at nearly the same time, producing the greater demand for checkout, which generates the longest queues. after that, the new agents will enter slowly as other agents exit the simulation, and thus the described behavior will relax. this dynamic lead to more queuing agents during the first hour of simulation and less during the second hour. because of it, we analyze separately the duration of encounters occurred during the first and the second simulation hour in fig. 8 . the different time scales and the number of cases in its both panels confirm that the first hour is dominated by particular long lines waiting for checkout, while in the second hour (fig. 8 b) the duration of social distance events less than 2 m are dominated by the shorter process, i.e.: the picking time at products. events in the queuing line are much long-lasting because of two reasons. first, the particular process at the checkout desk takes between 2 to 4 minutes (instead of 1 to 1.5 minutes in the picking process). second, a line with n l agents will produce that the last agents will spend about n l times t co , which for a few agents, namely n l = 5, it could represent 20 minutes of waiting time at a distance of 1.5 m from another agent. this problem of high exposure time between pairs of agents at queuing lines could be avoided if a slower ramp of inflow of agents was adopted in the start of the process, let's say something above the maximum average outflow of the system (eight agents in three minutes, i.e. ∼ 1 agent every 23 s). we did not adopt this in the simulations because it would take too long for simulations to reach the desired occupation n . but it is clear that the problem noted above at the beginning could be solved in a real operation by allowing a low flow rate of agents at the opening (of about twice the capacity of the checkout). also, this problem would be a transient behavior only at opening, being the most part of the day operation as described in our second simulation hour. furthermore, fig.8 shows that, as expected, fewer social distance events occur when the time thresholds increase. and in all cases, the number of events seems to grow quadratically with n . now, looking for a criterion that determines which would be a reasonable allowed capacity in the ideal supermarket, we define the social distance coefficient (sd c (t e )), for the threshold distance of 2 m, as where t e is the minimum duration of a particular social distance event (r ij ≤ 2 m), n e (t e ) is the number of these events which last at least t e , and n p is the total number of agents processed by the system in the same period of time in which n e is computed. factor 2 is needed for taking into account the number of agents in the numerator since 2 agents (i and j) participate in each event. this coefficient allows us to compare how many agents have participated in social distance events of duration greater than t e concerning the number of people who have passed by the system. thus, a value of sd c (t e > 2min.) = 1 indicates that, on average, each agent has participated in one event of social distance less than 2 m that lasts at least 2 minutes. if sd c (t e > 2min.) < 1, it would indicate that only a fraction of the agents have participated in such events. having established in subsec. 4.3 that the event duration of the first simulation hour is dominated by checkout lines, we now concentrate on looking at the second hour of simulation when the impact of these lines is very low and stationary. this situation is representative of the daily operation of the supermarket, and it is shown in fig.9 displaying the social distance coefficient as a function of the occupation for different event duration limits t e . first, we note in fig. 9 a that the curve corresponding to t e > 1 min. grows steeply with n. this could be related to the fact that the picking time range between 1 minute and 1.5 minutes, and that the products are spaced by 1 m, so if two agents must go simultaneously to the same product or the first or second nearest product, they could generate a 2 m social event lasting at most 1.5 minutes, in particular many events lasting more than 1 minute would occur. also, the social distance coefficient seems to follow a linear relation with n for this particular time limit t e . a change of regime can be observed for t e > 1.5 min. in which curves are more similar between themselves for the different t e presented, and they follow a quadratic relation with n . because the maximum picking time is 1.5 minutes, this is the maximum possible overlapping time for two agents in neighbor (or the same) products. greater time events will arise when more than two agents are waiting for the same or near products, as is in the cases of products near any of the short lines for checking out. finally, we could use fig. 9 b, as a guide for deciding on the allowed occupancy. if based on epidemiological knowledge or criteria, it was decided that it would be acceptable to allow that all agents participate once in a 2-m social event lasting at most 1 minute, then the allowed occupation would be very small, n ∼ 10. alternatively, if events up to 1.5 minutes were accepted, then the allowed occupation would be n = 40. in the case of t e = 2 min., the capacity could rise to n = 70. also, it could be established that even for n = 90 the events of the social distance of 2 m, lasting more than 3 minutes, would only affect the 40% of the processed agents. of course, the same fig. 9 b could be used for finding another allowed occupancy if the criterion would consider that, for example, only 25% of the agents can participate in the analyzed events. in this work, we investigate and characterize the social distance realization in an everyday pedestrian system by simulating the dynamics of an ideal supermarket. many sources of complexity were successfully taken into account with a multi-level model, which allows us to simulate not only translation but also more complex behaviors like waiting times when picking for particular products and queuing for checkout points. the main process that keeps pedestrians near one to another is the queuing lines for checkout. therefore, advice for the operation would be to keep these lines as small as possible either by increasing the number of checkout points or by decreasing the occupancy. different operational models, display similar macroscopic observables regarding social distances at values greater than 2 m indicating that the results are robust with respect to microscopic collision avoidance resolution and also suggesting that the simulated paths of the particles are more influenced by the geometry, shopping list, and time-consuming process, than by the particular avoidance mechanism. however, first-order models as the cpm presented in [8] and sec.2 seem more suitable for simulation of high populated scenarios with complex behavioral agents. taking a social distance threshold of 2 m [5] , probabilities and duration of such events were studied. the social distance coefficient was defined as an indicator of the fraction of the population passing by the system that is involved in one or many of these events lasting at least a certain time threshold t e . the same analysis can be done for a different set of parameters and, of course, for other pedestrian facilities, being other specific supermarkets or any different systems (transport, entertainment, etc.). of course, existing facilities can be monitored with measurement methods [4] providing high-quality trajectory data. this kind of data could also be interpreted in terms of the analysis performed in the present work. by setting a distance and a time threshold, and what fraction of the population would be tolerable to be exposed to these conditions, the maximum occupancy can be established in any particular system by performing simulations and using the observables defined in this work. therefore, this analysis could be used for determining which is the ideal occupation of a pedestrian facility, keeping low contagion risk, and maximizing the number of users per unit of time. world health organization. coronavirus disease (covid-19) advice for the public a microscopic approach to study the onset of a highly infectious disease spreading agent-based simulation of pedestrian dynamics for exposure time estimation in epidemic risk assessment monitoring physical distancing for crowd management: real-time trajectory and group analysis on coughing and airborne droplet transmission to humans pedestrian route-choice and activity scheduling theory and models path planning for virtual human motion using improved a* star algorithm continuous-space automaton model for pedestrian dynamics pedestrian collision avoidance with a local dynamic goal simulating dynamical features of escape panic specification of the social force pedestrian model by evolutionary adjustment to video tracking data a predictive collision avoidance model for pedestrian simulation the authors acknowledge the information and data provided by the argentinean supermarket chain "la anónima". this work was founded by project pid2015-003 (anpcyt) argentina. key: cord-300731-i2ow33bk authors: cowan, fred m.; broomfield, clarence a.; stojiljkovic, milos p.; smith, william j. title: a review of multi-threat medical countermeasures against chemical warfare and terrorism date: 2004-11-17 journal: mil med doi: 10.7205/milmed.169.11.850 sha: doc_id: 300731 cord_uid: i2ow33bk the multi-threat medical countermeasure (mtmc) hypothesis has been proposed with the aim of developing a single countermeasure drug with efficacy against different pathologies caused by multiple classes of chemical warfare agents. although sites and mechanisms of action and the pathologies caused by different chemical insults vary, common biochemical signaling pathways, molecular mediators, and cellular processes provide targets for mtmc drugs. this article will review the mtmc hypothesis for blister and nerve agents and will expand the scope of the concept to include other chemicals as well as briefly consider biological agents. the article will also consider how common biochemical signaling pathways, molecular mediators, and cellular processes that contribute to clinical pathologies and syndromes may relate to the toxicity of threat agents. discovery of mtmc provides the opportunity for the integration of diverse researchers and clinicians, and for the exploitation of cutting-edge technologies and drug discovery. the broad-spectrum nature of mtmc can augment military and civil defense to combat chemical warfare and chemical terrorism. t he list of military chemical threat agents has lengthened over the last century. world war i saw the development and deployment of edemagenic and choking agents such as chlorine, phosgene, and cyanide, and blister gases such as sulfur mustard (sm). [1] [2] [3] world war ii cultivated germany's development of nerve gases such as tabun, sarin, and soman, [3] [4] [5] [6] and this class of agent was used in the 1980 to 1988 iraq-iran conflict 7, 8 and in the 1994 to 1995 terrorist attacks in japan. [9] [10] [11] the cold war yielded more lethality refinements in nerve agents to include development of vx. 12 consistent with past experience, new threat agents with novel mechanisms of toxicity may evolve. countermeasures for defending deployed military personnel against weaponized chemical warfare agents do not seamlessly transition to military force protection and civil defense against chemical terrorism. terrorist threat agents include all classes of chemical warfare agents, toxins, and a broader menu of toxic industrial chemicals. the large number of threats and potential for use of multiple toxic substances in combination can complicate agent identification and medical response. a terrorist assault might more closely resemble the chemical accident at bhopal, india, than a chemical warfare attack. symptoms may be closer to smoke inhalation and acute respiratory distress syndrome (ards) than to the cholinergic crisis caused by nerve agent poisoning. research into medical countermeasures against weaponized battlefield chemical warfare agents has focused on distinct classes of nerve, blister, choking, or edemagenic agents and pathologies specific to these agents, e.g., nerve agent-induced seizure, phosgene-induced ards, and sm blistering. the present challenge of defending against the myriad of threat chemicals and toxins necessitates a more collective and comprehensive strategy that more generally addresses the symptoms and pathology of chemical toxicity rather than a specific chemical mechanism. the multi-threat medical countermeasure (mtmc) hypothesis for chemical toxicity is analogous to the broad-spectrum approach used for protecting against infectious agents authored by dr. ken alibek, former deputy chief of biopreparat, the civilian arm of the former soviet union's biological weapons program. 13, 14 it may be possible to exploit the fact that the diverse chemistry, sites, and mechanisms of action of toxic chemicals often involve common underlying biochemical signal pathways. these common pathways are key in initiating the cellular processes that contribute to clinical pathologies caused by chemical exposures. clinically defined syndromes or symptoms that represent pathologies of respiratory, neuronal, cardiac, cutaneous, or other systems to include anaphylactic shock, bronchospasm, ards, seizure, excitotoxin-like neuronal degeneration, and blistering are observed after exposure to chemical warfare agents, hazardous chemicals, toxins, and even infectious agents. the mtmc hypothesis suggests that common biochemical pathways contribute to the toxicity that causes the pathologies associated with different classes of toxic chemicals. mtmc drugs that target these key pathways might create a physiological "quiescent" state that protects against various chemical insults. 14 the fact that inflammation causes or contributes to numerous pathologies and that many toxic chemicals provoke an inflammatory response is well established (for review, see refs. [14] [15] [16] . although evidence for a significant role for inflammation in chemical toxicity continues to accumulate, inflammatory response is still often regarded as merely collateral to chemical insult and secondary to toxicity. [14] [15] [16] the evolution over the last decade of the concept that inflammatory response is a key element of chemical toxicity that allows for the development of anti-inflammatory drugs as multi-agent countermeasures is unique to the mtmc hypothesis. [14] [15] [16] the mtmc hypothesis encompasses the interaction of multiple toxic agents and diverse countermeasure drugs on mediators and signaling pathways that cause or sustain inflammation. whether inflamma-tion is a collateral biomarker or causative contributor to chemical toxicity is controversial. however, in either case inflammation can provide a common denominator to facilitate deciphering the relationships between agents, drugs, signaling pathways, and cellular processes pertinent to predicting candidate mtmc drugs. 14 a major tenet of the pharmacology of anti-inflammatory drugs is that drug action can result from the specific inhibition of molecular mediators of inflammation or from modulation of the underlying biochemistry that initiates or sustains an inflammatory response. different inflammatory mediators, pathways, and cell populations can be induced or recruited by different conditions or insults. this multifaceted nature of inflammation may require the use of numerous pharmacologically distinct antiinflammatory drugs to contend with the different manifestations of inflammatory pathology. however, enough uniformity of biochemical pathways of inflammation exists to allow for development of broad-spectrum anti-inflammatory drugs. the biochemical pathways associated with chemical toxicity can involve proteases, inflammatory cytokines such as tumor necrosis factor (tnf), interleukin (il)-1, il-8, and other molecules such as platelet activating factor (paf), n-methyl-d-aspartate (nmda) glutamate receptors, acetylcholine (ach), substance p, and poly(adp-ribose) polymerase (parp) (for review, see ref. 14). these mediators and receptors can influence inflammatory responses associated with cellular processes such as degranulation, apoptosis, and necrosis that contribute to pathologies caused by chemical agents. therefore, many classes of compounds used as countermeasures to chemical warfare agents such as parp inhibitors, proteases inhibitors, adenosine agonists, and nmda receptor antagonist, although not chiefly thought of as anti-inflammatory drugs, have anti-inflammatory pharmacology (table i ) (for review, see ref. 14) . the diverse biochemical pathways that influence inflammatory responses may explain why many pathologies associated with intoxication by chemical agents have an inflammatory component. along this line of reasoning, it is not surprising that the majority of countermeasure drugs that have shown efficacy against edemagenic, vesicant, and nerve agents have anti-inflammatory pharmacology (for review, see ref. 14) . the presence of inflammatory responses and anti-inflammatory actions of countermeasures implicates inflammation as a cause or a resulting biomarker of the toxicity of these agents. the extent to which common inflammatory mechanisms are coupled to chemical agent toxicity will determine the feasibility of developing anti-inflammatory mtmc. inflammatory response induced by the prototypic edemagenic agent phosgene is the primary cause of the frequently lethal ards. 17, 18 ards is also associated with pulmonary toxicity of the acetyl-cholinesterase-inhibiting organophosphorus insecticide thionazine, 19 sm, 14 and other toxic chemicals, as well as toxins, infectious-agent pneumonia, shock, sepsis, burns, and other trauma (table ii) . sm dermal-epidermal separation is similar to that caused by proteolysis and certain bullous diseases, and this has fostered the hypotheses that sm vesication involves proteolytic and/or inflammatory responses. [15] [16] [17] [18] [19] [20] experimental evidence has accumulated over the last decade demonstrating sm-increased proteases and the expression of inflammatory enzymes, gene products, mediators, and receptors in tissue or cell cultures. 14, 15, [19] [20] [21] [22] [23] inflammatory enzymes and cytokines have also been reported in the skin of sm-exposed animals. 24, 25 cholinergic crisis is the paramount toxicological event in nerve agent intoxication, causing acute toxicity and precipitating seizure and neuronal degeneration. 26 however, along with cholinergic crisis anaphylactoid reactions, pathological proteolytic activity and inflammatory cytokines have also been reported in nerve agent-intoxicated animals. 14,27-31 the initial observations of anaphylactoid reactions associated with soman poisoning reported by doebler et al. 28 and newball et al. 29 nearly a quarter century ago have recently again been considered by gilat et al. 30 in sarin-intoxicated animals. anaphylactoid reactions, by definition, occur upon initial exposure to a chemical independent of antibody production. anaphylactoid reactions and nerve agent intoxication can generate acute symptoms such as convulsions, bronchoconstriction, respiratory failure, circulatory collapse, and death within a few minutes. 14,16 thus, anaphylactoid reactions, 14, 16 like the more extensively investigated cholinergic crisis and excitotoxin-like neuronal degeneration, are a clinically well-defined and potentially lethal syndrome associated with nerve agent intoxication. mcleod 32 noted neuronal and neurophil degeneration and necrosis associated with soman-induced brain lesions. the pattern of injury was similar to that caused by epilepsy and ischacute lung/organ dysfunction emic brain injury. some animals also had cardiac lesions characterized by acute necrosis with subsequent mild inflammation and fibrosis. inflammatory cytokines such as tnf, il-1, and il-8 are associated with recruitment of inflammatory cells and cellular processes such as degranulation, apoptosis, and necrosis. inflammatory cytokines are causative factors in phosgeneassociated ards, they contribute to sm injury, and they are implicated in nerve agent seizure and neuronal degeneration. 14, 17, 18, 24, 25, 31, 32 skin from mice sm exposure sites and the brains of soman-intoxicated rats demonstrate increased il-1␤ mrna. 22,23,31 il-1 is an inflammatory cytokine that can precipitate other cytokines such as tnf, il-6, and il-8, and such a cytokine cascade is implicated in sm toxicity. 22 svensson et al. 31 have demonstrated that soman induces the inflammatory cytokine il-1␤ in rat brain, and this activity is highly correlated with convulsions. williams et al. 33 confirmed il-1␤, and further observed an acute and transient upregulation of other inflammatory gene response, i.e., tnf, il-6, e-selectin, and intercellular adhesion molecule-1 caused by exposure to soman. the authors suggest that these molecules may be involved in soman-induced brain injury. eriksson et al. 34 demonstrated that nmda antagonists could suppress excitotoxin-induced il-1␤ mrna in the rat brain. furthermore, vezzani et al. 35 have protected mice against nmda receptor-dependent seizures with soluble il-1 receptors that specifically antagonize il-1 activity. the authors suggest that the interaction between il-1 and il-1 receptors represents a crucial mechanism of seizure and a target for development of anticonvulsant drugs. the role of inflammatory cytokines in edemagenic, nerve, and blister agent toxicity supports cytokine antagonists as candidate mtmc. anti-inflammatory drugs such as ibuprofen are the drugs of choice in reducing phosgene toxicity. 17, 18 the majority of drugs that have shown efficacy against sm in animal models have anti-inflammatory pharmacology (for review, see ref. 14). our institute has screened more than 400 compounds in the mouse ear vesicant model (mevm) for cutaneous injury; 19 compounds reduced sm histopathology greater than 50%. 36 -38 the 19 compounds include seven listed as antiinflammatory drugs, consisting of five capsaicin analogs, a single cyclooxygenase inhibitor, indomethacin, and a calmodulin antagonist, fluphenazine. 36 -38 the three protease inhibitors and three parp inhibitors included in the list of antivesicant drugs also have the potential to inhibit inflammatory responses. 14, 39 the remaining six anti-vesicant drugs, sodium 3-sulfonatopropyl glutathionyl disulfide, 3% hydrogen peroxide gel, dimercaprol, and three other mercaptopyridines analogs are listed as sm scavengers. 36, 38 however, mercaptopyridine-like compounds have demonstrated anti-inflammatory pharmacology to include inhibition of the cellular release of the inflammatory cytokine il-1. 40 therefore, 17 of the 19 compounds with efficacy against sm toxicity have anti-inflammatory pharmacology, the exceptions being two of the listed sm scavengers. furthermore, antivesicant drugs with anti-inflammatory pharmacology generally demonstrate inhibition of hd-increased il-8 in human epidermal keratinocyte (hek) cultures (ref. 41; fm cowan, ca broomfield, and wj smith, unpublished observations). collectively, the mevm and in vitro sm-increased il-8 hek assays effectively screen, not just for vesicant countermeasures, but also for anti-inflammatory drugs that inhibit il-8. these models further identified drugs such as capsaicin analogs as better antivesicants and anticytokine drugs than major classes of currently available corticosteroids and nonsteroidal anti-inflammatory drugs. the anti-inflammatory action of nerve agent countermeasures to include atropine, carbamates, and benzodiazepines has been extensively reviewed. 14, 16, 42 for example, a role for synergy between mediators of anaphylactic response and cholinergic status "hypersensitivity" is accepted for asthma-related lethal bronchoconstriction, and has been implicated for nerve agentinduced lethal bronchoconstriction. 14, 16, 43, 44 kotev 44 has demonstrated that soman-induced bronchial spasm was greatly intensified not only by ach, but also by histamine. hypersensitivity to inhalation of the cholinergic compounds such as methylcholine or the autacoid histamine is a diagnostic test for asthma, and anticholinergics are sometimes used as supplementary treatment. 45 inflammatory, neurological, and tissue pathology pathways involve proteolytic processes. 14 synthesized mediators such as paf and serine proteases such as tryptase and tissue plasminogen activator (t-pa) in addition to autacoids such as histamine are significant factors in inflammatory and anaphylactic responses. 14, 16, 46 carbamates can inhibit serine proteases that mediate anaphylaxis and prevent anaphylactic response. 47 the benzodiazepine diazepam is food and drug administration approved for the treatment of seizures associated with nerve agent poisoning. 48 benzodiazepines can, in addition to acting as ␥-aminobutyric acid receptor agonists, also inhibit paf receptors. 49 paf is a dual proinflammatory and neuromessenger phospholipid molecule that participates in pathological phenomena that include protease synthesis, lethal anaphylaxis, vesication and nmda glutamate receptor excitotoxicity, seizure, and neuronal degeneration (for review, see ref. 14) . mice are protected from lethal anaphylaxis by serine protease inhibitors or paf antagonists. 46, 49 the ␥-aminobutyric acid receptor agonist and paf receptor antagonist actions of benzodiazepines might synergistically protect against nerve agent cholinergic and inflammatory toxicities. furthermore, proteolytic and inflammatory pathways are also strongly implicated in nmda receptor-mediated seizure and neuronal degeneration. knockout mice deficient in the protease t-pa were resistant to seizure and neuronal pathology caused by excitotoxins (for review, see ref. 50) . furthermore, the synthetic protease inhibitor tpa-stop (american diagnostica, greenwich, ct) suppressed excitotoxin-induced seizure and neuronal degeneration (for review, see ref. 50 ). experimental evidence generated in in vitro and in vivo models of chemical agent toxicity supports the concept of anti-inflammatory mtmc drugs. sm exposure increases the inflammatory il-8 in hek and human small airway cell (hsac) cultures. 40,51 sm-increased il-8 in hek cultures is a biomarker of dermal inflammatory pathology that is used as an in vitro drug screening and treatment model to complement in vivo studies in the mevm. 40 likewise, sm-increased il-8 in hsac has recently been developed as an in vitro model for treatment of inhalation toxicity. 51 hsac cultures exposed to 25 to 200 m sm or 0.2 to 6.4 ppm/min phosgene demonstrated significantly increased il-8. 52 a maximum increase of approximately 1,000 pg/ml il-8 in hsac cultures exposed to phosgene or sm was observed. 52 the pattern of il-8 response, increase to maximum levels followed by inhibition at higher cytotoxic doses, was similar for both agents. ibuprofen (62, 125, 250, 500, and 1 ,000 m) significantly diminished phosgene-increased il-8 in sac cultures exposed to 2 ppm/min phosgene. 52 furthermore, the doses of ibuprofen that decreased phosgene-increased il-8 approximately 50% in hsac (125 and 250 m) also inhibited sm-increased il-8 in hek cultures to about the same extent. 51, 53 serine protease inhibitors and parp inhibitors can have antiinflammatory pharmacology, and members of these two distinct classes of compounds have shown efficacy for nerve and blister agents in vivo. 14, 36, 46, 53 serine protease inhibitors can prolong the survival of animals intoxicated with the nerve agent soman, 41 and can also protect against vesication caused by the blister agent sm. 36 parp inhibitors can reduce soman-induced neuronal degeneration 53 and sm-induced epidermal necrosis. 36 candidate anti-inflammatory mtmc drugs to include protease inhibitors, paf antagonists, parp inhibitors, nmda receptor antagonist, and adenosine agonists have been suggested (table i) . 14 adenosine receptor agonists illustrate the potential and the complexity of mtmc drugs. van helden et al. 54 used the adenosine a1 receptor agonist n6-cyclopentyl adenosine as a countermeasure to soman poisoning. without any supportive treatment with atropine, oxime, or diazepam, n6-cyclopentyl adenosine protected rats from convulsive activity and respiratory distress and improved 24-hour survival. 54 reduced ach levels in rat brains were reported in this study, and the protection was attributed to cholinergic mechanisms wherein the adenosine a1 receptor agonists binding to the a1 adenosine receptor caused inhibition of ach release. 54 however, the cholinesterase-inhibiting nerve agents sarin, tabun, and soman can interact directly with brain a1 adenosine receptors and may competitively alter the action of these agents at adenosine receptor sites. 14, 55 adenosine receptors can moderate the activation of other receptors that influence neurotransmitter release and/or synaptic transmission, e.g., ach and nmda receptors. 54,56 a2, a3, and a1 adenosine receptor agonists can further influence mast cell degranulation, and adenosine-directed treatment modalities have been suggested for asthma. 57 neutrophil infiltration of dermal inflammatory sites, similar to that also observed in sm pathogenesis, was inhibited by administration of adenosine a1 receptor agonists. 14, 58 this anti-inflammatory action of adenosine agonists was reversed by an nmda agonist and was mimicked by a glutamate nmda receptor antagonist. 58 hence, in this model, central nmda receptor activity can control peripheral neutrophil accumulation, and adenosine a1 receptor agonists can influence this nmda receptor excitation-mediated inflam-mation. 58 virag and szabo 59 have further demonstrated that adenosine can inhibit the activation of parp, and they propose that this may affect cell death and inflammation. thus, adenosine agonists can influence multiple inflammatory biochemical pathways associated with nerve and blister agent toxicity and are candidate mtmc (table iii) . the activation of inflammatory responses that contribute to pathology is not restricted to chemical agents. t-2 toxin, as a major trichothecene mycotoxin, has some radiomimetic properties, and the results of its action, like the blistering induced by sm, can be alleviated by treatment with glucocorticoid hormones. 60, 61 moreover, paf seems to have an important role in the pathophysiology of t-2 toxicosis because its selective antagonist bn 520121 can prolong survival of rats exposed to lethal doses of this toxin. 62 trichothecene mycotoxin also induce mesenteric 63 and subcutaneous mast cell degranulation 64 and increased secretion of interleukins il-1␤ and il-6 in vitro. this suggests that superinduction of cytokines might be one of the mechanisms of t-2 toxin-induced tissue damage. 65 the inflammagenic properties of mycotoxin may further contribute to core pathologies such as ards. likewise, inflammatory cytokines and the complication of ards are associated with the pathology of infectious agents such as anthrax, smallpox, ebola, and, of course, the corona virus infection that causes severe acute respiratory syndrome. [66] [67] [68] the mtmc hypothesis suggests that inflammation is not just a biomarker, but is also a major cause of toxicity for many chemical warfare agents and toxic chemicals. as a counterpoint to the anti-inflammatory mtmc drugs, substances that increase inflammation can synergistically augment chemical toxicity. roth and coworkers (for review, see ref. 69 ) demonstrated that a small dose of the inflammagenic endotoxin lipopolysaccharide (lps), which is without effect by itself, markedly enhances the hepatotoxic effects of aflatoxin b 1 . a similar effect of lps endotoxin occurs with other toxic chemicals and other target organs (for review, see ref. 69 ). stone et al. 70 have reported that lps or the inflammatory cytokines tnf or il-1 enhanced the cytotoxic effects of the sm stimulant 2-chloroethyl ethyl sulfide. whether the results of coexposure to an inflammagen such as lps or an intrinsic property of a toxic agent, the inflammatory response can augment sensitivity to chemical toxicities. thus, inflammation associated with chemical toxicity is a target for anti-inflammatory mtmc. the mtmc hypothesis is supported by experimental evidence that distinct classes of drugs such as protease inhibitors and anti-inflammatory parp inhibitors have demonstrated efficacy against the nerve agent soman and the blister agent sm. 33, 36, 41, 53 moreover, t-2 mycotoxin and sm injury can be alleviated by treatment with glucocorticoid hormones. 60, 61 by focusing on key biochemical signal pathways and cellular processes that ultimately contribute to pathologies associated with chemical toxicity, the mtmc hypothesis provides the possibility of developing single countermeasure drugs with prophylactic or therapeutic efficacy against distinct pathologies caused by multiple classes of toxic chemical agents. 14 furthermore, the mtmc concept may not be limited to chemical agents. many biological toxins and infectious agents cause inflammatory responses that contribute to pathology. finally, because mtmc addresses clinically recognized pathologies such as ards, collateral improvements in general health care and emergency medicine are possible. once one looks past the diverse chemistry, sites of action, and symptoms, common biochemical pathways may exist for many chemical threat agents. even when threats extend to mixtures of agents and toxic industrial chemicals, or new threat agents evolve, such biochemical signaling pathways may remain more constant and amenable to medical intervention. defining and interdicting these pathways provide strategies for developing civil defense and military preparedness against chemical threats. of perhaps equal significance, the mtmc hypothesis provides an incentive for dialogue and cross-fertilization between diverse chemical defense, academic, clinical, and drug discovery research interests to combat chemical warfare and chemical terrorism. using mtmc to blunt pathological actions of chemical agents is an attractive possibility. a higher form of killing: the secret story of gas and germ warfare chemical and biological weapons: the silent killers organophosphorus compounds as chemical warfare agents organophosphate poisoning: an overview iraq's chemical warfare: case proved analysis of two chemical weapons samples from the iran-iraq war nerve gases as means of chemical terrorism in japan anatomy of aum shinrikyo's organization and terrorist attacks with chemical and biological weapons the first mass chemical terrorism using sarin in matsumoto at war with chemistry putative role of proteolysis and inflammatory response in the toxicity of nerve and blister chemical warfare agents: implications for multi-threat medical countermeasures putative roles of inflammation in the dermatopathology of sulfur mustard a hypothesis for synergistic toxicity of organophosphorus poisoning-induced cholinergic crisis and anaphylactoid reactions efficacy of ibuprofen and pentoxifylline in the treatment of phosgene-induced acute lung injury the temporal profile of cytokines in the bronchoalveolar lavage fluid in mice exposed to the industrial gas phosgene severe poisoning with the organophosphorus insecticide thionazine-2 cases with the development of ards (adult respiratory distress syndrome) molecular basis for mustard-induced vesication protease released in organ culture by acute inflammatory lesions produced in vivo in rabbit skin by sulfur mustard: hydrolysis of synthetic peptide substrates for trypsin-like and chymotrypsin-like enzymes response of normal human keratinocytes to sulfur mustard (hd): cytokine release using a non-enzymatic detachment procedure inflammatory gene expression in mouse skin following sulfur mustard exposure mcp-1, il-1a , and gro in rabbit inflammatory skin lesions produced by the chemical irritant sulfur mustard inflammatory cytokine response in sulfur mustard-exposed mouse skin neuropharmacological mechanisms of nerve agent induced seizure and neuropathology biochemical changes associated with muscle fibre necrosis after experimental organophosphate poisoning quantitative cytophotometric analyses of mesenteric mast cell granulation in acute soman intoxicated rats organophosphateinduced histamine release from mast cells comparative efficacy of antidotal oximes: protection and inflammatory markers following exposure of guinea-pigs to sarin vapor soman-induced interleukin-1a mrma and protein in rat brain pathology of nerve agents: perspectives on medical management central neuro-inflammatory gene response following soman exposure in the rat inhibition of kainic acid induced expression of interleukin-1 a and interleukin-1 receptor antagonist mrna in the rat brain by nmda receptor antagonists powerful anticonvulsant action of il-1 receptor antagonist on intracerebral injection and astrocytic overexpression in mice development of medical countermeasures to sulfur mustard vesication systemic administration of candidate antivesicants to protect against topically applied sulfur mustard in the mouse ear vesicant model (mevm) in vivo therapeutic prophylactic protection against cutaneous sulfur mustard injury using the mouse ear vesicant model (mevm) army medical research institute of chemical defense antiinflammatory effects of a novel, potent inhibitor of poly (adp-ribose) polymerase 4-substituted-5-acetyl-2-methyl-6-phenyl-3(2h)pyridazinones as pge2 and il-1 release inhibitors from mouse adherent macrophages suppression of sulfur mustard-increased il-8 in human keratinocyte cell cultures by serine protease inhibitors: implications for toxicity and medical countermeasures protective action of the serine protease inhibitor n-tosyl-l-lysine chloromethyl ketone (tlck) against acute soman poisoning mechanisms of and experimental therapy for acute soman-induced bronchial spasm the cholinergic "pitfall": acetylcholine, a universal cell molecule in biological systems, including humans anticholinergic therapy for airway diseases proteases and protease inhibitors in experimental shock states action of 2-pam and of acetylcholinesterase inhibitors on anaphylactic shock in the rabbit vesicants and nerve agents in chemical warfare. decontamination and treatment strategies for a changed world paf-acether-induced mortality in mice: protection by benzodiazepines neuronal death in the hippocampus is promoted by plasmin-catalyzed degradation of laminin suppression of sulfur mustard-increased il-8 in human keratinocyte cell cultures by the cox inhibitor indomethacin, the poly(adp-ribose) polymerase (parp) inhibitor 3-(4ј-bromophenyl)ureidobenzamide and the calmodulin antagonist fluphenazine sulfur mustard-and phosgene-increased il-8 in human small airway cell cultures: implications for medical countermeasures against inhalation toxicity adp-ribose) polymerase inhibitor, is neuroprotective against soman-induced seizure-related brain damage new generic approach to the treatment of organophosphate poisoning: adenosine receptor mediated inhibition of ach-release binding of some organophosphorus compounds at adenosine receptors in guinea pig brain membranes adenosine a2a receptor interactions with receptors for other neurotransmitters and neuromodulators adenosine, mast cells and asthma spinal cord adenosine receptor stimulation in rats inhibits peripheral neutrophil accumulation. the role of n-methyl-daspartate receptors purines inhibit poly(adp-ribose) polymerase activation and modulate oxidant-induced cell death attenuation of general toxic effect of t-2 toxin in rats treated with various formulations of methylprednisolone potential antiinflammatory treatments against cutaneous sulfur mustard injury using the mouse ear vesicant model protective effect of paf-acether antagonist, bn 52021, in trichothecene toxicosis mesenteric mast cell degranulation in acute t-2 toxin poisoning cutaneous injury by topical t-2 toxin: involvement of microvessels and mast cells the effect of t-2 toxin on il-1a and il-6 secretion in human fetal chondrocytes acute respiratory distress syndrome cutting edge: impairment of dendritic cells and adaptive immunity by ebola and lassa viruses on the role of macrophages in anthrax bacterial lipopolysaccharide exposure alters aflatoxin b(1) hepatotoxicity: benchmark dose analysis for markers of liver injury lipopolysaccharide enhances the cytotoxicity of 2-chloroethyl ethyl sulfide key: cord-288348-b10e023s authors: estes, mary kolb; graham, david yates title: epidemic viral gastroenteritis date: 1979-06-30 journal: the american journal of medicine doi: 10.1016/0002-9343(79)90457-1 sha: doc_id: 288348 cord_uid: b10e023s abstract epidemic viral gastroenteritis is a significant world wide problem. in developed countries, gastroenteritis accounts for significant morbidity and loss of time from work; in the third world it is the leading cause of mortality among infants and children. recent technologic advances have been associated with an explosion of research activity. two virus groups, the norwalk-like agents and the rotaviruses, are currently accepted as causative agents of viral gastroenteritis in man. the problem of viral gastroenteritis is reviewed both from a current and a historic perspective. epidemic viral gastroenteritis is a significant world wide problem. in developed countries, gastroenteritis accounts for significant morbidity and loss of time from work; in the third world it is the leading cause of mortality among infants and children. recent technologic advances have been associated with an explosion of research activity. two virus groups, the norwalk-like agents and the rotaviruses, are currently accepted as causative agents of viral gastroenteritis in man. the problem of viral gastroenteritis is reviewed both from a current and a historic perspective. viral gastroenteritis, more correctly termed acute enteritis or enteropathy, is a common cause of illness in man. in the united states, the illness is usually not of significant severity or duration to cause the patient to seek medical attention. therefore, even widespread outbreaks remain unrecognized unless they are brought to the attention of the health authorities because of the large number of people involved. viral gastroenteritis is a more significant problem in underdeveloped countries; estimates are that it is responsible for more than 500 million episodes of diarrhea and 5 to 18 million deaths annually [il. a large number of viral agents have been implicated as causative factors of gastroenteritis. many studies have consisted simply of screening stool specimens of diarrhea1 patients for viral agents; the frequency of positive cultures may or may not have been compared with results obtained in a control population. such studies have accumulated data to foster epidemiologic hypotheses which require further testing; they have not provided definite evidence of causation. current knowledge of the viral stool-shedding patterns is rudimentary. for example, in a recent study virus excretions in the stools of 27 babies were followed for one year [2] . the study revealed that babies frequently shed viruses and that the viral excretion patterns change rapidly. fifteen per cent of the stools from healthy babies contained viruses, including rotaviruses, adenoviruses, astroviruses, caliciviruses and small unidentified viruses. during hospitalization, the proportion of infants who excreted viruses increased to 44 per cent, with many of the viruses appearing to be hospital-acquired. this study emphasized the current problems in interpreting data from studies of fecal viral excretion, including the need for concurrent strict control populations. with these problems in mind, one can critically evaluate previous studies and carefully plan future studies to differentiate between diarrheal-associated viruses and proved enteritis-inducing agents. proof of causation of infectious disease was codified by robert koch in 1890 [3]. koch's postulates stated that an organism must be regularly isolated from patients with the illness and in a logical relationship to the pathologic changes observed. the isolated organism should be grown in pure culture and shown to be capable of causing the same illness in susceptible animals or human subjects. finally, the organism should be reisolated from the subject with the experimentallyinduced disease. noting that many viruses cannot be cultured, thomas rivers [4] and later robert huebner [5] proposed modifications of koch's postulates that would allow proof of causation of viral illnesses. they suggested that for viral diseases the specific virus must be regularly associated with the disease, that the illness must be transmissible to susceptible hosts from material known to be free from nonviral agents, and that control and immunologic studies be carried out to exclude that the virus was fortuitously present or picked up from the experimental host. supporting immunologic critera are also important, including evidence that specific viral antibody is absent before infection and that it is produced by the infection; the absence or presence of antibody should correlate with disease susceptibility and protection, respectively. of the many viruses identified in stools, only two groups have met the criteria as definite etiologic agents of epidemic gastroenteritis in human subjects: rotaviruses and the small 27 nm agents [norwalk-like agents) ( of investigations which began with an epidemic of gastroenteritis occurring in the newborn in the baltimore-washington area in the fall of 1941. they were able to obtain multiple passages of the infectious agent (the baltimore agent) by administering filtered stool to calves. calves were found to be uniformly susceptible with an incubation period of two to five days. the illness in calves had a mortality rate of 13 per cent, and infection was followed by homologous immunity. horizontal infection was evident, and strict isolation procedures were employed. the baltimore agent is the only isolate from the early studies that was saved, and recent characterizations revealed that it is probably a rotavirus [71. in 1945 reimann, hodges and price [8] reported the results of an investigation of an epidemic occurring in philadelphia between september and december 1943 and recurring in 1944. they also provided the first overview of the problem. they were the first to use volunteer subjects to passage their material. however, the results were inconclusive because an epidemic of viral enteritis was ongoing in the community, and no attempts were made to isolate the subjects. on december 16,1946, an epidemic of gastroenteritis began in the marcy state hospital near utica, new york. subsequently, gordon and his associates [9,10] published a series of papers on the marcy strain that included volunteer studies showing the transmission of a nonbacterial gastroenteritis. at the same time, an outbreak of gastroenteritis occurred in a group of families being followed in cleveland as part of a long-term study of illnesses. this group observed that nonbacterial gastroenteritis was the second most common disease in their families. the family study (fs) gastroenteritis agent was also passed in vol-unteer subjects and compared with the marcy strain [ll] . clear differences were found between the two isolates with respect to symptoms, incubation period, disease duration and immunity. infection with each agent conferred homologous but not heterologous immunity. japanese workers showed that infection with a japanese agent, the niigata strain, conferred protection against infection with the marcy strain [el. work was largely abandoned during the 1950s and 1960s because of the inability to either propagate or to visualize the infectious agents. advances in knowledge and methodology for the detection of fastidious and noncultivatable viral agents in the 1960s led to a renewal of interest in gastroenteritis viruses in the 1970s. the application of immune electron microscopy, the technique used to identify hepatitis a virus particles in stools, to diarrhea1 samples brought about the successful identification of two enteritis viruses in human subjects. the recent development of rapid and very sensitive tests (radioimmunoassay, immune adherence hemagglutination and enzyme-linked immunosorbent assays) which can detect symptomatic and asymptomatic infections should quickly expand our knowledge of the natural history and epidemiology of these diseases. the history of investigations in viral gastroenteritis is a good example of the fact that major scientific advances often require and parallel new technologic opportunities. 27 nm particles (norwalk-like agents). on october 30,1968, an attack of viral gastroenteritis occurred in an elementary school in norwalk, ohio [13] . in this attack, 50 per cent of the teachers and students were affected within a 24-hour period with an illness characterized by nausea, vomiting and abdominal cramps. the symptoms lasted from 12 to 24 hours. no patient required hospitalization, and recovery was complete. because of the suddenness of the attack and the large number of persons involved, public health authorities including the center for disease control were asked to investigate the outbreak. no evidence of food or water spread was found. no pathogenic bacteria were discovered, and the symptomatic characteristics of the illness were different from common bacterial diarrheas. investigators at the national institutes of health began work with this "norwalk agent." the disease was transmitted to volunteer subjects which provided enough infectious material (stool filtrates] for the performance of detailed clinical studies and preliminary laboratory analyses of the physical and biochemical properties of the agent [14, 15] . stool filtrates were characterized as being free from bacteria, fungi, bacteriophages and enterotoxin. although no viral agent could be propagated in a variety of tissue culture systems or in laboratory animals [including primates), the studies in volunteer subjects demonstrated that the infectious agent would pass through a 60 pm filter but that it was retained by a 20 frn filter. it was also heat stable, resistant to ether and acid treatments, and produced ho-mologous immunity following infection [15] . subsequently, the agent was propagated briefly in culture of material from intestinal organs [16] and in 1972 kapikian and co-workers [l7] succeeded in identifying a 27 nm particle by immune electron microscopy of stools from infected volunteer subjects. in 1973, two groups described histologic studies of small bowel mucosa in volunteer subjects infected with the agent [18, 19] . histologic changes included abnormalities in the mucosal absorptive cells, mucosal inflammation, villus shortening and secondary crypt hypertrophy. in addition, there were decreases in specific activities of intestinal brush border enzymes. these changes were associated with malabsorption of the carbohydrates, d-xylose and lactose, and transient steatorrhea. the histologic findings became apparent several hours prior to the onset of symptoms and persisted for several days. histologic abnormalities were also found in patients without symptoms, but severe lesions tended to be associated with the severe illnesses. two other norwalk-like agents have been obtained in outbreaks of gastroenteritis: the h agent in honolulu in march 1971 and the mc agent, in an attack in montgomery county, maryland, in june 1971. studies in volunteer subjects suggested that the norwalk and hawaii agents were antigenically dissimilar, whereas the norwalk agent conferred immunity to challenge with the mc agent [zo] . rechallenge studies in volunteer subjects with the norwalk agent revealed a pattern of susceptibility to re-infection with the same agent [zl] . six of 12 subjects became ill with the first challenge, and, when rechallenged many months ltiter, the same six subjects became ill again. subsequent rechalienge of five of the six "susceptible" subjects several weeks later revealed immunity in four of them. an increase in serum antibody titer was demonstrable in those patients who became ill but apparently offered no protection. work with these agents has been hampered by the lack of a reliable cultivation method in vitro. in recent studies the transmission of infection to chimpanzees has been reported but, unfortunately, no signs of clinical illness (such as vomiting or diarrhea] occurred [22] . the illness in chimpanzees was identified by use of a highly-sensitive solid phase microtiter radioimmunoassay [23] which detected both virus excretion in the stool and increase in serum antibody. in recent immunologic studies it has been demonstrated that serum antibody to the norwalk agent is acquired gradually; by the fifth decade, one half of the adults tested had antibody [23, 24] . similar 27 nm viral agents have been under investigation in the united kingdom and in japan [25, 26] . the 27 nm particles have been called "parvovirus-like agents," dna viruses, based on their small size, density in cesium chloride of 1.38 to 1.41 g/ml, and stability to ether, acid and heat. however, this provisional designation may be a misnomer. correct classification of these agents must await propagation of the agent(s) to sufficient titers to allow biochemical characterization, including type of nucleic acid. the limited information available about 27 nm agents is also consistent with their classification as caliciviruses, rna viruses which have recently been demonstrated to cause enteritis in animals [27] . until a reproducible system is devised for the propagation of the 27 nm agents, future investigations will continue to progress at a slow rate. rotaviruses. in 1973, bishop et al. [28] reported finding viral particles in biopsy specimens of duodenal mucosa from children with gastroenteritis in melbourne, australia. this same year flewett and associates [29] in england described similar viral particles in diarrhea1 feces. it was soon recognized that the particles are morphologically similar to the epizootic diarrhea virus of infant mice and to the nebraska calf diarrhea virus [30] . rotaviruses are a major cause of diarrhea1 illnesses in young mammals of a wide variety of species. these isolates have also been termed infantile gastroenteritis virus, reo-like virus and duovirus; however, the designation of "rotavirus" as a separate genus of the reoviridae family has recently been accepted by the international committee on taxonomy of viruses as their official name. the derivation of that terminology is based on the electron microscopic appearance of a wheel with radiating spokes. rotaviruses have been established as enteritis viruses by isolation and purification from stools of subjects suffering from gastroenteritis, and by induction of disease and seroconversion in both animals and volunteer subjects with purified preparations, epidemiologic studies on the prevalence of rotavirus infections have shown these ubiquitous agents to be a major cause of gastroenteritis in children. in a typical study, davidson et al. [31] found that 52 per cent of the cases of acute gastroenteritis in 378 children hospitalized in australia were caused by rotavirus. an etiologic agent was identified in 76 per cent of these cases, and the remaining causes were shigella 1 per cent, escherichia coli 2 per cent, enterovirus 2 per cent, adenovirus 7 per cent and salmonella 11 per cent. similar findings have been repeated the world over. rotavirus infections usually predominate during the winter season with an incubation period of two to four days. symptomatic infections are most common in children aged six months to six years, and transmission of rotavirus gastroenteritis appears tb be by the fecal-oral route. clinical features of rotavirus gastroenteritis in infants and children include diarrhea, vomiting, fever and abdominal pain [32] . in contrast to the clinical illness seen in hospitalized infants, adult contacts often become infected, as evidenced by-seroconversion, but they suffer only mild symptoms and the virus is rarely detected in the stool [32, 33] . the failure to identify virus in the stool may reflect the insensitive methods used for viral detection. the ubiquitous nature of rotaviruses can also be demonstrated by measuring the prevalence of immunity in the population. by age six, 60 to 90 per cent of children have serum antibody titers [34, 35] . the protective nature of the circulating antibodies remains unclear, since it is known that both human subjects and animals can become infected even when they possess detectable immunity [33, 36] . local immune factors, such as secretory immunoglobulin a or interferon, may therefore be important in protection against rotavirus infection, alternatively, reinfection in the presence of circulating antibody could reflect the presence of multiple serotypes of virus [37] ; at least four agents in human subjects have been characterized to date [38-411. asymptomatic infections are common in infants before the age of six months, the time during which protective antibody acquired passively by newborns should be present and active [42, 43] , and breast-fed babies excrete fewer particles/g feces than bottle-fed babies although both groups of babies become infected [42] . rotavirus antibody has been detected in colostrum for up to nine months postpartum [44] . other investigations have emphasized characterization of rotavirus particles and replication patterns in an attempt to understand the mechanism of infection and pathogenesis. such studies have shown that the rotavirus agents in human subjects are very similar in physical properties to the diarrhea1 agents in animals recognized by veterinary researchers since 1969 [30] . rotaviruses contain segmented double-stranded rna genomes with the pattern of rna segments varying according to the species of origin [45] . the rna pattern is becoming increasingly important as a method of identifying species of origin of rotavirus isolates as laboratory contamination and cross-species infections occur. the rotaviruses are double-shelled particles with an average diameter of 70 nm. the outer-shelled particle contains type-specific antigenic components, whereas the inner shell contains type-common antigenic determinants [46, 47] . the outer shell presumably contains a hemagglutinin [48] and glycoproteins [49] , but the role(s) of the various protein components in viral infectivity or virus neutralization remain(s) to be elucidated. the double-shelled particles exhibit a density of 1.36 g/ml in cesium chloride and have been referred to as smooth particles. these particles appear to be the infectious particles as determined by their ability to induce viral antigen detectable by immunofluorescence in cells in tissue culture [50] , particles lacking the outer shell have a density of 1.38 g/ml in cesium chloride and have been referred to as rough particles. the single-shell particles contain an rna-dependent rna polymerase activity which can be assayed directly. alternatively, the polymerase can be activated from the double-shell particles by treatment with chelating agents which reportedly remove the outer shell from the particles [51] . the rotaviruses appear to be stable entities. their resistance to acid, like most enteric viruses, guarantees their survival after transversing the acidic environment of the stomach, in addition, these viruses are stable to freezing, sonication, treatment with liquid solvents and heating (50"c). the rotaviruses, unlike the polioviruses, are not stable to heating in the presence of 1 m magnesium chloride, although they are stabilized in the presence of magnesium sulfate [52] . such stabilization could be useful to prepare vaccines not requiring refrigeration for distribution in tropical climates. biologically, the rotaviruses are fastidious organisms. the agents from human subjects have not been successfully cultured in vitro in most laboratories, although wyatt et al. [53] reported limited success in growing one isolate in human embryonic kidney cells for up to 14 passages. rotavirus agents isolated from animals, including calves, pigs and monkeys, have been able to be cultured or adapted to growth in tissue culture in the laboratory. success with cultivation and development of plaque assays for the calf and simian agents have required the use of proteolytic enzymes, such as pancreatin or trypsin [54, 55] . the agents from the calf and monkey, now easily grown in tissue culture, are being used as model viruses to study rotavirus replication, and to enhance and develop methods for rotavirus detection. the inner coat of all known rotaviruses contains common antigens. this property has allowed the development of diagnostic procedures for infected human subjects using calf and simian viruses in place of the more fastidious agent isolated from human subjects. rotaviruses from many species, including human subjects, have been used successfully in experimental infections of colostrum-deprived gnotobiotic animals [56, 67] . such studies have suggested that cross infection between species may occur. our knowledge of the histopathology and pathophysiology of rotavirus infections has come from analyses of such infections in animals and from limited studies of mucosal biopsy specimens from infected children. the general pattern of infection involves virus penetration and infection of the differentiated enterocytes in the villi of the small intestine [58] . rotaviruses multiply in the cytoplasm of these cells and damage the absorptive cells, resulting in damage to both the digestive and the absorptive functions. available evidence suggests that such damaged cells are sloughed into the small intestine; lysis of the infected cells releases the virus into the intestine, resulting in the large quantities of virus detected in stools of infected subjects. these studies suggest that the diarrhea caused by rotavirus infection is due to malabsorption which also includes impaired glucose-sodium absorption. the highly differentiated absorptive villous cells are replaced by immature crypt cells that are not able to immediately compensate for the absorptive defect [59] . treatment of viral gastroenteritis. treatment of viral gastroenteritis is supportive. death is associated with loss of electrolytes and water, leading to dehydration, acidosis and shock; it is not due to an irreversible effect of the causative agent. due to the damage to the intestinal digestive and absorptive functions, initial therapy should include withdrawal of milk or lactose-containing products, dehydration should be combatted with the oral replacement of fluids and electrolytes. recent studies have suggested that oral therapy, using sugar and electrolyte solutions, is as effective as intravenous therapy in most infants and children. of great interest are two recent controlled doubleblind studies that compared the use of sucrose-electrolyte solutions with glucose-electrolyte solutions; both sugar electrolyte solutions were found to be effective in the treatment of rotavirus diarrhea [60, 61] . this is an important observation since sucrose is less expensive and more readily available than glucose, particularly in developing countries in which the mortality from rotavirus infection is a significant problem. in addition, it appeared that oral glucose-electrolyte solutions are effective therapy for diarrhea1 disease of both bacterial and viral etiology. although limited knowledge exists, widely administered antidiarrheal agents, such as kaolin-pectin or lomotip, do not appear to be useful in the relief of viral diarrhea in children [62] . areas for further research. despite the tremendous amount of investigative work carried out in viral gastroenteritis, an understanding of the natural history and epidemiology of this disease is still lacking. both rotaviruses and the norwalk-like viral agents have been established as etiologic agents in some cases of viral gastroenteritis. the role and clinical significance of other viral agents that have been associated with diarrhea in man, including adenoviruses, enteroviruses, astroviruses, coronaviruses and caliciviruses, remain to be elucidated. the inability to cultivate rotaviruses and norwalk-like viruses from human subjects has hampered progress in elucidating the natural history, virology and epidemiology of these illnesses. to date, no small animal model has been identified which would permit the study of the replication of these viruses in vivo. due to the inability to cultivate the viruses in vitro, there has been slow development of rapid, reliable and sensitive means for detecting the agenfs and comparing the antigenic variations of viral isolates or strains recovered from differing geographic areas. four strains of rotavirus from human subjects and at least two strains of norwalk-like viruses have been reported; future studies need to determine the clinical significance of these strain differences. it remains to be established whether avirulent as well as virulent strains exist, whether defective strains are produced and whether recombinant viruses occur. recombinant viruses are a particular possibility if one considers that the segmented rna genomes of the rotaviruses are similar to those of the influenza viruses. the role of the immune response in viral gastroenteritis requires further analysis, including the possible protective effects of serum and local intestinal antibody, interferon and cellular immunity. in addition, the effect of host nutritional status is unknown. our knowledge of the epidemiology of viral gastroenteritis infections is still in its infancy. most studies have focused on hospitalized children and adults; these studies may reflect a considerable bias in relatic iship to the occurrence of infections in the population at arge, since most of the illnesses are of insignificant severity and duration for the patient to seek medical attention. transmission of both the norwalk-like agent and rotaviruses appears to be by fecal-oral routes. the role of contaminated food and water in the transmission of gastroenteritis requires further examination, as does the effectiveness of our present sanitation procedures for eradication of the agents. the role of rotaviruses in persistent and recurrent infections is unknown. it has been demonstrated that rotaviruses can establish persistent infections in cell culture [63] , but the control mechanisms responsible for chronic infections in animals and/or cell lines have yet to be unraveled. lastly, the hosts and reservoirs of these viruses remain to be established. unfortunately, possible cross-species infections and the lack of reliable sensitive methods to detect rotavirus strain differences do not make these goals easily attainable. finally, the possible role of these viral agents in the establishment and maintenance of chronic diarrhea1 disease warrants further investigation, both with respect to chronic diarrhea of childhood and the relationship to inflammatory bowel disease. although prevention of gastroenteritis by vaccination is suggested in many publications, the primitive state of our knowledge of gastroenteritis viruses makes such discussions truly academic at the present time. acute infectious nonbacterial gastroenteritis: intestinal histopathology. histologic and enzymatic alterations during illness produced by the norwalk agent in man comparison of three aeents of acute infectious nonbacterial eastroenteritis bv c;oss-challenge in volunteers experimental infection of chimpanzees with the norwalk agent of epidemic viral gastroenteritis solid-phase microtiter radioimmunoassav for detection of the norwalk strain of acute nonbacterial, "epidemic gastroenteritis virus and its antibodies prevalence of antibodv to the norwalk aeent bv a newlv develoned immune adherence hemaggl<tination assay parvovirus-like agent detected by em in infantile diarrhea and its relationship to epidemic gastroenteritis in school children in japan (abstract) a virus from epidemic vomiting disease bridaer ic: isolation of small viruses resembline astroviruses and caliciviruses from acute enteritis of calve virus particles in enithelial cells of duodenal mucosa from children with acuie non-bacterial gastroenteritis virus narticles in eastroenteritis relation between viruses from acute gastroenteritis of children and newborn calves importance of a new virus in acute sporadic enteritis in children clinical, laboratory and epidemiological features of a viral gastroenteritis braska calf diarrhea virus (bovine rotavirus). intervirol in infants and children human reovirus-like -" agent as the major pathogen assoicated with "winter" 37. gastroenteritis in hospitalized infants and young children epidemiology of human n engl mm: distribution and titres of rotavirus antibodies in rotavirus types 1 and 2 as studied by enzyme-linked imdifferent age groups holmes ih: comparison of an enmunosorbent assay rotavirus antibodies with complement fixation in an epi different serotypes of human rotavidemiological surgery eoizootiologv of bovine rotavirus infections. vet ruses serotypes of human rotavirus further biochemical &ltivation of human reovirus-like agent of infantile diarcharacterization, including the detection of surface glycoproteins, of human, calf, and simian rotaviruses separation and infectivity of two particle types 54. of human rotavirus a plaque assay for ribonucleic acid polymerase activity associated with simian rotavirus sall rotavirus stabdity and inactivation plaque assay of neonatal calf diarrhea virus and the neutralizing antibody in human sera sequential enteric illnesses associated with different rotavirus serotypes more serotypes of human rotavirus asymptomatic endemic rotavirus infections in the newborn lancet 2: 1149,1977. shedding of human reovirus-like agent in gnotobiotic newborn pigfets with experimentally-induced diarrhea naturally occurring and experimentally induced rotaviral infections of domestic and laboratory animals pathology of neonatal calf diarrhea induced by a reo-like virus secretory iga mechanisms in the pathogenesis of diarrhea: a review oral hydration in ro comparison of human and animal rotavirus strains by gel electrophoresis tavirus diarrhoea: a double blind comparison of sucrose of viral rna comparison of sucrose morphological and antinenic relationships between viruses lrotavirusesl from with glucose in oral therapy of infant diarrhoea. lancet 1: 277.1978. ac& gastroenteritisbf children, calves,.piglets. mice and 62 antidiarrheal agents in the treatment of acute diarrhea in children antigenic relationship 844.1976. between human and simian rotaviruses hemagglutination and hemaggfutination-inhibition studies with a strain of ne key: cord-027337-eorjnma3 authors: fratrič, peter; sileno, giovanni; van engers, tom; klous, sander title: integrating agent-based modelling with copula theory: preliminary insights and open problems date: 2020-05-22 journal: computational science iccs 2020 doi: 10.1007/978-3-030-50420-5_16 sha: doc_id: 27337 cord_uid: eorjnma3 the paper sketches and elaborates on a framework integrating agent-based modelling with advanced quantitative probabilistic methods based on copula theory. the motivation for such a framework is illustrated on a artificial market functioning with canonical asset pricing models, showing that dependencies specified by copulas can enrich agent-based models to capture both micro-macro effects (e.g. herding behaviour) and macro-level dependencies (e.g. asset price dependencies). in doing that, the paper highlights the theoretical challenges and extensions that would complete and improve the proposal as a tool for risk analysis. complex systems like markets are known to exhibit properties and phenomenal patterns at different levels (e.g. trader decision-making at micro-level and average asset price at macro-level, individual defaults and contagion of defaults, etc.). in general, such stratifications are irreducible: descriptions at the micro-level cannot fully reproduce phenomena observed at macro-level, plausibly because additional variables are failed to be captured or cannot be so. yet, anomalies of behaviour at macro-level are typically originated by the accumulation and/or structuration of divergences of behaviour occurring at micro-level, (see e.g. [1] for trade elasticities). therefore, at least in principle, it should be possible to use micro-divergences as a means to evaluate and possibly calibrate the macro-level model. one of the crucial aspects for such an exercise would be to map which features of the micro-level models impact (and do not impact) the macro-level model. from a conceptual (better explainability) and a computational (better tractability) point of view, such mapping would enable a practical decomposition of the elements at stake, thus facilitating parameter calibration and estimation from data. moreover, supposing these parameters to be adequately extracted, one could put the system in stress conditions and see what kind of systematic response would be entailed by the identified dependence structure. the overall approach could provide an additional analytical tool for systematic risk. with the purpose of studying and potentially providing a possible solution to these requirements, we are currently working on establishing a theoretical framework integrating agent-based modelling (abm) with advanced quantitative probabilistic methods based on copula theory. the intuition behind this choice is the possibility to connect the causal, agentive dependencies captured by agentbased models with the structural dependencies statistically captured by copulas, in order to facilitate the micro-macro mappings, as well as the extraction of dependencies observable at macro-level. to the best of our knowledge, even if many research efforts targeting hybrid qualitative-quantitative methods exist in the computational science and artificial intelligence literature, the methodological connection of abm with copula theory is still an underexplored topic. a large-scale agent-based model of trader agents incorporating serial dependence analysis, copula theory and coevolutionary artificial market, allowing traders to change their behaviour during crisis periods, has been developed in [2] ; the authors rely on copula to capture cross-market linkages on macro-level. a similar approach is taken in [4] . examples of risk analysis in network-based setting can be found for instance in [5, 6] , in which the mechanisms of defaults and default contagion are separated from other dependencies observed in the market. in [3] , copula is used to model low-level dependencies of natural hazards with agent-based models, in order to study their impact at macro-level. in the present paper, we will use copula to model dependencies among agents at micro-level, and we will propose a method to combine aggregated micro-correlations at market scale. the paper is structured as follows. section 2 provides some background: it elaborates on the combined need of agent-based modeling and of quantitative methods, illustrating the challenges on a running example based on canonical trader models for asset pricing, and gives a short presentation on copula theory. section 3 reports on the simulation of one specific hand-crafted instanciation of copula producing a relevant result from the running example, and will elaborate on extensions and theoretical challenges that remain to be solved for the proposal to be operable. a note on future developments ends the paper. in financial modelling, when statistical models are constructed from time series data, it is common practice to separately estimate serial dependencies and crosssectional dependencies. the standard approach to capture serial dependence (also referred to as autocorrelation) is to use autoregressive models [10] . if the time series exhibit volatility clustering-i.e. large changes tend to be followed by large changes, of either sign, small changes tend to be followed by small changesthen it is typical to use the generalized autoregressive conditional heteroskedasticity model (garch) [11] , or one of its variants [12] . once the garch model is estimated, the cross-sectional dependence analysis can be performed on the residuals. unfortunately, autoregressive models provide only little information useful for interpretation; this is no surprise since these models are purely quantitative, and suffer problems similar to all data-driven methods. as low interpretability goes along with limited possibility of performing counterfactual or what-if reasoning (see e.g. the discussion in [9] ), such models are weakly justifiable in policy-making contexts, as for instance in establishing sound risk balancing measures to be held by economic actors. an alternative approach is given by agent-based modelling (abm), through which several interacting heterogeneous agents can be used to replicate patterns in data (see e.g. [2] ). typically, agent models are manually specified from known or plausible templates of behaviour. to a certain extent, their parameters can be set or refined by means of some statistical methods. model validation is then performed by comparing the model execution results against some expected theoretical outcome or observational data. these models can be also used to discover potential states of the system not yet observed [23] , thus becoming a powerful tool for policy-making. independently from their construction, agent models specify, at least qualitatively, both serial dependencies (for the functional dependence between actions) and cross-sectional dependencies (for the topological relationships between components), and are explainable in nature. however, they do not complete the full picture of the social system, as the focus of the designers of agent-based models is typically on the construction of its micro-level components. nevertheless, to elaborate on the connection between micro-level and macro-level components of a social system we still need to start from capturing behavioural variables associated to micro-level system components, assuming them to have the strongest effect at micro-level (otherwise the micro-level would be mostly determined by the macro-level). running example: asset market. we will then consider a paradigmatic scenario for abm: an asset market, in which traders concurrently sell, buy or hold their assets. our running example is based on canonical asset pricing models, proceeding along [17] . fundamental value. the target asset has a publicly available fundamental value given by a random walk process: where η is a normally distributed random variable with mean zero and standard deviation σ η . market-maker agent. at the end of each trading day, a market-maker agent sets the price at which a trader agent can buy or sell the asset according to a simple rule: where: the variable d(t) denotes the number of buy orders at time t, s(t) denotes number of sell orders at time t and δ is a normally distributed random variable with zero mean and constant standard deviation σ δ . the positive coefficient a can be interpreted as the speed of price adjustment. fundamental traders. fundamental traders operate with the assumption that the price of an asset eventually returns to its fundamental value. therefore, for them it is rational to sell if the value of an asset is above its fundamental value and buy if the value of an asset is below its fundamental value. their price expectation can be written as: where α is a normally distributed random variable with mean zero and standard deviation σ α . x fund can be interpreted as the strength of a mean-reverting belief (i.e. the belief that the average price will return to the fundamental value). technical traders. in contrast, technical traders, also referred to as chartists, decide on the basis of past trend in data. they will buy if the value of an asset is on the rise, because they expect this rise to continue and sell if the value is on decline. their expectation can be written as: where β is a normally distributed random variable with mean zero and standard deviation σ β . x tech can be interpreted as a strength of reaction to the trend. relevant scenario: herding behaviour. because they are intrinsic characteristics of each agent, x fund and x tech can be seen as capturing the behavioural variables we intended to focus on at the beginning of this section. now, if for all traders x fund or x tech happen to be realized with unexpectedly large values at the same time, the effect of α and β will be diminished, and this will result in higher (lower) expected value of the asset price and then in the consequent decision of traders to buy (sell). purchases (sales) in turn will lead the market-maker agent to set the price higher (lower) at the next time step, thus reinforcing the previous pattern and triggering herding behaviour. such chain of events are known to occur in markets, resulting in periods of rapid increase of asset prices followed by periods of dramatic falls. note however that this scenario is not directly described by the agent-based models, but is entailed as a possible consequence of specific classes of instantiations. herding behaviour is recognized to be a destabilizing factor in markets, although extreme time-varying volatility is usually both a cause and an effect of its occurrence. in the general case, factors contributing to herding behaviour are: the situation on the global market, the situation in specific market sectors, policies implemented by policy makers, etc. all these factors are somehow processed by each human trader. however, because such mental reasoning is only partially similar amongst the agents, and often includes non-deterministic components (including the uncertainty related to the observational input), it is unlikely that it can be specified by predefined, deterministic rules. for these reasons, probabilistic models are a suitable candidate tool to recover the impossibility to go beyond a certain level of model depth, in particular to capture the mechanisms behind behavioural variables as x fund and x tech . in the following, we will therefore consider two normally distributed random variables x fund and x tech realizing them. 1 this means that the traders will perceive the price difference in parenthesis of eqs. (3) and (4) differently, attributing to it different importance at each time step. looking at eqs. (3) and (4) we can see that the essence of an agent's decision making lies in balancing his decision rule (e.g. for the fundamental trader with the uncertainty about the asset price (e.g. α). if for instance the strength of the mean-reversing belief x fund happens to be low (in probabilistic terms, it would be a value from the lower tail), then the uncertainty α will dominate the trader's decision. in contrast, if x fund happens to be very high (i.e. from the upper tail), then the trader will be less uncertain and trader's decision to buy or sell will be determined by (f t − p t ). similar considerations apply to technical traders. assuming that behavioural random variables are normally distributed, obtaining values from the upper tail is rather unlikely and, even if some agent's behavioural variable is high, it will not influence the asset price very much since the asset price is influenced collectively. however, if all traders have strong beliefs about the rise or fall of the price of the asset, then the price will change dramatically. the dependence of the price on a collective increase in certainty cannot be directly modeled by the standard toolkit of probability, and motivates the use of copulas. copula theory is a sub-field of probability theory dealing with describing dependencies holding between random variables. application-wise, copulas are well established tools for quantitative analysis in many domains, as e.g. economic time series [15] and hydrological data [16] . consider a random variable u = (u 1 , ..., u d ) described as a d-dimensional vector. if all components of u are independent, we can compute its joint probability distribution function as f u (u 1 , .., u d ) = f u1 (u 1 ) · ... · f u d (u d ), i.e. the product of marginal distributions. in case of dependence among components we need some function that specifies this dependence. the concept of copula is essentially presented as a specific class of such functions, specifically defined on uniform marginals [13] : to obtain a uniform marginal distribution from any random variable we can perform a probability integral transformation u i = f i (x i ), where f i is the marginal distribution function of the random variable x i . in practice, when we estimate the copula from data we estimate the marginals and copula components separately. we can then introduce the most important theorem of copula theory: if f 1 , ..., f d are continuous this copula is unique. if we consider the partial derivatives of eq. (5) we obtain the density function in the form: where c is density of the copula and f i is marginal density of random variable x i . the reason why copulas gained popularity is that the cumulative distribution function f i contains all the information about the marginal, while the copula contains the information about the structure of dependence, enabling a principled decomposition for estimations. correspondingly to the high variety of dependence structures observed in the real world, there exists many parametric families of copulas, specializing for specific types of dependence. the most interesting type for economic applications is tail dependence. for example, if there is nothing unusual happening on the market and time series revolve around its mean value, then the time series might seem only weakly correlated. however, co-movements far away from mean value tend to be correlated much more strongly. in probabilistic terms, the copula describing such dependence between random variables has strong tail dependence. tail dependence does not have to be always symmetrical. certain types of copulas have strong upper tail dependence and weaker lower tail dependence. in simple terms this means that there is higher probability of observing random variables together realized in the upper quantile of their distribution than in the lower quantile. one of the parametric copulas having such properties is the joe copula [13] , illustrated in fig. 1 . this section aims to show how copulas can be effectively used to enrich stochastic agent-based models with additional dependencies, relative to micro-and macrolevels. in particular, we will focus on the phenomenon of dependence of the asset price to a collective increase in strength of belief associated to herding behaviour scenarios, observed in sect. 2.2. to model the balance between certainty and uncertainty of each trader with respect to current price of the asset, we need to set the marginal distribution functions of x fund and x tech to have a mean value and standard deviation such that if herding behaviour occurs then the uncertainty parameters α and β will play essentially no role, but if herding behaviour is not occurring, then α and β will stop traders from massive buying or selling. therefore the parameters for x fund , x tech and α, β are not entirely arbitrary. to illustrate the influence of dependence structures of behavioural random variables we will compare simulations of the market with independent behavioural random variables to simulations of the market whose random variables have a dependence structure defined by joe copula. it is important to make clear that this exercise does not have any empirical claim: it is meant just to show that copula can be used for a probabilistic characterization of the social behaviour of the system. consider a group of n total = 1000 trader agents consisting of n fund = 300 fundamental traders and n tech = 700 technical traders (this ration roughly reflects the situation of a real market). let us denote with x the vector collecting all behavioural random variables of the traders. table 1 reports all parameters we have used for our simulation. the sequence of fundamental values {f t } t t=1 is generated by eq. (1) at the beginning and remains the same for all simulations. independence scenario. at first, we will assume that the (normally distributed) behavioural random variable assigned to each trader is independent of other normally distributed behavioural variables. this means that the probability density function capturing all behavioural variables of market can be written as a simple product of marginal density functions: we have considered t = 502 time steps with initialization p 1 = 10 and p 2 = 10.02574. figure 2 illustrates the output of 100 simulations of the market with a probability density function given by the eq. (7). the simulations on average follow the fundamental price of an asset. the marginal distribution of the increments δp t taking realizations of the generated time series (fig. 2) clearly follows a normal distribution (left of fig. 4) . dependence scenario. let us keep same number of agents and exactly the same parameters, but this time consider a dependence structure between the behavioural variables described by a joe copula with parameter equal to 8. as shown in fig. 1 , this copula has strong right upper tail dependence. the joe copula is an archimedean copula, which means it attributes an univariate generator, hence drawing samples from this copula is not time consuming, even for large dimensions. in our case, each sample will be a n total dimensional vector u with components u i from the unit interval. for each agent, a quantile transformation will be made x a,i = q xa,i (u i ) by the quantile function q xa,i of i-the agent a = {fund, tech} to obtain the realization from the agent's density function. here the i-th agent a's behavioural random variable is again distributed following what specified in table 1 . running 100 market simulations, the time-series we obtain this time are much more unstable (fig. 3 ). this is due to the structural change in the marginal distribution function of δp t , which has now much fatter tails. the fatter tails can be seen on the right histogram on fig. 4 , and on the comparison of both histograms by normal qq-plots in fig. 5 . we see that for independence the increments follow a normal distribution very closely, but for dependence defined by joe copula the tails of marginal distribution deviate greatly and approximates a normal distribution only around the mean. for applications in finance it would be desirable to extend this model to consider broader contexts: e.g. a market with many assets, whose prices in general may exhibit mutual dependencies. a way to obtain this extension is to introduce adequate aggregators, following for instance what was suggested in [14, 3.11.3] . in order to apply this method, however, we need to make explicit a few assumptions. step-wise computation assumption. the increase or the decrease of the price of the asset δp t is obtained via the simulation of the agent-based market model. as it can be seen in formulas (3) and (4), the increment δp t at time t depends on the realization of random variables x at time t with agents observing previous market values p t−1 and p t−2 . this means that δp t is also a continuous random variable and its probability density function should be written as f δpt|pt−1,pt−2,x . note that this function can be entirely different at each time t depending on what values of p t−1 , p t−2 the agents observe. since the time steps are discrete, then the density functions forms a sequence {f δpt|pt−1,pt−2,x } t=t t=t0 . in this paper, for simplicity, we will not describe the full dynamics of this sequence; we will focus only on one density function in one time step, assuming therefore that the computation can be performed step-wise. by fixing time, p t−1 , p t−2 are also fixed (they have already occurred), so we can omit them and write just f δpt|x . generalizing to multiple assets. consider m non-overlapping groups of absolutely continuous random variables x 1 , ..., x m , where each group consists of behavioural random variables, or, to make our interpretation more general, predictor variables that determine the value of an asset. each group x g forms a random vector and has a scalar aggregation random variable v g = h g (x g ). this means that each value of an asset is determined by a mechanism specified by the function h g , which might be an agent-based model similar to the one explored in the previous section, but this time each group of predictor random variables will have its own distribution function. we can write then: where f denotes the marginal (joint) probability density function of the corresponding variable (variables) written as a subscript. the validity of (8) relies on two assumptions: (a) conditional independence of the groups given aggregation variables v 1 , .., v m and (b) conditional distribution function of the group x g conditioned on v 1 , ..., v m is the same as the conditional distribution of x g conditioned on v g (for a 2-dimensional proof see [14] ). these assumptions are in principle not problematic in our application because we are assuming that all interactions on micro-level of the agents are sufficiently well captured by the distribution of aggregation variables. hence formula (8) should be viewed as a crucial means for simplification, because it enables a principled decomposition. expressing the density function of v in (8) using formula (6) as copula, we obtain: this formula provides us a way to integrate in the same model mechanisms associated to the different assets in the market by means of a copula at aggregate level. in other words, by this formula, it is possible to calculate the probability of rare events, and therefore estimate systematic risk, based on the dependencies of aggregation variables and on the knowledge of micro-behaviour specified by group density functions of the agent-based models. the marginal distribution functions f vi (v i ) can be estimated either from real world data (e.g. asset price time series), or from simulations. note that whether we estimate from real world data or an agent-based market model should not matter in principle, since, if well constructed, the agent-based model should generate the same distribution of δp as the distribution estimated from real world data. the density function f xg (x g ) of an individual random vector x g can be defined as we did in our simulation study. however, to bring this approach into practice, three problems remain be investigated: estimation of copula. we need to consider possible structural time dependencies and serial dependencies in the individual aggregation variables. additionally, the agents might change their behavioural script (e.g. traders might pass from technical to fundamental at certain thresholds conditions). high dimensionality of eqs. (8) and (9) . if we consider n predictor variables for each group g = 1, ..., m we will end up with a n · m-dimensional density function. interpolation of function h g . calculating high-dimensional integrals that occur for instance in formula (8), with function h g being implicitly computed by the simulation of an agent-based market model, is clearly intractable. for the first problem, we observe that time dependence with respect to copulas is still an active area of research. most estimation methods do not allow for serial dependence of random variables. one approach to solve this is to filter the serial dependence by an autoregressive model as described at the beginning of sect. 2. another approach is to consider a dynamic copula, similarly as in [14, 20, 21] . a very interesting related work is presented in [22] , where arma-garch and arma-egarch are used to filter serial dependence, but the regime switching copula is considered on the basis of two-states markov models. using an agent-based model instead of (or integrated with) a markov model would be a very interesting research direction, because the change of regime would also have a qualitative interpretation. for the second problem, although in our example we have used 1000 agents, in general this might be not necessary, considering that abms might be not as heterogeneous, and aggregators might work with intermediate layers between micro and macro-levels. for the third problem, a better approach would be to interpolate the abm simulation by some function with closed form. in future works, we are going to evaluate the use of neural networks (nns), which means creating a model of our agent-based model, that is, a meta-model. the general concept of meta-models is a well-established design pattern [18] and the usage of nns for such purposes dates back to [19] . in our example the basic idea would be to obtain samples from the distribution f xg (x g ) as input, the results of an abm simulation v g as output, and then feed both input and output to train a dedicated nn, to be used at runtime. this would be done for each group g. the biggest advantage of this approach, if applicable in our case, is that we will have both a quick way to evaluate a function approximating h g , but we will also have the interpretative power of the agent-based market model, resulting in an overall powerful modelling architecture. agent-based models are a natural means to integrate expert (typically qualitative) knowledge, and directly support the interpretability of computational analysis. however, both the calibration on real-data and the model exploration phases cannot be conducted by symbolic means only. the paper sketched a framework integrating agent-based models with advanced quantitative probabilistic methods based on copula theory, which comes with a series of data-driven tools for dealing with dependencies. the framework has been illustrated with canonical asset pricing models, exploring dependencies at micro-and macrolevels, showing that it is indeed possible to capture quantitatively social characteristic of the systems. this also provided us with a novel view on market destabilization, usually explained in terms of strategy switching [24, 25] . second, the paper formally sketched a principled model decomposition, based on theoretical contributions presented in the literature. the ultimate goal of integrating agent-based models, advanced statistical methods (and possibly neural networks) is to obtain an unified model for risk evaluation, crucially centered around eq. (9) . clearly, additional theoretical challenges for such a result remains to be investigated, amongst which: (a) probabilistic models other than copulas to be related to the agent's decision mechanism, (b) structural changes of dependence structures, (c) potential causal mechanisms on the aggregation variables and related concepts as time dependencies (memory effects, hysteresis, etc.) and latency of responses. these directions, together with the development of a prototype testing the applicability of the approach, set our future research agenda. from micro to macro: demand, supply, and heterogeneity in the trade elasticity financial contagion: a propagation simulation mechanism when does a disaster become a systemic event? estimating indirect economic losses from natural disasters large scale extreme risk assessment using copulas: an application to drought events under climate change for austria integrating systemic risk and risk analysis using copulas incorporating contagion in portfolio credit risk models using network theory an agent-based model of an endangered population of the arctic fox from mednyi island cognitive maps and bayesian networks for knowledge representation and reasoning causal inference in statistics: an overview introduction to time series and forecasting. springer texts in statistics autoregressive conditional heteroscedasticity with estimates of the variance of united kingdom inflation volatility and time series econometrics: essays in honor of robert engle an introduction to copulas dependence modeling with copulas a review of copula models for economic time series copulas and its application in hydrology and water resources on the unstable behaviour of stock exchanges a metamodeling approach based on neural networks time-dependent copulas analysing financial contagion and asymmetric market dependence with volatility indices via copulas regime switching vine copula models for global equity and volatility indices why model? simple agent-based financial market model: direct interactions and comparisons of trading profits an agent-based simulation of stock market to analyze the influence of trader characteristics on financial market phenomena circlize implements and enhances circular visualization in r acknowledgments. the authors would like to thank drona kandhai for comments and discussions on preliminary versions of the paper. this work has been partly funded by the marie sklodowska-curie itn horizon 2020-funded project insights (call h2020-msca-itn-2017, grant agreement n.765710). key: cord-249836-s303s1tm authors: potter, lucas; ayala, orlando; palmer, xavier-lewis title: biocybersecurity -a converging threat as an auxiliary to war date: 2020-10-01 journal: nan doi: nan sha: doc_id: 249836 cord_uid: s303s1tm biodefense is the discipline of ensuring biosecurity with respect to select groups of organisms and limiting their spread. this field has increasingly been challenged by novel threats from nature that have been weaponized such as sars, anthrax, and similar pathogens, but has emerged victorious through collaboration of national and world health groups. however, it may come under additional stress in the 21st century as the field intersects with the cyberworld -a world where governments have already been struggling to keep up with cyber attacks from small to state-level actors as cyberthreats have been relied on to level the playing field in international disputes. disruptions to military logistics and economies through cyberattacks have been able to be done at a mere fraction of economic and moral costs through conventional military means, making it an increasingly tempting means of disruption. in the field of biocybersecurity (bcs), the strengths within biotechnology and cybersecurity merge, along with many of their vulnerabilities, and this could spell increased trouble for biodefense, as novel threats can be synthesized and disseminated in ways that fuse the routes of attacks seen in biosecurity and cybersecurity. herein, we offer an exploration of how threats in the domain of biocybersecurity may emerge through less foreseen routes as it might be an attractive auxiliary to conventional war. this is done through an analysis of potential payload and delivery methods to develop notional threat vectorizations. we conclude with several paradigms through which to view bcs-based threats. the increasing intersections between the cyberworld and biotechnology have introduced novel threats, of which most can be defined as bcs threats. these are threats at the nexus of biological, cyber, and physical apparatuses, allowing for malware to be distributed biologically or a pandemic-worthy agent via an autonomous drone guided over wi-fi networks. teams across the spectrum of hacking, from the individual to the state level are looking at novel ways to attack and defend against cyberthreats from multiple angles, leading to deeper discussions on how. with the ability to modify aspects of nature such as photosynthesis to improve on food production or energy storage ever-improving the importance of nefarious purposes is worth consideration [1] recently, many teams have brought these under-addressed possibilities to light, highlighting the many ways in which threats may emerge [2] [3] . through discussed vulnerabilities, exploits, and solutions differ, they point to a concerning issue: the morphing definition and basis of a payload that is transformed by this intersection. by payload, we refer to the functional aspect that causes damage to the systems that they are introduced to. within, we discuss types of payloads and explore what could be used as the next delivery mechanism for potential future payloads, and ways of framing the use of bioweapons in hostile contexts. a popularly used tool by penetration testers called metasploit lists three types of payloads for the description of threat delivery that we can use here, singles, stagers, and stages [4] . singles are self-contained; stagers link targets of interest and their attacker(s), and stages equip stagers. physical examples of this, in order, can be direct exposures to a pathogen like an injection or sneeze (single), poorly developed applications that unnecessarily group people (stagers), and poorly ventilated buildings, and poorly planned infrastructures that amplify or provide the threats (stages). adapting cybersecurity frameworks can give us a useful perspective on understanding intersectional threats. the conventional payloads of biological warfare are bacterial and viral loads. these loads are typically extant microbes that could then be genetically modified to either incapacitate in the case of tularemia or affect a large number of fatalities, (for instance, weaponized smallpox [5] [6] . the issue when it comes to deploying these agents as an auxiliary to warfare is, in one domain, well documented. conventional dispersal of anthrax, for instance, dates as far back to 1916 though it was deemed largely ineffective at the time [7] . the signing of the biological weapons convention has eliminated most uses of biological weapons from conventional warfare [8] . yet, the idea of using agents such as this for groups not bound by this treaty has never been greater, for the reasons elaborated below, including the lowered cost and increased accessibility to genetic modification, greater accessibility to targeting techniques (see delivery systems, below), centralized supply modalities (see in delivery systems), and an increasing number mechanisms by which to deliver a hypothetical biological agent (see threat vectorization) [9] . the question of each threat's damage comes down to the mode of the exposure of the payload to the intended target. is it intermittent like gusts down a hallway between bathroom breaks? is it stealthy like hiding on a pen? is it apparent, but unavoidable, like an elevator to an otherwise inaccessible penthouse? is it segmented in delivery like meal courses at a company cafeteria? these must be pondered in considering the means and mode. the many methods by which a biological agent can be delivered is an intriguing notion when compared to the conventional threat analysis. yet, the fundamental difference from the perspective of conventional threat analysis to threat analysis of bcs is that the delivery system for most conventional threats is built-for-purpose. the system of a bow delivering an arrow, or firearm delivering a bullet, or even a missile delivering a warhead, must be intricately designed to precise specifications for the delivery to be made. the delivery systems for biological threats can be entirely different. as seen in the covid-19 pandemic, sneezing, coughing, touching, and even talking transform benign surfaces into vectors [10] [11] . knowing this, the fundamental problem of delivering a biological payload with modifications to make it more robust then turns not to be what the physical method of delivery is but rather targeting. specifically, the problem is targeting people most likely to either spread or suffer from such an agent. this targeting can be surgical, involving one or a relatively few people, or it can be indiscriminate and en masse, for causing disruptions and indirect damage, making defense and containment difficult. a middle ground could be to identify genetic markers and utilize existing delivery systems to maximize cases of exposure. this could be useful for countering a condition for which many are asymptomatic but can be fatal in others. take for instance the following scenario: a terrorist organization has made a bacterium that is especially virulent in the presence of a diabetic comorbidity. assuming their goal is to maximize the number of casualties before any biosecurity groups can be alarmed, how might they best deliver their payload? some ideas come to mind such as finding diabetic treatment centers and infecting shared surfaces--doorknobs or waiting rooms. however, assume that they possess few resources to disseminate this agent and they want to minimize both their means of being tracked. this would require the minimal exposure of their members in spreading this contaminant. in this scenario, the possibilities become more limited, and they would need to find someone else to deliver their payload. in this case, the centralization of shipping and logistics under the umbrella of electronic distributors becomes interesting. it is no secret that online retailers use information about their customers for marketing purposes [12] [13] . if one makes an account on the site of an online retailer and browses items specific to people with diabetes (for instance test strips) they will quickly be shown hundreds of other items that people with diabetes may purchase. here, assume there is an item of clothing that is exceptionally popular with people with diabetes. the factory that supplies this item, which is likely not subject to biosafety regulations or familiar with the threat then becomes a perfect target. the distribution and targeting of the payloads become simple questions of infiltrating the supply chain. below are examples of everyday objects listed in a threat vectorization, combining which populations could match with which items, and so be targeted with a biological agent. it is useful to examine how various objects in our present society could be used. in figure 1 , we list some everyday objects to demonstrate how they could unknowingly be abused within a bcs context. these vectors are to be convenient delivery systems that fall in-between current threats. blood/organ transplants and foodstuffs are excluded as they are subject to a battery of tests and routinely monitored for biological agents. a: tufts of "fur" can trap dirt, dust, and host organisms at different scales. this can be amplified when they are wet or otherwise accumulate sweat and dead skin cells from kids who fail to frequently wash them. long amounts of skin to skin contact is possible, and this can be especially dangerous for immunocompromised (specifically young) individuals. if these are made into smart-toys, it is important that sensors embedded and data transfer capacities are scrutinized to eliminate unnecessary leaks regarding health information of users, especially children. b: toothbrushes are often wet and are rarely cleaned between uses, aside from rinsing. it is commonly guided for them to be simply disposed of after a few months. in terms of smart-products, these could be maliciously used to analyze personal data to identify optimal microbial threats in between uses, so caution is warranted with any sensors embedded and data collected. c: otoscopes or similar medical peripherals: devices like these are frequently used, and in times of fatigue, healthcare providers may fail to clean these, making for the possibility of pathogen transfer. residual matter is commonly left on surfaces when healthcare providers fail to wash hands or change gloves, and in addition to failure to properly sterilize the non-disposable surfaces, and merely discarding the disposable parts can miss pathogens that have transferred at the disposable and non-disposable intersections. they sometimes interact with other devices, in a wired or wireless capacity. in wired capacity or slotted capacities, those devices also require cleaning, especially when connecting with other devices. data hygiene should also be considered if there is data transfer and saving between smart devices, to minimize leaks of healthcare data. concerning otg (on the go) devices used in telehealth, there aforementioned risk of contamination can be multiplied for the patient wherein insufficient training is given on the use and cleaning of said device as well as the risk of overly sharing or holding sensitive health information in terms of pictures of numerical health data. d: smartphones are not often cleaned well and often taken to the bathroom, accumulating even more personal pathogens. it is common for these devices to touch and interact with a variety of surfaces such as laps, inner pocket linings, tables and desks, walls (when propped) and various holders, and more. further, they hold many personal records and have sensors that can give data about the holder such as temperature or heart rate. given the wide penetration of smartphones throughout the world, they are a ubiquitous and stealthy vector of compromise. long amounts of contact with face and hands (which allows indirect contact with face) can improve transmission. e: countertops and similar furnishings are exposed to the surfaces of many objects that are placed on it. their vector potential can be exacerbated by ridge patterns or carpet-like material, which require more dedicated cleanings. in sum, surfaces may contain harmful agents due to cross-contamination or poor sanitation, and pores in surfaces may present difficulties for cleaning. within bcs, designed agents can be spread and hidden on such surfaces as a means of attack or means of hiding data you can't slip a paper into a micro-sized pore, but you can slip bacteria with the contents of that paper in dna within that same pore. f: newspapers and similar magazines can not only host pathogens on their surfaces, but they can also disrupt airflow as their pages are turned due to their large surface area. when the user is alone, this is not problematic, but they need to be aware of surfaces that they have been exposed to such as when they transport these. in public, users need to be aware not to overly fan, to avoid fanning aerosols unknowingly across the public. more importantly, newspapers are never sanitized after leaving a central location and can interact with a wide number of people, resulting in the subtle, mass transfer of pathogens. above the idea of using modern supply chains and linking biological agents with cyber systems has been stated. now, an elaboration of how these systems can be used is in order: bcs as siege warfare, bcs, and compromised telehealth, bcs in surveillance warfare. in popular culture, an act of bioterrorism is frequently thought of as a dagger in the back -something, one barely knew was there before it was utilized. it is time to add another way of thinking about biological threats. ironically, biological warfare could see a return (especially if the current virus that has captured the attention of the world, covid-19, could be weaponized) to its original form in the idea of siege warfare. where once defenders flung infected masses over castle walls or by alternative methods, now the thought of using a biological contaminant to cause nations to shut down their borders, to cease internal traffic, and to occupy the skilled defenders has come full circle [14] [15] . the notion of a biological agent overwhelming healthcare facilities and followed by infrastructural damage enhancing casualties or risk, has shown its effectiveness even as an accidental occurrence [16] . this becomes an even more important idea as more targeted methods of biological warfare become more available. research into microfluidics, while nominally concerning itself with diagnosis, could be turned towards developing more virulent strains of biological warfare. take the following case as an example of how the notional concepts of siege warfare could be used for a biological threat. if a group was able to effectively modify a biological threat to quarantine a population, they could then use targeted biological data (for instance a microfluidics unit embedded in an electric toothbrush) to release secondary or alternate biological agents with a greater affinity for the quarantined population to incite panic via a potpourri of specific biological agents. variations of this could expand to different means of warfare which continue to occur with biology as a common link. we may see novel microbes designed and delivered to contaminate work sites that need to remain sterile, to degrade energy or food sources. for instance, if a power plant serving a quarantined population has a large number of people that would be susceptible to a specific disease then after the quarantine was in place the plant itself could be targeted with an alternate vector. this would be much more effective than either of these two actions separately. this concept can be used in parallel with crafted information warfare campaigns. we could see carefully designed information campaigns designed to expose us to or augment current threats. this can be done more efficiently through populations that lack high science literacy and possess significant distrust in scientific institutions. as noted by the philosopher amitabha palmer, anti-science propagandists can prey on trust and our reliance on others for information, which creates a ripe atmosphere for sowing discord with well-resourced campaigns [17] . recent happenings within information warfare, concerning covid-19, serve as a clear example of where this risk is and can be further elevated. wherein, foreign bots are encouraging individuals within america to re-open their economy and perhaps remove restrictions, hampering efforts to conduct an extended lock-down [18] . this can lead to an increase in infections, hampering not only the health of the populace but that of the economy, which of course affects a nation's military. bcs creates an indirect means of attack that cannot be understated without consequence. that is to say that such warfare is already here, intended or not, and it is up to us to consider how we may address it. this might be thought of as the only payload that is controllable as programs for informational dispersal can be shut off and on as needed, but once the material is out of the hands of the user or suffers some narrative disruptions, conspiracies can persist [19] . a question that merits asking those that consider deploying any biocyber threat is if they can justify or live with the threat persisting and eventually finding its way back to the country of origin. the payload can also be thought of digitally in the vein of telehealth, wherein patients use their phones or other computing telecommunication devices to consult with health professionals. major points of vulnerability lie in the hardware used, but also infections that could confound the telehealth technologies used. for example, compromised sensors, compromised software, or the combination thief could allow compromised healthcare readings and outcomes. deep learning enhanced malware was demonstrated to be effective in tricking radiologists on reading health outcomes for patients receiving a ct-scans, but this was for larger machines [20] . who is to say that such could not eventually occur on smartphones and otg accessories like otoscopes and auxiliary thermometers as machine learning becomes easier to adapt? in terms of sexual health, danger could apply to personal devices that could collect reproductive information, which additionally produces questions of care for marginalized sexual communities. further down the line, within the intersection of synthetic biology and cybersecurity, it is not unimaginable that state-level actors could create designer viruses that attack the dermis or epithelial tissues, producing difficult to detect qr-code analogs seen in the systems of birds, which indicate sex, hunger, and health qualities [21] [22] [23] . these could have the capability of compromising software designed to address and identify what a visual sensor picks up, especially that of software that uses system cameras similar to that of qr code scanning applications. one fix is through saving some aspects of care for in-person treatment, but a complication that may arise can also exist through exploiting vulnerabilities in means of indirect treatment through robots, as trailed in select cases [24] [25] . the threat of in-person and telehealth sabotage remains and may grow in delivered medical goods, especially for future innovations in the development of new electronic prosthetics, the restoration of function in the form of electronic skin that can transmit sensation, and the beginning of research that may transmit taste [26] [27] [28] . we may see novel, but general attacks for the populace that eventually reach healthcare works with their prosthesis, confounding their ability to adequately assess a patient. one attack could be the equivalent of a digital novocain or reflexive jerk, either of which could harm a patient, indirectly, through the provider's inability to feel for the use of their instrument on the patient, or directly through physical damage caused by involuntary movements. herein is a scenario in which patients and/or healthcare workers are harmed and resources are wasted. thus, concerns about telehealth at the intersection of bcs cannot be written off. the earlier mention of virus enabled skin modification addresses the intersection of enhanced surveillance. through propagation of computationally optimized viruses that label individuals, such technology could enhance the detection and hunt for targets that might otherwise escape. one can imagine patterns caused on one's skin created by a virus that is difficult to see human eyes but is instead easily viewable and readable by special cameras. this code can morph with and report changes in one's metabolism, detailing drug use, people contacted, or beyond, and be generated on a commonly exposed area of skin. these might one day be transmittable, like a sexually transmitted disease. we illustrate the notion of reproducing tattoos in figure 2 below. pictures that capture exposed skin with such code would allow for their surveillance as an individual is eventually to end up a mode of digitally accessible recording through social media. photos can be geotagged, but this tag can be removed. however, subtle patterns in the skin of an individual would be harder to remove and plan for its removal in the same way as geotagging is watched for due to the lack of knowledge that one is aware that someone in their picture is tagged in such a surreptitious fashion. often when an individual takes a selfie or picture within metropolitan areas, people in the background may be captured. early on, this was not as much of a problem, owing to no or just low-resolution photography being available with the introduction of early cell phones. however, smartphone camera technology has been improving, allowing for the equivalent of a worldwide panopticon, especially given that cell phones of improving caliber are increasingly penetrating developing markets. one could be tracked in everyday photo documentation on social media without even knocking about it, around the world, due to accidental photobombing. in military applications, a target of interest could have a device allowing tracking on a level with livestock as with rfids, qr codes, and or similar iot technologies, but in a more, undetectable manner that has fused with your biology [29] [30] . in this respect, resources allocated to track could be severely reduced, without sacrificing quality. careful checks would be needed to curtail abuses such as ethics and security panels, improved public science literacy, and advocates sponsoring technology that counters the unauthorized and nonconsensual use of such. groups working on technology to counter such is a given, and much like with the cold war this could be one front of the bio-surveillance arms race. further, emergent problems will need close inspections, as leaks and crosstalk occur, due to the increase in bio-insecurities produced, from this and the generation of information itself [31] . in terms of products to come out of this domain, a helpful means of public reassurance can come out of iot capacity labeling as demonstrated by carnegie mellon's cylab; with such products can be labeled for their surveillance capacity [31] . disagreement is had over the nature to which bio-enabled surveillance and digitalization of biological aspects may be had, but what's undeniable is that institutions will need to reexamine the insecurities increased through the data generation and through laxes in best practices [31, 33] . actors, no matter their level, must tread lightly in this domain. cybersecurity as a means of conducting auxiliary actions to conventional warfare is a field with impressive depth and breadth-there are entire journals dedicated to cyber warfare such as the cyber defense review. the exact protocols by which cybersecurity can be used adjacent to conventional military operations is too large of an issue to be constrained in this single article, though its importance to future defense operations is not to be understated. simply detailing the different vectors by which one may obtain confidential information like passcodes could fill an entire text. this is an exercise for the reader, who can search the key phrase "cybersecurity textbooks" to find a plethora of writings. searching "biocybersecurity textbook" will (at time of publication) likely return no or very few results. however, it should be noted that a similar analysis of what is done here (particularly in figure 1 ) could be done for the cybersecurity domain. the authors actively encourage such research and would be curious to note what similarities, differences, analogies, and the possible new threats that could emerge from such an analysis. in the world of bcs, and the greater universe of threat analysis, there are rarely simple convenient solutions. engineering rarely gives a clear-cut solution to any given problem, and instead leaves one with several piecemeal parts that solve added parts of the problem until the issue itself is negligible. an unsatisfying solution to threats mounted against the bio-cyber nexus is simply to end the cyber domain. but ask one of your colleagues if they could do their job without access to networked computers and you will find this solution untenable (and you would, likely and rightly, be mocked). the threat of bcs is here to stay. conventional solutions are the same ones we use for agricultural security. regular monitoring, dedicated personnel, and robust legislation to protect the world's citizens from threats that they cannot control. one never truly considers that in the developed world, there are legions of people and libraries of regulations that control how their edible goods are gathered, processed, treated, and transported [34] . the primary issue is identifying which items should be compelled to be inspected. this would require a constant surveillance bioweapons program focused on longevity and robustness of bioweapons, constantly attempting to find the long-term viability of bioweapons on goods given the contemporary logistics network conditions. to conclude, there is a need to build upon our imagination as to what threat will emerge within the growing field of cybersecurity. threats can emerge through the most mundane encounters and can unintentionally be delivered or be directed to. the question that remains is what each country will be doing to predict these routes, prevent, and counter them. failure of a country to produce a reliable answer may spell their decline. cyber-spinach turns sunlight into sugar cyberbiosecurity: an emerging new discipline to help safeguard the bioeconomy cyberbiosecurity: from naive trust to risk awareness an analysis of the ids penetration tool: metasploit. the infosec writers text library epidemiology of tularemia review: the risk of contracting anthrax from spore-contaminated soil -a military medical perspective current trends in the modern trends in plant genome editing: an inclusive review of the crispr/cas9 toolbox coronavirus disease 2019 (covid-19) and pregnancy: what obstetricians need to know talking can generate coronavirus droplets that linger up to 14 minutes technocreep: the surrender of privacy and the capitalization of intimacy the use of observational technology to study in-store behavior: consumer choice, video surveillance, and retail analytics biological warfare plan in the 17th century-the siege of candia biological warfare at the 1346 siege of caffa patients with cancer hit hard by deadly explosions in beirut. the lancet oncology how and why anti-science propaganda works nearly half of the twitter accounts discussing 'reopening america' may be bots an automated pipeline for the discovery of conspiracy and conspiracy theory narrative frameworks: bridgegate, pizzagate and storytelling on the web ct-gan: malicious tampering of 3d medical imagery using deep learning house sparrow, passer domesticus, parents preferentially feed nestlings with mouth colours that appear carotenoid-rich sex difference in mouth coloration and begging calls of barn swallow nestlings mouth coloration of nestlings covaries with offspring quality and influences parental feeding behavior rwanda has enlisted anti-epidemic robots in its fight against coronavirus a biomimetic eye with a hemispherical perovskite nanowire array retina electronic skin and electrocutaneous stimulation to restore the sense of touch in hand prosthetics electronic skin: recent progress and future prospects for skin-attachable devices for health monitoring, robotics, and prosthetics norimaki synthesizer: taste display using ion electrophoresis in five gels safety traceability system of livestock and poultry industrial chain from human body digitization to internet of bodies toward a new dimension of military operations bioveillance: a techno-security infrastructure to preempt the dangers of informationalised biology iot labels will help consumers figure out which devices are spying on them computer security, privacy, and {dna} sequencing: compromising computers with synthesized {dna}, privacy leaks, and more. retrieved fearing food: the influence of risk perceptions on public preferences for uniform and centralized risk regulation key: cord-214774-yro1iw80 authors: srivastava, anuj title: agent-level pandemic simulation (alps) for analyzing effects of lockdown measures date: 2020-04-25 journal: nan doi: nan sha: doc_id: 214774 cord_uid: yro1iw80 this paper develops an agent-level simulation model, termed alps, for simulating the spread of an infectious disease in a confined community. the mechanism of transmission is agent-to-agent contact, using parameters reported for corona covid-19 pandemic. the main goal of the alps simulation is analyze effects of preventive measures -imposition and lifting of lockdown norms -on the rates of infections, fatalities and recoveries. the model assumptions and choices represent a balance between competing demands of being realistic and being efficient for real-time inferences. the model provides quantification of gains in reducing casualties by imposition and maintenance of restrictive measures in place. there is a great interest in statistical modeling and analysis of medical, economical and epidemiological data resulting from current the covid-19 pandemic. from an epidemiological perspective, as large amount of infection, containment, and recovery data from the this pandemic becomes available over time, the community is currently relying essentially on simulation models to help assess situations and to evaluate options [1] . naturally, simulation systems that follow precise mathematical and statistical models are playing an important role in understanding this dynamic and complex situation [2] . in this paper we develop a mathematical simulation model, termed alps, to replicate the spread of an infectious disease, such as covid-19, in a confined community and to study the influence of some governmental interventions on final outcomes. since alps is purely a simulation model, the underlying assumptions and choices of statistical distributions for random quantities become critical in its success. on one hand it is important to capture the intricacies the observed phenomena as closely as possible, using sophisticated modeling tools. on the other hand, it is important to keep the model efficient and tractable by using simplifying assumptions. one can, of course, relax these assumptions and obtain more and more realistic models as desired, but at the cost of increasing computational complexity. there have been a large models proposed in the literatures relating to the the spread of epidemics through human contacts or otherwise. they can be broadly categorized in two main classes (a more detailed taxonomy of simulation models can be found in [3] ): 1. population-level coarse modeling: a large number of epidemiological models have focused on the population-level variables -counts of infected (i), 0 1/15 susceptible (s), removed or recovered (r), etc. the most popular model of this type is the susceptible-infected-removed (sir) model [4] proposed by kermack and mckendrick in 1927 . this model uses ordinary differential equations to model constrained growth of the counts in these three categories: ds dt = −βi(t)s(t), di dt = βi(t)s(t) − γi(t), dr dt = γi(t) . the two parameters β and γ control the dynamics of infections, and the condition ds dt + di dt + dr dt = 0 ensures constancy of the community size. a number of other papers have studied variants of these models and have adapted them for different epidemics, such as ebola and sars [5] . while there are spatial versions of sir models, they are usually limited in their modeling of spatial dynamics. they typically use a uniform static grid to represent the spread of infections, from a location to its neighbors, over time. in general these models do not explicitly model people dynamics as residents move around in a community. several recent simulation models, focusing directly on covid-19 illness, also rely on such coarser community level models [6] . 2. agent-level modeling: while population-level dynamical evolutions of population variables are simple and very effective for overall assessment, they do not take into account any social dynamics, human behavior, government-mandated restrictions, and complexities of human interactions explicitly. the models that study these human-level factors and variables , while tracking disease at an individual level, are called agent-level models [7] . here one models the mobility, health status, and interactions of individual subjects (agents) in order to construct an overall population-level picture in a bottom-up way. the advantages of agent-based models are that they are more detailed and one can vary the parameters of restriction measures, such as social distancing, at a granular level to infer overall outcomes. agent-based models have been discussed in several papers, including [3, [8] [9] [10] and so on. the importance of simulations based analysis of epidemic spread is emphasized in [11] but with a focus on infection models within host. some aspect of spreading of diseases using network contact is also discussed. hunter et al. [10] construct a detailed agent-based model for spread of infectious diseases, taking into account population demographics and other social conditions, but they do not consider countermeasures such as lockdowns in their simulations. a broad organization of different agent-based simulation methods have been presented in [3] . there are numerous other papers on the topic of agent-based simulations for simulating spread of infections that are not referenced. the main distinction of the current paper from the past literature is its focus on agent-level transmission of infections, and the influence of social dynamics and lockdown-type restrictions on these transmissions. in this paper we assume a closed community with the infection started by a single agent at initial time. the infections are transmitted through physical exposure of susceptible agents to the infected agents. the infected agents go through a period of sickness with two eventual outcomes -full recovery for most and death for a small fraction. once recovered, the agents can no longer be infected. the social dynamical model used here is based on fixed domicile, i.e. each agent has a fixed housing unit. under unrestricted conditions, or no lockdown, the agents are free to move over the full domain using a simple motion model. these motions are independent across agents and encourage smooth paths. under lockdown conditions, most of the agents head directly to their housing units and generally stay there during that period. a very small fraction of agents are allowed to move freely under the restrictions. the rest of this paper is organized as follows. section 2 develops the proposed agent-level pandemic simulation (alps) model, specifying the underlying assumptions and motivating model choices. it also discusses choices of model parameters and present a validation using comparisons with the sir model. section 3 presents some illustrative examples and discussed computational complexity of alps. the use of alps in understanding influences of countermeasures is presented in section 4. the paper ends by discussing model limitations and suggesting some future directions. in this section we develop our simulation model for agent-level interactions and spread of the infections across a population in a well-defined geographical domain. in terms of the model design, there are competing requirements for such a simulation to be useful. our main considerations in the design of alps are as follows. on one hand, we want to capture detailed properties of agents and their pertinent environments so as to render a realistic model of pandemic evolution with or without countermeasures. on the other hand, we want to keep model complexity reasonably low, in order to utilize it for analysis under variable conditions and countermeasures. also, to obtain statistical summaries of pandemic conditions under different scenarios, we want to run a large number of simulations and compute averages. this also requires keeping the overall model tractable from a computational perspective to allow for multiple runs of alps in short time. the overall setting of the simulation model is the following. we assuming that the community is based in a square geographical region d with h household units arranged in a uniformly-spaced square grid. we assume that there are n total agents in the community and the configuration updates every unit interval (hour) counted by the variable t. the agents are fully mobile to traverse all of d when unconstrained but are largely restricted to their home units under restrictions. next, we specify the assumptions/models being used at the moment. • independent agents: each agent (person) has an independent motion model and independent infection probability. the actual infection event is of course dependent on being in a close proximity of an infected carrier (within a certain distance, say ≈ 6 feet) for a certain exposure time. but the probability of an agent being infected is independent of such events for other persons. • full mobility in absence of restrictive measures: we assume that each agent is fully mobile and moves across the domain freely when no restrictive measures are imposed. in other words, there no agents with restricted mobility due to age or health. also, we do not impose any day/night schedules on the motions, we simply follow a fixed motion model at all times. some papers, including [12] , provide a two-or three-state models where the agents transition between some stable states (home, workplace, shopping, etc) in a predetermined manner. • homestay during restrictive measures: we assume that most agents stay at home at all times during the restrictive conditions. only a small percentage (set as a parameter ρ 0 ) of population is allowed to move freely but the large majority stays at home. • sealed region boundaries: in order to avoid complications of introducing a transportation model in the system, we assume that there is no transfer of agents into and out of the region d. the region is modeled to have reflecting boundaries to ensure that all the citizens stay with in the region. the only way the population of d is changed is through death of agent(s). • fixed domicile: the whole community is divided into a certain number of living units (households/buildings). these units are placed in square blocks with uniform spacing. each agent has a fixed domicile at one of the units. during a lockdown period, the agents proceed to and stay at home with a certain fidelity. we assume that all agents within a unit are exposed to each other, i.e. they are in close proximity, and can potentially infect others. • no re-infection: we assume that once a person has recovered from the diseased then he/she can not be infected again for the remaining observation period. while this is an important unresolved issue for the current covid-19 infections, it has been a valid assumption for the past corona virus infections. • single patient ground zero: the infection is introduced in the population using a single carrier, termed patient ground zero at time t = 0. this patient is selected randomly from the population and the time variable is indicated related to this introduction event. • constant immunity level: the probability of infection of agents, under the exposure conditions, remains same over time. we do not assume any increase or decrease in agent immunity over time. also, we do not assign any age or ethnicity to the agents and all agents are assumed to have equal immunity levels. there are several parts of the model that require individual specifications. these parts include a model on dynamics of individual agents (with or without restrictions in place), the mechanisms of transmitting infections from agent to agents, and the process of recovery and fatality for infected agents. a full listing of the model parameters and some typical values are given in table 1 in the appendix. • motion model: the movement of a subject follows a simple model where the instantaneous velocity v i (t) is a weighted sum of three components: (1) velocity at the previous time, i.e. v i (t − 1), (2) a directed component guiding them to their home, α(h i − x i (t − 1)), and (3) an independent gaussian increment σw i (t), w i (t) ∼ n (0, 1). note that the motions of different agents are kept independent of each other. the located h i ∈ d denotes the home unit (or stable state) of the i th person. using mathematical notation, the model for instantaneous position x i (t) and velocity v i (t) of the i th agent are given by: here α ∈ r + determines how fast one moves towards their home, and µ quantifies the degree to which one follows the directive to stay home. if µ = 0, then the person reaches home and stays there except for the random component w i . however, if µ = 0.5, then a significant fraction of motion represents continuity irrespective of the home location. the value µ = 1 implies that either there is not restriction in place or the person does not follow the directive. reflecting boundary: when a subject reaches boundary of the domain d, the motion is reflected and the motion continues in the opposite direction. fig. 1 shows examples of random agent motions under different simulation conditions. the leftmost case shows when there is no lockdown and the agents are moving freely throughout. the middle case shows the case when the restrictions are imposed on day 10 and the restrictions stay in place after that. the last plot shows the case where a lockdown is imposed on day 10 and then lifted on day 20. • restrictions (or lockdown) model: once the restriction period starts, at time t 0 , all agents are directed towards their homes and asked to stay there. we assume that ρ% of the subject follow this directive while the others ((100 − ρ)%) follow a different motion model. the variable ρ changes over time according to: we note that the people who do not follow restrictions follow the prescribed motion model with µ = 1. • exposure-infection model: the event of infection of an agent depends on the level of exposure to another infected person. this process is controlled by the following parameters: -the distance between the subject and the infected person should be less than r 0 . -the amount of exposure in terms of the number of time units should be at least τ 0 . at the moment we use the cumulative exposure over the whole history, rather than just the recent history. -given that these conditions are satisfied, the probability of catching the disease at each time t is an independent bernoulli random variable with probability p i . • recovery-death model: once a subject is infected, we randomly associate the nature of the illness immediately. either an infected agent is going to recover (non-fatal type, or nft) or the person is eventually going to die (fatal type, or ft). the probability of having a fatal infection given that a person is infection is -recovery: a subject with a non-fatal type (nft) is sick for a period of t r days. after this period, the person can recover at any time according a bernoulli random variable with probability p r . -fatality: a subject with a fatal type (ft) is sick for a period of t d days. after this period, the person can die at any time independently according a bernoulli random variable with probability p d . a complete listing of the alps model parameters is provided in the appendix in table 1 . in this section, we motivate the values chosen for those parameters in these simulations. these values are motivated by the current reports of corona virus pandemic. • we have used a square domain with size 2 miles × 2 miles for a community with population of n agents. for n = 900, the model represents a population density of 225 people/mile 2 . the community contains h living units (buildings) with a domicile of n/h people per unit. in case n/h is high, a unit represents a tall building in metropolitan areas, but when n/h is small, a unit represents a single family home in a suburban area. the time unit for updating configurations is one hour and occurrence of major events is specified in days. for example, the lockdown can start on day 1 and end on day 60. • the standard deviation for accelerations in agent mobility are approximately 1-5 feet/hour (fph). through integration over time, this results in agent speeds up to 1000 fph. we assume that ρ 0 = 0.98, i.e. 98% of the people follow the restriction directives. • the physical distance between agents to catch infection should at most r 0 ≈ 6f t and the exposure time should be at least τ 0 = 5 time units (hours). the probability p i of getting infected, under the right exposure conditions, is set at 5% at each time unit (hour) independently. • once infected, the probability of having a fatal outcome is set at 5-10%. the time period of recovery for agents with non-fatal outcomes starts at 7 days. the probability of reaching full recovery for those agents is p r = 0.001 at each subsequence time unit (hour). similarly, for the agents with fatal outcomes, the period of being infected is set to be 7 days and after that the probability of death at each time unit (hour) is set to be p d = 0.1. although alps is perhaps too simple model to capture intricate dynamics of an active society, it does provide an efficient tool for analyzing effects of countermeasures during the spread of a pandemic. before it can be used in practice, there is an important need to validate it in some way. as described in [3] , there are several ways to validate a simulation model. one is to use real data (an observed census of infections over time) in a community to estimate model parameters, followed by a statistical model testing. while such data may emerge for covid-19 for public use in the future (especially with the advent of tracking apps being deployed in many countries), there is currently no such agent-level data available for covid-19. the other approach for validation is to consider coarse population-level variables and their dynamics, and compare them against established models such as sir and its variations. we take this latter approach. fig. 2 shows plots of the evolutions of global infection counts (susceptible, infection, recovered) in a community under the well-known sir model (on the left) and the proposed alps model (on the right). in the alps model the counts for recovered and fatalities are kept separate, while in sir model these two categories are combined. one can see a remarkable similarity in the shapes of corresponding curves and this provides a certain validation to the alps model. in fact, given the dynamical models of agent-level mobility and infections, one can in principle derive parameters of the population-level differential equations used in the sir model. we have left that out for future developments. we illustrate the use of alps model by showing its outcomes under a few typical scenarios. further, we discuss the computational cost of running alps on a regular laptop computer. we start by showing outputs of alps under some interesting settings. in these examples, we use a relatively small number of agents (n = 900) and household units (h = 9), with t = 100days, in order to improve visibility of displays. fig. 3 shows a sequence of temporal snapshots representing thecommunity at different times over the observation period. in this example, the population is fully mobile over the observation period and no social distancing restrictions are imposed. the snapshots show the situations at 10, 20, 50, and 100 days. the corresponding time evolution of global count measures is shown in the bottom right panel. the infection starts to spread rapidly around the 10th day and reaches a peak infection level of 81% around day 35. then the recovery starts and continues until very few infected people are left. in this simulation, the number of fatalities is found to be 11%. in the second example, the restrictions are introduced on day 5 after the infection and these measures stay in place after that. the results are shown in fig. 4 . the bottom right panel shows temporal evolutions of the population-level infection counts: susceptible (blue curve), infected (red), recovered (green), and deceased (magenta). as the picture shows, the infections start growing initially but the gains of lockdown measures start appearing around day 15 -it takes about 10-12 for the restrictions to show results. the subsequent bumps in the infected counts is due to the new infections transmitted by roaming agents. in this run of alps, we see an overall fatality rate of 3% and an uninfected population size of 67%. in this example, with results shown in fig. 5 , the restrictions are introduced on day 5 after the infection and lifted on day 30. so the restriction period of 25 days is surrounded by unrestricted mobility on both sides. as the plot of global variables indicate, the early restriction helps reduce the infection rates but these gains are lost soon after lifting of the restrictions. the percentage of infected people goes back up and the rate of fatalities resemble the unrestricted situation in example 1. since computational efficiency of the simulator is of vital important, we next study the computational cost of running alps for different variable sizes. in these experiments, from these results we are see that the computational cost is linear in t with slope 1. for a change in n the number of agents, while keeping other variables fixed, the change in the computational cost is linear also. however, the rate of change is 2 for smaller values but increases to 4 for the larger values. the changes in computational cost due to changes in the number of households, with other variables held fixed, are minimal. there are several ways to utilize this model for prediction, planning and decision making. we illustrate some of these ideas using examples. at first we show individual simulations under different scenarios and then present results on average behavior obtained using hundreds of simulations. in these illustrations, we have used n = 972 agents and h = 81 households. effect of timing of imposition of restrictions: fig. 6 shows some examples of alps outputs when a lockdown is imposed on the community but at different times. from top-left to bottom-right, the plots show lockdowns starting on day 1, day 5, day 10, and day 20, respectively. once the restrictions are imposed, they are not removed in these examples. the best results are obtained for the earliest imposition of restrictions. in the top-left case, the peak infection rises to 22% of the community, on day 18, and then comes down steadily. the fraction of fatalities is 2% and the fraction of community never infected is 77%. in case the restrictions are imposed on day 5, and all other parameters held same, there is a small change in the situation. the peak infection rises to 55%, the fatalities increase to 7% and the fraction of uninfected goes down to 40%. we can see that an early imposition of lockdown measures also helps reduce peak infection rates in the community. sometimes we notice a saw-tooth shape in the curve for infected people. this implies that the even the small portion of mobile agents can break through and spread infections at other home units despite full restrictions being in place. this saw-tooth shape underscores the need for severely limiting mobility. even a small fraction of population being mobile can spread infections to the immobile agents and cause infections. the bottom two panels in fig. 6 show results for a delayed lockdown, with the restrictions being imposed on day 20 and day 30. one can see that the peak infection rate becomes quite high (82-84%) and casualties mount to 10-11%. the fraction of uninfected population falls to 0% in these cases. this shows that, under the chosen parameter settings, day 20 is quite late in imposing lockdown conditions on the community, and the results are very similar to any later imposition. if the restrictions are imposed after 15-20 days, then there are no uninfected people left in the community in a typical run of alps. effect of timing of removal of restrictions: in the next set of simulations, we study the effects of lifting restrictions and thus re-allowing full mobility in the community. some sample results are shown in fig. 7 . each plot shows the evolution for a different starting time t 0 and the end time fig. 8 shows an example of the configuration with 25 households and 200 agents in the community. we vary the start time t 0 (start day of restrictions) from 1 to 30 and then to 150 and study the resulting outcomes using 100 runes of alps. (the value of t 0 = 150 implies that the restrictions are never imposed in that setting.) the remaining panels in fig. 8 show box plots of the three variables changing with t 0 . • death rate: top right shows the percentage fatalities increasing from around 1% to almost 11% as t 0 changes from 1 to 30. the largest rate of increase is observed when t 0 is between 10 to 30 days. • number of uninfected: in the bottom left panel we see a decrease in the number of uninfected population from around 90% to 0% as t 0 increases. in case the restrictions are not imposed in first 30 days after the first infection, there is no agent left uninfected in the community. • peak infection rate: in case the restrictions are imposed on the first day after the infection, the peak infection rate is contained to be 10%. as t 0 is increased and the restrictions are delayed the peak infection rate rises to almost 80% of the community. in fig. 9 we study the effect of changing t 1 while t 0 = 1 is kept fixed (and other experimental conditions being same as in the last experiment). the results show that there is no difference in the eventual number of deaths and the peak infection rates when t 1 is changed from 10 to 40. this is because agent immunity and other infection the strengths and limitations of alps model are the following. alps provides an efficient yet comprehensive modeling of the spread of infections in a self-contained community, using simple model assumptions. the model can prove very useful in evaluating costs and effects of imposing social lockdown measures in a society. in the current version, the initial placement of agents is set to be normally distributed with means given by their home units and fixed variance. this variance is kept large to allow for near arbitrary placements of agents in the community. in practice, however, agents typically follow semi-rigid daily schedules of being at work, performing chores, or being at home. thus, at the time of imposition of a lockdown, the agents can be better placed in the scenes according to their regular schedules rather than being placed arbitrarily. in terms of future directions, there are many ways to develop this simulation model to capture more realistic scenarios: (1) it is possible to model multiple, interactive communities instead of a single isolated community. (2) one can include typical daily schedules for agents in the simulations. a typical agent may leave home in the morning, spent time in the office during the day, and return to home in the evening. (3) it is possible to provide an age demographics to the community and assign immunity to agents according to their demographic labels [13] . (4) as more data becomes available in the future, one can change immunity levels of agents over time according to the spread and seasons. (5) in practice, when an agent is infected, he/she goes through different stages of the disease, associated with varying degrees of mobility [9] . one can introduce an additional variable to track these stages of infections in the model and change agent mobility accordingly. this paper develops an agent-based simulation model, called alps, for modeling spread of an infectious disease in a closed community. a number of simplifying and reasonable assumptions makes the alps efficient and effective for statistical analysis. the model is validated at a population level by comparing with the popular sir model in epidemiology. these results indicate that: (1) early imposition of lockdown measures (right after the first infection) significantly reduce infection rates and fatalities; (2) 0 13/15 lifting of lockdown measures recommences the spread of the disease and the infections eventually reach the same level as the unrestricted community; (3) in absence of any extraneous solutions (a medical treatment/cure, a weakening mutation of the virus, or a natural development of agent immunity), the only viable option for preventing large infections is the judicious use of lockdown measures. special report: the simulations driving the world's response to covid-19. how epidemiologists rushed to model the coronavirus pandemic flute, a publicly available stochastic influenza epidemic simulation model a taxonomy for agent-based models in human infectious disease epidemiology a contribution to the mathematical theory of epidemics population-based simulations of influenza pandemics: validity and significance for public health policy estimates of the severity of coronavirus disease 2019: a model-based analysis agent-based models growing artificial societies. social science from the bottom up an agent-based approach for modeling dynamics of contagious disease spread. international journal of health geographics an open-data-driven agent-based model to simulate infectious disease outbreaks high-resolution epidemic simulation using within-host infection and contact data an agent-based spatially explicit epidemiological model in mason modelling transmission and control of the covid-19 pandemic in australia this research was supported in part by the grants nsf dms-1621787 and nsf dms-1953087. a listing of alps parameters table 1 provides a listing of all the parameters one can adjust in alps to achieve different scenarios. it also provides some typical values used in the experiments presented in the paper. key: cord-123804-cgvikrwm authors: liu, changliu title: a microscopic epidemic model and pandemic prediction using multi-agent reinforcement learning date: 2020-04-27 journal: nan doi: nan sha: doc_id: 123804 cord_uid: cgvikrwm this paper introduces a microscopic approach to model epidemics, which can explicitly consider the consequences of individual's decisions on the spread of the disease. we first formulate a microscopic multi-agent epidemic model where every agent can choose its activity level that affects the spread of the disease. then by minimizing agents' cost functions, we solve for the optimal decisions for individual agents in the framework of game theory and multi-agent reinforcement learning. given the optimal decisions of all agents, we can make predictions about the spread of the disease. we show that there are negative externalities in the sense that infected agents do not have enough incentives to protect others, which then necessitates external interventions to regulate agents' behaviors. in the discussion section, future directions are pointed out to make the model more realistic. with the covid-19 pandemic souring across the world, a reliable model is needed to describe the observed spread of the disease, make predictions about future, and guide public policy design to control the spread. existing epidemic models there are many existing macroscopic epidemic models. 3 for example, the si model describes the growth of 3 daryl j daley and joe gani. epidemic modelling: an introduction, volume 15. cambridge university press, 2001 infection rate as the product of the current infection rate and the current susceptible rate. the sir model further incorporates the effect of recovery into the model, i.e., when the infected population turns into immune population after a certain period of time. the sirs model considers the case that immunity is not for lifetime and that the immune population can become susceptible population again. in addition to these models, the seir model incorporates the incubation period into analysis. incubation period refers to the duration before symptoms show up. 4 the most important factor in all those models suppose there are m agents in the environment. initially, m 0 agents are infected. agents are indexed from 1 to m. every agent has its own state and control input. the model is in discrete time. the time interval is set to be one day. the evolution of the infection rate for consecutive days depends on agents' actions. the questions of interest are: how many agents will eventually be infected? how fast they will be infected? how can we slow down the growth of the infection rate? we consider two state values for an agent, e.g., for agent i, x i = 0 means healthy (susceptible), x i = 1 means infected. everyday, every agent i decides its level of activities u i ∈ [0, 1]. the level of activities for agent i can be understood as the expected percentage of other agents in the system that agent i wants to meet. for example, u i = 1/m means agent i expects to meet one other agent. the actual number of agents that agent i meets depends not only on agent i's activity level, but also on other agents' activity level. for example, if all microscopic epidemic model and marl 3 other agents choose an activity level 0, then agent i will not be able to meet any other agent no matter what u i it chooses. mathematically, the chance for agent i and agent j to meet each other depends on the minimum of the activity levels of these two agents, i.e., min{u i , u j }. in the extreme cases, if agent i decides to meet everyone in the system by choosing u i = 1, then the chance for agent j to meet with agent i is u j . if agent i decides to not meet anyone in the system by choosing u i = 0, then the chance for agent j to meet with agent i is 0. before we derive the system dynamic model, the assumptions are listed below: 6 6 these assumptions can all be relaxed in future work. they are introduced mainly for the simplicity of the discussion. 1. in the agent model, we only consider two states: healthy (susceptible) and infected. all healthy agents are susceptible to the disease. there is no recovery and no death for infected agents. there is no incubation period for infected agents, i.e., once infected, the agent can start to infect other healthy agents. to relax this assumption, we may introduce more states for every agent. 2. the interactions among agents are assumed to be uniform, although it is not true in the real world. in the real world, given a fixed activity level, agents are more likely to meet with close families, friends, colleagues than strangers on the street. to incorporate this non-uniformity into the model, we need to redefine the chance for agent i and agent j to meet each other to be β i,j min{u i , u j }, where β i,j ∈ [0, 1] is a coefficient that encodes the proximity between agent i and agent j and will affect the chance for them to meet with each other. for simplicity, we assume that the interaction patterns are uniform in this paper. 3. meeting with infected agents will result in immediate infection. to relax this assumption, we may introduce an infection probability to describe how likely it is for a healthy agent to be infected if it meets with an infected agent. on day k, denote agent i's state and control as x i,k ∈ x and u i,k ∈ u . by definition, the agent state space is x = {0, 1} and the agent control space is u = [0, 1]. the system state space is denoted x m := x × · · · × x . the system control space is denoted u m := u × · · · × u . define m k = ∑ i x i,k as the number of infected agents at time k. the set of infected agents is denoted: the state transition probability for the multi-agent system is a microscopic epidemic model and marl 4 according to the assumptions, an infected agent will always remain infected. hence the state transition probability for an infected agent i does not depend on other agents' states or any control. however, the state transition probability for a healthy agent i depends on others. the chance for a healthy agent i to not meet an infected agent j ∈ i k is 1 − min{u i , u j }. a healthy agent can stay healthy if and only if it does not meet any infected agent, the probability of which is π j∈i k (1 − min{u i , u j }). then the probability for a healthy agent to be infected is 1 − π j∈i k (1 − min{u i , u j }). from the expression π j∈i k (1 − min{u i , u j }), we can infer that: the chance for a healthy agent i to stay health is higher if • the agent i limits its own activity by choosing a smaller u i ; • the number of infected agents is smaller; • the infected agents in i k limit their activities. the state transition probability for an agent i is summarized in table 1. example consider a four-agent system shown in fig. 1 . only agent 1 is infected. and the agents choose the following activity levels: u 1 = 0.1, u 2 = 0.2, u 3 = 0.3, u 4 = 0.4. then the chance p i,j for agents i and j to meet with each other is p 1,2 = p 1,3 = p 1,4 = 0.1, p 2,3 = p 2,4 = 0.2, and p 3,4 = 0.3. note that p i,j = p j,i . the chance for agents 2, 3, and 4 to stay healthy is 0.9, although they have different activity levels. other agents are healthy. the numbers on the links denote the probability for agents to meet with each other, which depend on the chosen activity levels of different agents. before we start to derive the optimal strategies for individual agents and analyze the closed-loop multi-agent system, we first characterize the (open-loop) multi-agent system dynamics by monte carlo simulation according to the state transition probability in table 1. suppose we have m = 1000 agents. at the beginning, only agent 1 is infected. we consider two levels of activities: normal activity level u and reduced activity level u * . the two activity levels are assigned to different agents following different strategies as described below. in particular, we consider "no intervention" case where all agents microscopic epidemic model and marl 5 continue to follow the normal activity level, "immediate isolation" case where the activity levels of infected agents immediately drop to the reduced level, "delayed isolation" case where the activity levels of infected agents drop to the reduced level after several days, and "lockdown" case where the activity levels of all agents drop to the reduced level immediately. the vertical axis corresponds to agent id i. the color in the graph represents the value of x i,k , blue for 0 (healthy) and yellow for 1 (infected). for each case, we simulate 200 system trajectories and compute the average, maximum, and minimum m k (number of infected agents) versus k from all trajectories. a system trajectory in the "no intervention" case is illustrated in fig. 2 , where u = 1/m for all agents. the m k trajectories under different cases are shown in fig. 3 , where the solid curves illustrate the average m k and the shaded area corresponds to the range from min m k to max m k . the results are explained below. • case 0: no intervention. all agents keep the normal activity level u. the scenarios for u = 1/m and u = 2/m are illustrated in fig. 3 . as expected, a higher activity level for all agents will lead to faster infection. the trajectory of m k has a s shape, whose growth rate is relatively slow when either the infected population is small or the healthy population is small, and is maximized when 50% agents are infected. it will be shown in the following discussion that (empirical) macroscopic models also generate s-curves. • case 1: immediate isolation of infected agents. the activity levels of infected agents immediately drop to u * , while others remain u. the scenario for u = 1/m and u * = 0.1/m is illustrated in fig. 3 . immediate isolation significantly slows down the growth of the infections rate. as expected, it has the best performance in terms of flattening the curve, same as the lockdown case. the trajectory also has a s shape. • case 2: delayed isolation of infected agents. the activity levels of infected agents drop to u * after t days, while others remain u. in the simulation, u = 1/m and u * = 0.1/m. the scenarios for t = 1 and t = 2 are illustrated in fig. 3 . as expected, the longer the delay, the faster the infection rate grows, though the growth of the infection rate is still slower than the "no intervention" case. moreover, the peak growth rate (when 50% agents are infected) is higher when the delay is longer. • case 3: lockdown. the activity levels of all agents drop to u * . the scenario for u * = 0.1/m is illustrated in fig. 3 . as expected, it has the best perfor-microscopic epidemic model and marl 6 mance in terms of flattening the curve, same as the immediate isolation case. 7 7 in the case that infected population can be asymptomatic or have a long incubation period before they show any symptom, like what we observe for covid-19, immediate identification of infected person and then immediate isolation is not achievable. then lockdown is the only best way to control the spread of the disease in our model. since the epidemic model is monotone, every agent will eventually be infected as long as the probability to meet infected agents does not drop to zero. moreover, we have not discussed decision making by individual agents yet. the activity levels are just predefined in the simulation. we simulate the system trajectory under different infection coefficients as shown in fig. 4 . the trajectories also have s shapes, similar to the ones in the microscopic model. however, since this macroscopic si model is deterministic, there is no "uncertainty" range as shown in the microscopic model. the infection coefficient β depends on the agents' choices of activity levels. however, there is not an explicit relationship yet. it is better to directly use the microscopic model to analyze the consequences of individual agents' choices. this section tries to answer the following question: in the microscopic multi-agent epidemic model, what is the best control strategy for individual agents? to answer that, we need to first specify the knowledge and observation models as well as the cost (reward) functions for individual agents. then we will derive the optimal choices of agents in a distributed manner. the resulting system dynamics correspond to a nash equilibrium of the system. a knowledge and observation model for agent i includes two aspects: what does agent i know about itself, and what does agent i know about others? the knowledge about any agent j includes the dynamic function of agent j and the cost function of agent j. the observation corresponds to run-time measurements, i.e., the observation of any agent j includes the run-time state x j,k and the run-time control u j,k . in the following discussion, regarding the knowledge and observation model, we make the following assumptions: • an agent knows its own dynamics and cost function; • all agents are homogeneous in the sense that they share the same dynamics and cost functions. and agents know that all agents are homogeneous, hence they know others' dynamics and cost functions; 8 • at time k, agents can measure x j,k for all j. but they cannot measure u j,k until time k + 1. hence, the agents are playing a simultaneous game. they need to infer others' decisions when making their own decisions at any time k. we consider two conflicting interests for every agent: 9 9 the identification of these two conflicting interests is purely empirical. to build realistic cost functions, we need to either study the real world data or conduct human subject experiments. • limit the activity level to minimize the chance to get infected; • maintain a certain activity level for living. we define the run-time cost for agent i at time k as where x i,k+1 corresponds to the first interest, p(u i,k ) corresponds to the second interest, and α i > 0 adjusts the preference between the two interests. the function p(u) is assumed to be smooth. 10 due to 10 the function p(u) can be a decreasing function on [0, 1], meaning that the higher the activity level, the better. the function p(u) can also be a convex parabolic function on [0, 1] with the minimum attained at some u * , meaning that the activity level should be maintained around u * . our homogeneity assumption on agents, they should have identical preferences, i.e., α i = α for all i. agent i chooses its action at time k by minimizing the expected cumulative cost in the future: where γ ∈ [0, 1] is a discount factor. the objective function depends on all agents' current and future actions. it is difficult to directly obtain an analytical solution of (5). later we will use multi-agent reinforcement learning to obtain a numerical solution. in this section, to simplify the problem, we consider a single stage game 11 where the agents have zero discount of the future, i.e., γ = 0. 11 the formulation (5) corresponds to a repeated game as opposed to the single stage game. repeated games capture the idea that an agent will have to take into account the impact of its current action on the future actions of others. this impact is called the agent's reputation. the interaction is more complex in a repeated game than that in a single stage game. hence the objective function is reduced to which only depends on the current actions of agents. according to the state transition probability in table 1, the expected cost is nash equilibrium according to (7), the expect cost for an infected agent only depends on its own action. hence the optimal choice for an infected agent is u i,k =ū := arg min u p(u). then the optimal choice for a healthy agent satisfies: note that the term 1 − (1 − min{u,ū}) m k is positive and is increasing for u ∈ [0,ū] and then constant for u ∈ [ū, 1]. hence, the optimal solution for (9) should be smaller thanū = arg min u p(u). 12 then 12 if u ≥ū, then (9) becomes since ∂ ∂u |ū j(u) > 0, the optimal solution satisfies that u <ū with cost j(u) < j(ū). note that j(ū) equals to the smallest cost for the case u ≤ū. hence the optimal solution for (9) satisfies that u <ū. the objective in (9) can be simplified as 1 − (1 − u) m k + α i p(u). in summary, the optimal actions for both the infected and the healthy agents in the nash equilibrium can be compactly written as example consider the previous example with four agents shown in fig. 1 . define microscopic epidemic model and marl 9 which is a monotonically decreasing function as illustrated in fig. 5 . then the optimal actions in the nash equilibrium for this specific problem satisfy: solving for (12), for infected agents, u i,k = 1. for healthy agents, the choice also depends on α i as illustrated in fig. 6 . we have assumed that α i = α which is identical for all agents. we further assume that α < 2 such that the optimal solution for healthy agents should be u i,k = 0. the optimal actions and the corresponding costs for all agents are listed in table 2. in the nash equilibrium, no agent will meet each other, since all agents except agent 1 reduce their activity levels to zero. the actual cost (received at the next time step) equals to the expected cost (computed at the current time step). however, let us consider another situation where the infected agent chooses 0 activity level and all other healthy agents choose 1 activity level. the resulting costs are summarized in table 3. obviously, the overall cost is reduced in the new situation. however, this better situation cannot be attained spontaneously by the agents, due to externality of the system which will be explained below. actual l i,k 1 1 0 1+α exp(−1) 1+α exp(−1) 2,3,4 0 1 0 0 total 1 + α exp(−1) table 3 : list of the agent decisions and associated costs in a situation better than the nash equilibrium in the four-agent example. for a multi-agent system, define the system cost as a summation of the individual costs: the system cost in the nash equilibrium is denoted l * k , which corresponds to the evaluation of l k under agent actions specified in (10). on the other hand, the optimal system cost is defined as microscopic epidemic model and marl 10 the optimization problem (14) is solved in a centralized manner, which is different from how the nash equilibrium is obtained. to obtain the nash equilibrium, all agents are solving their own optimization problems independently. although their objective functions depend on other agents' actions, they are not jointly make the decisions, but only "infer" what others will do. by definition, l o k ≤ l * k . in the example above, l * k = 1 + 3α exp(−1) and l o k = 1 + α exp(−1). the difference l * k − l o k is called the loss of social welfare. in the epidemic model, the loss of social welfare is due to the fact that bad consequences (i.e., infecting others) are not penalized in the cost functions of the infected agents. those unpenalized consequences are called externality. there can be both positive externality and negative externality. under positive externality, agents are lacking motivations to do things that are good for the society. under negative externality, agents are lacking motivations to prevent things that are bad for the society. in the epidemic model, there are negative externality with infected agents. to improve social welfare, we need to "internalize" externality, i.e., add penalty for "spreading" the disease. now let us redefine agent i's run-time cost asl where q(·) is a monotonically increasing function. the last term x i,k q(u i,k ) does not affect healthy agents since x i,k = 0, but adds a penalty for infected agents if they choose large activity level. one candidate function for q(u) is 1 − (1 − u) m k . in the real world, such "cost shaping" using q can be achieved through social norms or government regulation. the expected cost becomes suppose the function q is well tuned such that the arg min u [α i p(u) + q(u)] = 0. then although the expected costs for infected agents are still independent from others, their decision is considerate to healthy agents. when the infected agents choose u = 0, then for healthy agents, the expected cost becomes α i p(u i,k ), meaning that they do not need to worry about getting infected. let us now compute the resulting nash equilibrium under the shaped costs using the previous example. example in the four-agent example, set q(u) = u. then arg min u [αp(u) + u] = 0. hence agent 1 will choose u 1,k = 0. for agents i = 2, 3, 4, they will choose u i,k = 1 since they are only minimizing p(u). the resulting costs are summarized in table 4. with the shaped costs, the microscopic epidemic model and marl 11 system enters into a better nash equilibrium which indeed aligns with the system optimum in (14). a few remarks: • cost shaping did not increase the overall cost for the multi-agent system. • the system optimum remains the same before and after cost shaping. • cost shaping helped agents to arrive at the system optimum without centralized optimization. we have shown how to compute the nash equilibrium of the multiagent epidemic model in a single stage. however, it is analytically intractable to compute the nash equilibrium when we consider repeated games (5). the complexity will further grow when the number of agents increases and when there are information asymmetry. nonetheless, we can apply multi-agent reinforcement learning 13 to 13 lucian buşoniu, robert babuška, and bart de schutter. multi-agent reinforcement learning: an overview. in innovations in multi-agent systems and applications-1, pages 183-221. springer, 2010 numerically compute the nash equilibrium. then the evolution of the pandemic can be predicted by simulating the system under the nash equilibrium. as evident from (10), the optimal action for agent i at time k is a function of x i,k and m k . hence we can define a q function (action value function) for agent i as according to the assumptions made in the observation model, all agents can observe m k at time k. for a single stage game, we have derived in (10) for repeated games (5), we can learn the q function using temporal different learning. at every time k, agent i chooses its action as microscopic epidemic model and marl 12 after taking the action u i,k , agent i observes x i,k+1 and m k+1 and receives the cost l i,k at time k + 1. then agent i updates its q function: where η is the learning gain and δ i,k is the temporal difference error. all agents can run the above algorithm to learn their q functions during the interaction with others. however, the algorithm introduced above has several problems: • exploration and limited rationality. there is no exploration in (18). indeed, q-learning is usually applied together with -greedy where with probability 1 − , the action u i,k is chosen to be the optimal action in (18), and with probability , the action is randomly chosen with a uniform distribution over the action space. the -greedy approach is introduced mainly from an algorithmic perspective to improve convergence of the learning process. when applied to the epidemic model, it has a unique societal implication. when agents are randomly choosing their behaviors, it represents the fact that agents have only limited rationality. hence in the learning process, we apply -greedy as a way to incorporate exploration for faster convergence as well as to take into account limited rationality of agents. • data efficiency and parameter sharing. keeping separated q functions for individual agents is not data efficient. an agent may not be able to collect enough samples to properly learn the desired q function. due to the homogeneity assumptions we made earlier about agents' cost functions, it is more data efficient to share the q function for all agents. its societal implication is that agents are sharing information and knowledge with each other. hence, we apply parameter sharing 14 as a way to 14 jayesh k gupta, maxim egorov, and mykel kochenderfer. cooperative multi-agent control using deep reinforcement learning. in international conference on autonomous agents and multiagent systems, pages 66-83. springer, 2017 improve data efficiency as well as to consider information sharing among agents during the learning process. 15 15 in a more complex situation where agents are not homogeneous, it is desired to have parameter sharing with a smaller group of agents, instead of parameter sharing will all agents. with the above modifications, the multi-agent q learning algorithm 16 is summarized below. 16 junling hu and michael p wellman. nash q-learning for general-sum stochastic games. journal of machine learning research, 4(nov): 2003 • for every time step k, agents choose their actions as: • at the next time step k + 1, agents observe the new states x i,k+1 and receive rewards l i,k for all i. then the q function is updated: example in this example, we consider m = 50 agents in the system. only one agent is infected in the beginning. the run-time cost is the same as in the example in the distributed optimal control section, i.e., l i,k = x i,k+1 + α exp( 1 u i,k −1 ) where α is chosen to be 1. for simplicity, the action space is discretized to be {0, 1/m, 10/m}, called as low, medium, and high. hence the q function can be stored as a 2 × m × 3 matrix. in the learning algorithm, the learning rate is set to η = 1. the exploration rate is set to decay in different episodes, i.e., = 0.5(1 − e/ max e) where e denotes the current episode and the maximum episode is max e = 200. the q function is initialized to be 10 for all entries. three different cases are considered. for each case, we illustrate the q function learned after 200 episodes as well as the system trajectories for episodes 10, 20, . . . , 200, blue for earlier episodes and red for later episodes. the results are shown in fig. 7 . • case 1: discount γ = 0 with runtime cost l i,k . with γ = 0, this case reduces to a single stage game as discussed in the distributed optimal control section. the result should align with the analytical nash equilibrium in (10). as shown in the left plot in fig. 7(a) , the optimal action for a healthy agent is always low (solid green), while the optimal action for an infected agent is always high (dashed magenta). the q values for infected agents do not depend on m k . the q values for healthy agents increase when m k increases if the activity level is not zero, due to the fact that: for a fixed activity level, the chance to get infected is higher when there are more infected agents in the system. all these results align with our previous theoretical analysis. moreover, as shown in the right plot in fig. 7(a) , the agents are learning to flatten the curve across different episodes. • case 2: discount γ = 0.5 with runtime cost l i,k . since the agents are now computing cumulative costs as in (5), the corresponding q values are higher than those in case 1. however, the optimal actions remain the same, low (solid green) for healthy agents, high (dashed magenta) for infected agents, as shown in the left plot in fig. 7(b) . the trends of the q curves also remain the same: the q values do not depend on m k for infected agents and for healthy agents whose activity levels are zero. however, as shown in the right plot in fig. 7(b) , the agents learned to flatten the curve faster than in case 1, mainly because healthy agents are more cautious (converge faster to low activity levels) when they start to consider cumulative costs. • case 3: discount γ = 0.5 with shaped runtime costl i,k in (15). the shaped cost changes the optimal actions for all agents as well as the resulting q values. as shown in the left plot in fig. 7(c) , the optimal action for an infected agent is low (dashed green), while that for a healthy agent is high (solid magenta) when m k is small and low (solid green) when m k is big. note that when m k is high, the healthy agents still prefer low activity level, though the optimal actions for infected agents are low. that is because: due to the randomization introduced in -greedy, there is still chance for infected agents to have medium or high activity levels. when m k is high, the healthy agents would rather limit their own activity levels to avoid the risk to meet with infected agents that are taking random actions. this result captures the fact that agents understand others may have limited rationality and prefer more conservative behaviors. we observe the same trends for the q curves as the previous two cases: the q values do not depend on m k for infected agents and for healthy agents whose activity levels are not zero. in terms of absolute values, the q values for infected agents are higher than those in case 2 due to the additional cost q(u) inl i,k . the q values for healthy agents are smaller than those in case 2 for medium and high activity levels, since the chance to get infected is smaller as infected agents now prefer low activity levels. the q values remain the same for healthy agents with zero activity levels. with shaped costs, the agents learned to flatten the curve even faster than in case 2, as shown in the right plot in fig. 7(c) , since the shaped cost encourages infected agents to lower their activity levels. agents vs humans the epidemic model can be used to analyze realworld societal problems. nonetheless, it is important to understand the differences between agents and humans. we can directly design and shape the cost function for agents, but not for humans. for agents, their behavior is predictable once we fully specify the problem (i.e., cost, dynamics, measurement, etc). hence we can optimize the design (i.e., the cost function) to get desired system trajectory. for humans, their behavior is not fully predictable due to limited rationality. we need to constantly modify the knowledge and observation model as well as the cost function to match the true human behavior. future work the proposed model is in its preliminary form. many future directions can be pursued. • relaxation of assumptions. we may add more agent states to consider recovery, incubation period, and death. we may consider the fact that the interaction patterns among agents are not uniform. we may consider a wide variety of agents who are not homogeneous. for example, health providers and equipment suppliers are key parts in fighting the disease. they should receive lower cost (higher reward) for maintaining or even expanding their activity levels than ordinary people. their services can then lead to higher recovery rate. in addition, we may relax the assumptions on agents' knowledge and observation models, to consider information asymmetry as well as partial observation. for example, agents cannot get immediate measurement whether they are infected or not, or how many agents are infected in the system. • realistic cost functions for agents. the cost functions for agents are currently hand-tuned. we may learn those cost functions from data through inverse reinforcement learning. those cost functions can vary for agents from different countries, different age groups, and different occupations. moreover, the cost functions carry important cultural, demographical, economical, and political information. a realistic cost function can help us understand why we observe significantly different outcomes of the pandemic around the world, as well as enable more realistic predictions into the future by fully considering those cultural, demographical, economical, and political factors. • incorporation of public policies. for now, the only external intervention we introduced is cost shaping. we may consider a wider range of public policies that can change the closed-loop system dynamics. for example, shut-down of transportation, isolation of infected agents, contact tracing, antibody testing, etc. • transient vs steady state system behaviors. we have focused on the steady state system behaviors in the nash equilibrium. however, as agents live in a highly dynamic world, it is not guaranteed that a nash equilibrium can always be attained. while agents are learning to deal with unforeseen situations, there are many interesting transient dynamics, some of which is captured in fig. 7, i.e., agents may learn to flatten the curve at different rates. methods to understand and predict transient dynamics may be developed in the future. • validation against real world historical data. to use the proposed model for prediction in the real world, we need to validate its fidelity again the historical data. the validation can be performed on the m k trajectories, i.e., for the same initial condition, the predicted m k trajectories should align with the ground truth m k trajectories. this paper introduced a microscopic multi-agent epidemic model, which explicitly considered the consequences of individual's decisions on the spread of the disease. in the model, every agent can choose its activity level to minimize its cost function consisting of two conflicting components: staying healthy by limiting activities and maintaining high activity levels for living. we solved for the optimal decisions for individual agents in the framework of game theory and multi-agent reinforcement learning. given the optimal decisions of all agents, we can make predictions about the spread of the disease. the system had negative externality in the sense that infected agents did not have enough incentives to protect others, which then required external interventions such as cost shaping. we identified future directions were pointed out to make the model more realistic. estimating the impact of public and private strategies for controlling an epidemic: a multi-agent approach multi-agent reinforcement learning: an overview epidemic modelling: an introduction cooperative multi-agent control using deep reinforcement learning nash q-learning for generalsum stochastic games seir epidemic model with delay key: cord-016819-6r4qf63o authors: radosavljevic, vladan title: a new method of differentiation between a biological attack and other epidemics date: 2012-08-31 journal: biopreparedness and public health doi: 10.1007/978-94-007-5273-3_3 sha: doc_id: 16819 cord_uid: 6r4qf63o the main obstacle in identifying a biological attack (ba), while preventing false alarms, epidemics of panic and unnecessary expenditures is the insufficient data on which to rely. тhis new method of outbreak analysis is based on our original model of bioterrorism risk assessment. the intention was to develop a model of quick and accurate evaluation of an unusual epidemiologic event (uee) that would save time, money, human and material resources and reduce confusion and panic. this uee analysis is a subtle and detailed differentiation through assessment of ba feasibility in comparison with three other types of outbreak scenarios. there are two types of differences between these four scenarios: qualitative and quantitative. qualitative and quantitative differences are defined with 23 and 10 indicators, respectively. both types of indicators can have three different values: n/a, 0 or 1. we have carried out a feasibility analysis for subtle and detailed differentiation among four outbreak scenarios. as a tool for feasibility analysis we have introduced a “system of elimination”. system elimination is applied if one component contains all indicators scored with 0 or as n/a – the related scenario is then eliminated from further consideration. the system was applied to four uees: (1) an intentional attack by a deliberate use of a biological agent (amerithrax), (2) a spontaneous outbreak of a new or re-emerging disease (“swine flu”), (3) a spontaneous outbreak by an accidental release of a pathogen (sverdlovsk anthrax), and (4) a spontaneous natural outbreak of a known endemic disease that may mimic bioterrorism or biowarfare (kosovo tularemia). it was found that “agent” was the most important and the most informative uee component of the new scoring system. this system might be helpful in the analysis of unusual epidemic events and a quick differentiation between biological attacks and other epidemics. most diseases caused by potential biological warfare agents have low natural incidence rates. the lack of clinician experience with these diseases can impede rapid diagnosis and reporting to public health authorities [ 5 ] . the main obstacle in identifying a biological attack (ba), while preventing false alarms, epidemics of panic and unnecessary expenditures is the lack of data to rely on [ 23 ] . we are trained to consider common causes for syndromes fi rst -and unless we have a high level of suspicion -we may not realize that we need to apply non-standard methods to identify an intentional use and to detect the kind of biological agents that a terrorist might use. basically, any unexpected occurrence of one or more patients or deaths in humans or animals which might have been caused by an intentional release of pathogens may be the fi rst clue of an unusual epidemic event (uee). also, the occurrence of a single case or death caused by an unknown or already eradicated disease or agent may be considered as "unusual". three systematic models of assessing differences between natural and deliberate epidemics have been published. grunow et al. [ 12 ] put emphasis on three groups of characteristics: (1) political, military and social analysis of the af fl icted region (two criteria), (2) speci fi c features of the pathogen (three criteria), and (3) characteristics of the epidemic and clinical manifestation (six criteria). dembek et al. [ 7 ] proposed 11 potential clues to a deliberate epidemic which are focused on epidemic characteristics. these two models are accurate, but time consuming. however, saving time is crucial in the case of an uee. radosavljevic et al. [ 20 ] ) suggested a model for early orientation and differentiation between natural and deliberate outbreak. our new method of outbreak analysis is based on an original model of bioterrorism risk assessment [ 19 ] . the intention was to develop a model of quick and accurate evaluation of a uee that would save time, money, human and material resources and reduce confusion and panic. this uee analysis is a subtle and detailed differentiation through assessment of ba feasibility in comparison with other outbreak scenarios, in particular: (1) a spontaneous outbreak of a new or re-emerging disease (nr) (such as "swine fl u"), (2) a spontaneous outbreak by an accidental release of a pathogen (ar) (such as the sverdlovsk anthrax outbreak), and (3) a spontaneous natural outbreak of a known endemic disease that may mimic bioterrorism or biowarfare (ne) (such as the kosovo tularemia outbreak). after identi fi cation of an uee we introduced a new method for the subtle and detailed differentiation of such an event. in our previous paper [ 19 ] a ba was de fi ned by four components, and now we propose their equivalent terms in an uee: reservoir/source of infection vs. perpetrator, pathogen vs. biological agent, transmission mechanisms and factors vs. media and means of delivery, and susceptible population vs. target. indicators of a deliberate outbreak can be conclusive and non-conclusive [ 12 ] . conclusive indicators are direct and comparatively objective indicators for an intentional event. non-conclusive indicators estimate only the likelihood of an intentional event on the basis of circumstantial evidence. there are three conclusive indicators of deliberate outbreaks: evidence of intelligence/secrecy activities coincident or related to an outbreak, con fi rmed presence of a known bio-agent (has characteristics of traditional biological weapons or genetically modi fi ed as agent), and evidence of a means of delivery (munitions, delivery systems or dispersion systems). all other indicators are more or less non-conclusive. equivalent to the perpetrator in a biological attack is the source, or reservoir of infection in a natural epidemic. perpetrators may behave in two ways. some bioterrorists want to avoid attribution for an attack, others want to claim credit for it, or, at least want the authorities to recognize that a disease outbreak was deliberate, and not of a natural origin. people who are accidentally included in natural outbreaks (as a source or reservoir of infection) and look like perpetrators at fi rst sight, are always highly afraid and cooperative. also, a source/reservoir of infection always completely behaves according to epidemiological characteristics (incubation period, period of communicability) [ 18 ] . if political, military, ethnic, religious or other motives can be identi fi ed, this would lend credence to the assumption that an attack using pathogens or toxins as biological agents has taken place. in natural outbreaks usually there is no motive, but if we fi nd them, motive(s) are commonplace and simple. in natural epidemics sources of infection may be discovered by usual epidemiological and microbiological routine investigations, and there are no tendencies to keep themselves unknown (no secrecy) [ 19 ] . there are some coincident points related to an outbreak that may be an indicator of secrecy/intelligence inclusion in a biological attack. intelligence presents an ability to get true and on-time information on a global and local level related to biological attacks [ 19 ] . if some activities related to a biological attack are kept unobserved before an attack as well as after an attack, it is a parameter of secrecy and should be considered in the context of a biological attack. it would also be conceivable that certain persons or groups may be given suf fi cient prior warning about a biological attack and could have been spared from an epidemic by preventive measures (e.g. receiving vaccinations, adhering strictly to instructions to boil drinking water). quantitative parameters. in natural outbreaks the number and distribution of sources of infection are related to the incubation period and period of disease communicability. such regularity is seldom seen in a biological attack. here we also have to consider special situations in natural outbreaks when the incubation times and communicability may be changed: for instance in the case of an exposure to massive doses of pathogens by contaminated water or food, such as may be the case with natural disasters or accidents in water treatment plants and distribution systems or hygienic failures in kitchens. then we may have epidemiological, microbiological and clinical patterns which could resemble the characteristics we would also expect in a biological attack. strategic (large-scale) biological attack. states' institutions such as military forces, intelligence services or well-funded and possibly state-supported organizations can be perpetrators in a strategic biological attack. the present threat analysis states that in the next years only a very few so called "rogue" countries with clandestine offensive biological warfare programmes would be able to launch strategic biological attacks [ 19 ] . such attacks include politically, military and/or ideologically motivated ones. in natural, large-scale epidemics, infection is mainly unintentionally and individually disseminated and strictly related through periods of incubation and communicability of disease. the period between deployment of a bioweapon and its effects, however, is long enough to give a terrorist a chance to escape. so, it could be very dif fi cult to fi nd a perpetrator. operational (middle-scale) biological attack. this type of attack could be carried out by all three types of perpetrators (government supported institutions/organizations, terrorist groups, individuals) [ 19 ] . if psychological effects (fear and panic) are greater than the biological losses (diseased and died) it might also be a biological attack [ 21, 22 ] . tactical (small-scale) biological attack. the terrorist or criminal groups dominate as perpetrators in this type of attack. if "hard" targets are hit, perpetrators likely have to be highly skilled, or with suicidal tendencies, and politically or ideologically motivated [ 19 ] . qualitative indicators. qualitative indicators related to ba have no equivalents (not applicable -n/a) in two cases: natural outbreak of a known endemic disease that may mimic bioterrorism (biowarfare), or outbreak of a new or re-emerging disease. in the case of an accidental release of a pathogen, motivation, ability and intelligence information are n/a. three quantitative indicators of a ba have their equivalents in the other three outbreak scenarios -ar, ne and nr. the most dif fi cult scenario of a ba for investigators is if an endemic pathogen was used. in such a case microbial forensic tools for identifying a deliberate outbreak should be given priority. type of agent. there are two types of biological agents: conventional (natural form of the pathogen) and biological warfare agent. a qualitative parameter may be if some pathogen species or strain (subspecies) is unusual, atypical or antiquated, e.g. is identi fi ed in the region concerned for the fi rst time ever or again after a long absence, or if an agent has certain characteristics like: a special genetic signature, mixed with a stabilizing agent, highly concentrated, fi lled in munitions, high toxicity, more virulent, resistant to antibiotics, and multiple modes of transmission. many potential biological warfare agents could be obtained from natural sources (infected animals, patients, or contaminated soil). many pathogens, perhaps the majority concerned, cause zoonoses, i.e. infect animals as well as humans [ 16 ] . the sudden occurrence of a zoonotic disease, such as brucellosis, in the absence of the natural animal host or reservoir and other likely sources of transmission may be suggestive of an unnatural cause. the so-called "zoonotic" potential should be considered in this differentiating evaluation. a regionalized animal die-off may provide a clue that something is present or may have been released that might also infect humans. this phenomenon of animal illness heralding human illness was observed during the west nile virus encephalitis outbreak in new york city in 1999, when many local crows, along with exotic birds at the bronx zoo, died [ 17, 24 ] . in the case of a so-called "reverse spread", where human disease precedes animal disease, or human and animal disease occur simultaneously, one should consider an unnatural spread. this is often also the case in plague or tularemia outbreaks and has led to speculations like in surat (india) in 1994 or kosovo. many strains isolated from nature have low virulence. therefore, a terrorist must isolate many different strains before fi nding one suf fi ciently potent as a warfare agent. considering the technical dif fi culties to obtain virulent microorganisms from nature, terrorists may fi nd it easier to steal well-characterized strains from a research laboratory, or to purchase the known pathogenic strains from a national culture collection or commercial supplier. between 1985 and 1989, the iraqi government ordered virulent strains of anthrax and other pathogens from culture collections in france and the united states, presumably for public health research -a purpose that was legal at the time, and approved by the u.s. department of commerce [ 25, 26 ] . it is speculated that one reason for the lack of success in causing illness following dissemination of anthrax spores by the cult aum shinrikyo was the inadvertent selection of a non-pathogenic strain of bacillus anthracis [ 14 ] . strategic (large-scale) biological attack. respiratory agents are almost always candidates for strategic use because of the possibility for their clandestine use, their high dispersal potential, and their high contagiousness. category a agents, and the agents causing sars, avian in fl uenza, and pandemic in fl uenza (including swine fl u) might be candidates for use at the strategic level [ 19 ] . operational (middle-scale) biological attack. for this type of attack, the spectrum of suitable agents is wider than for large-scale attacks, and possibilities include (in addition to the agents mentioned above) hanta viruses, multi drug resistant mycobacterium tuberculosis , hepatitis a virus, noroviruses, cryptosporidium spp. , and toxins [ 19 ] . consequently, measures of detection and identi fi cation are more dif fi cult. also, the accessibility of these agents for terrorists is easier, and the amounts of the available agent are larger. tactical (small-scale) biological attack. the agents from all three categories and emerging biological agents are potential candidates for this purpose. biological agents are still the preferred materials of hoax perpetrators at the tactical level, probably because perpetrators could easily produce and safely handle these simulants of potential biological warfare agents. if there is a non-proportional, large amount of biological agent present in sources/reservoirs or environmental sampling, or an epidemiologically unexplainable transmission or distribution of an agent, there is a high probability of a biological attack. all four outbreak scenarios could have the same values for indicators related to agent. if we fi nd some kind of munition, delivery system or dispersion system (means of delivery) at the outbreak focus it should be the proof of a deliberate epidemic. if food, water, or fomites are the media of delivery it should be possible to trace it and fi nd out the source of infection and type (natural or arti fi cial) of epidemic. but, air as a medium of delivery remains the most complicated for investigation. looking for people who came in contact with the agent through this method of exposure may be very dif fi cult. however, several other parameters may help (type, strain and approximate amount of released agent, period of incubation, and period of communicability). also, natural epidemics will feature paths of transmission that are typical for the pathogen and its natural hosts. such deviations from natural paths of infection could indicate that biological agents have been deliberately disseminated. many diseases exhibit vastly different clinical presentations, periods of incubation, and mortality rates depending upon the route of transmission. therefore, outbreaks due to an atypical route of transmission, particularly aerosol transmission, such as in the inhalation anthrax example above, are more suggestive of an intentional use. weather factors, especially wind direction, temperature, and humidity, are important determinants of pathogen dissemination and disease occurrence and must be taken into account in an outbreak investigation. a disease outbreak occurring downwind (or downriver) of a suspected biological warfare agent production facility, such as in the sverdlovsk anthrax disaster, provides compelling evidence for an accidental or intentional release [ 27 ] . it is useful to plot locations where cases occur on a geographic map. if affected cases are clustered in a downwind pattern, an aerosol release should be considered like in the 1979 anthrax outbreak in sverdlovsk [ 7 ] . many people are becoming increasingly mobile and interactive with indigenous populations. as a result, they have a correspondingly greater potential for translocating diseases to previously non-endemic areas through unknown transportation of infected vectors or people during incubation or clinical symptom stages. therefore, knowledge of the at-risk population's travel and contact histories may be essential in determining the etiology and likely source of an outbreak. tourists, military personnel, traders, settlers and immigrants, and travel adventurers may carry new pathogens to unsuspecting and susceptible populations. people, storms, and fl oods can transport arthropods, rodents, snails, birds, and other creatures that can also bring new infections to previously unaffected areas. changes in human behaviour, technologic devices, the environment, institutional living, and poor nutrition or vitamin de fi cits can spark new epidemics. the speed at which an epidemic spreads is determined by the virulence, resistance and concentration of the pathogen, the contagiousness of the disease and the intensity of the transmission process, on the one hand, and on the susceptibility and disposition of the exposed population, on the other. it is unclear how changes in household sizes, working patterns, and mobility would affect transmission patterns today. incorporating detailed data on demographics and human mobility into spatially explicit models offers one method by which such extrapolation can be made more reliable, but the scale of changes mean that much uncertainty will inevitably remain. qualitative indicators: air, water, food and fomites could be the media or means of delivery for all four outbreak scenarios. quantitative indicators: three quantitative indicators of ba -munitions, delivery systems, and dispersion systems -will not exist in the other three uee scenarios -ar, ne, and nr. in natural outbreaks there is no target, but there is a susceptible (affected or endangered) population. in both natural and deliberate epidemics there can be two types of consequences: direct (death and/or illness), and indirect (political and economic). however, in a deliberate outbreak, indirect (political and economic) effects are usually intended and have great impact. in natural epidemics indirect (political and economic) effects every time are "collateral damage" or sometimes expected consequences of disasters. in addition, the use and even the threatened use of certain biological agents can have intense psychological effects on the population at large [ 19, 21 ] . in naturally occurring epidemics "soft" targets are mainly affected, because "hard" targets (e.g. heads of state or other vips) are better protected than "soft" targets (e.g. the unprotected population). there are no signs or indicators of intelligence/ secret activities (e.g. repeated visits by individuals or vehicles identi fi ed as out of place, prior warning of a possible biological attack such as active or passive immunoprophylaxis or chemo-prophylaxis of a non-target population, threats, or hoaxes). there is no suspicious behaviour: unexplained contamination of a media (air, food, and water), or use of unusual fomites (of fi ce equipment, postal letters). there is also no obvious target in a natural outbreak. some parameters of an outbreak (location of the exposure/target site, importance and number of people in the site, and distribution of people from the site) may also point to a deliberate attack. large-scale attack. nowadays, one of the main objectives of a bioterrorist is to propagate fear, anxiety, uncertainty, and depression within the population, induce mistrust of the government, in fl ict economic damage, and disrupt travel and commerce [ 19, 21 ] . causing signi fi cant outbreaks of disease may be a secondary objective. the ultimate goal of biological attacks is to cause political consequences. bioterrorists want to produce an epidemic of fear and panic [ 19, 21 ] . this cannot be evoked in such manner if the attack is clandestine and mimics a natural outbreak. naturally occurring large-scale epidemics or pandemics are only possible by aerosol transmissible agents. all other large-scale outbreaks should raise suspicion as a potential deliberate outbreak. middle-scale and small-scale attack. in the case of "hard targets" (highly prominent and protected institutions like governmental buildings, media centres, and persons such as politicians, scientist, or high military of fi cials) being affected, the probability of a deliberate outbreak is high. consequences even in small-scale attacks can be of strategic importance. "soft targets" are considered ordinary people in public places (e.g. respiratory agents in crowded and closed places like theatres, cinemas, sports events, and political meetings). small-scale outbreaks in "soft targets" are more dif fi cult to differentiate and may be of less strategic importance. except for the most blatant violations of natural principles, bioterrorism will continue to remain dif fi cult to differentiate from naturally occurring outbreaks. certain attributes of a disease outbreak, while perhaps not pathognomonic for a biological attack when considered singly, may in combination with other attributes provide convincing evidence for intentional causation. the possibility of mixed epidemics must always be taken into consideration when assessing the outbreak of a disease, since they complicate the epidemiologic situation and can present additional dif fi culties for the investigation of unusual outbreaks. there are two types of differences between these four scenarios: qualitative and quantitative. qualitative differences are de fi ned by 23 indicators, and quantitative by ten indicators. both types of indicators could have three different values: n/a, 0, or 1. in our previous article [ 19 ] numerous parameters -indicators were de fi ned. by using them we have carried out feasibility analysis for subtle and detailed differentiation among four outbreak scenarios. as a tool for feasibility analysis we have introduced a "system of elimination". system elimination is necessitated if one component contains all indicators scored with 0 or as n/a, then the related scenario is eliminated from further consideration. (table 3. reservoirs. reservoirs of infection are pigs [ 4, 6, 9 ] and turkeys, and all six qualitative indicators are not applicable (ne, nr scenarios). in consideration of ba scenario there are likely no terrorists who intend to create an uncontrolled pandemic originating in a mexico rural area. later events also showed that there was no misuse or intent either for commercial purposes by pharmaceutical industries or from military experiments. therefore this scenario has been eliminated. with pigs and turkeys as the reservoir, accidental release of the pathogen is also not likely, and therefore this scenario is eliminated. agent. undoubtedly, this is a new and emerging pathogen [ 6, 10 ] . this clearly eliminates a natural outbreak of a known endemic disease. air and fomites could be the media of delivery in a new or a re-emerging disease [ 1, 2, 13, 15 ] . susceptible population vs. target. in the "swine fl u" pandemic, intelligence and secrecy are both scored with 1 because of the early detection of the outbreak, and identi fi cation of the agent and reservoirs of infection. conclusion. "system elimination" clearly discriminates a spontaneous natural outbreak of a known endemic disease, a biological attack, outbreak by an accidental release of a pathogen because they do not have the components "agent" or "perpetrator". considering the fi rst component (perpetrator or reservoirs/sources) in the third scenario, the fi rst six qualitative indicators are n/a. the three quantitative indicators are each scored 1. taking into account the scores of 1 for intelligence and secrecy, as well as the absolute absence of material evidence of biological attack, we should accept the scenario as a spontaneous outbreak of a new or reemerging disease. emerging diseases, both new to a region like west nile virus encephalitis, and totally "new" like sars and avian in fl uenza, have occurred in the last decade. examples include the appearance of west nile virus encephalitis in new york city in 1999 [ 8 ] , bubonic plague cases in new york city in 2002 [ 3 ] , or monkey pox outbreak in the usa in 2003. the west nile virus encephalitis outbreak in new york city in 1999 constituted a true emerging infection, as the disease became established in a new location, while the plague cases were simply imported by out-of-state residents. until the epidemic in new york city in 1999, west nile virus had never been isolated in the western hemisphere. many diseases, such as dengue fever in cuba having been imported from vietnam, or vivax malaria in korea, represent a re-establishment of endemic transmission in areas from which they were once eradicated. about 40 new pathogens have been found in the last 35 years [ 11 ] . the united states was caught off-guard by the increasing aids epidemic that began in the early 1980s. today, the aids epidemic -at the beginning of the twenty-fi rst century -is worse than the worst-case scenarios that were predicted in the early 1990s. tuberculosis, re-emerged in the united states in the 1980s after decades of decline, and includes newer multidrug-resistant strains. perpetrator. there were repeated and separate bas using this agent (e.g. multiple letters sent), which is not very probable in an accidental release. agent. the causative agent in this scenario is a category a agent (ames strain from the us army medical research institute for infectious diseases, fort detrick) that was misused by an experienced insider and specially prepared and released deliberately in a signi fi cant amount. because of this, a natural outbreak of a known endemic disease and a spontaneous outbreak of a new or re-emerging disease are clearly eliminated from further consideration. transmission mechanisms and devices vs. media and means of delivery. the perpetrator used postal letters (fomites) and the american postal service (delivery system) for the ba. susceptible population vs. target. three indicators -intelligence, secrecy and personal control (of employees with access to the agent) -were not successfully applied at the initial phase of the ba. intelligence is a cornerstone of prevention. information is provided using electronic surveillance methods, local intelligence systems, and observation of possible targets. repeated visits by individuals or vehicles must be identi fi ed. the impact of secrecy has been evident in some recent incidents. such an incident occurred in the aftermath of the 2001 anthrax letter attacks. although the us postal service and the cdc knew that the brentwood postal facility in washington, d.c., was contaminated, they waited for 4 days before closing the facility and treating workers with antibiotics. by that time, one worker had died of anthrax, another was close to death, and two were gravely ill. another example is china in 2003, when the government denied the sars epidemic for 6 weeks, causing international alarm and spread of the disease. these examples illustrate that government secrecy is a persistent jeopardy, leaves the public in ignorance, and allows narrow-minded political agendas to undermine healthcare goals. personal control includes physical control of people (their health status) and behavioral control (cv review, control of suspect behavior, control of contacts) [ 19 ] . conclusion. "system elimination" clearly eliminates the other three scenarios by the "agent" and "perpetrator" components. therefore, we should accept the ba scenario as the likely event. during 1900-2001, 77 biological "events" (i.e. episodes involving the deliberate use of a biological agent to harm people) were perpetrated. of these, just four post-1945 events generated more than ten casualties [ 28 ] . besides this, about a thousand anthrax hoaxes occurred alone between 1996 and 2001 which concerned the public, administration, and public health authorities, prompting excessive decontamination and post-exposure measures and intensive forensic and laboratory investigations in order to discriminate the events as false alarms. reservoirs/sources. in the kosovo tularemia outbreak, only an insider could be a possible perpetrator as others would not have the ability or knowledge because of the unpredictable war and after-war events. the secrecy and capacity needed would be possible only from highly sophisticated insiders. the qualitative indicators from ba and natural epidemic scenarios are not similar (rodents were reservoirs) and their differences are assessed in the rest of the indicators. quantitative indicators are also different. the number of perpetrators should be numerous but were not identi fi ed, however the number of rodents as reservoirs were numerous. comparing the accessibility to sources of the agent with the distribution sources of the agent, as well as accessibility to the target by perpetrators (humans) and rodents as reservoirs, there are signi fi cant differences. because of the timing and geographically very dispersed occurrence of cases, quantitative indicators related to a perpetrator were scored with 0, and those related to a natural epidemic were scored with 1. a spontaneous outbreak by an accidental release of a pathogen and ba were not likely scenarios because of the timing, geographic separation, and repeated occurrence of cases. agent. the implicated agent was francisella tularensis holarctica , that causes a milder form of tularemia and is endemic in the balkan region. because of this, a spontaneous outbreak of a new or re-emerging disease is clearly eliminated. transmission mechanisms and devices vs. media and means of delivery. there was no convincing and conclusive evidence for devices of delivery [ 12 ] . it is well known tularemia could be spread by multiple natural transmission sources like water, food, or animals, as was the case in this outbreak. susceptible population vs. target. this component should provide the fi nal information to solve the conundrum between a ba and a natural epidemic scenario. there was no intelligence information (no convincing or conclusive indicator was documented regarding a possible perpetrator), no secrecy (no attempts to control information after the fi rst diagnosed cases), no control of means/media, no physical protection, chemical protection, or immunological protection (all three types of protection and ways to control transmission were absent or implemented late), and a lack of signi fi cance from a military/terrorist ba logistical standpoint in the importance and location of the target, and the number and distribution of the people affected. conclusion. "system elimination" clearly excluded a spontaneous outbreak of a new or re-emerging disease because of the agent type (not new or re-emerging). the scenario of spontaneous outbreak by an accidental release of a pathogen was also not likely because of timing, geographic dispersion, and repeated occurrence of cases without any convincing and conclusive indicator. the total score supports a scenario of a spontaneous natural outbreak of a known endemic disease that could mimic bioterrorism or bio-warfare. perpetrator. in the sverdlovsk anthrax outbreak, the large amount of agent (enough to contaminate a city with thousands of inhabitants and the surrounding area) was not likely spread from a natural source or reservoir of infection unobserved and in such a short time. these facts eliminate two scenarios: natural epidemic and a spontaneous outbreak of a new or re-emerging disease. regarding the circumstances in the 1960s in the former soviet union with an isolated city (sverdlovsk) in siberia, for a ba scenario, capacity and secrecy are scored with 0. agent. there was a large amount of spores of a virulent strain of bacillus anthracis (category a agent) as the causative agent. accordingly, we scored those two indicators with 1. transmission mechanisms and devices vs. media and means of delivery. there were no delivery devices identi fi ed. however, the only way to spread such large quantities of anthrax spores during this short period of time was by air. susceptible population vs. target. this component should solve the doubt between a ba and accidental pathogen release scenario. in terms of intelligence, no conclusive or inconclusive perpetrator activities or other evidence related to ba were documented, but there was very conclusive evidence related to an accidental release. in terms of secrecy, there was prolonged and stringent secrecy and disinformation supported by soviet of fi cials about the event. personal control, control of means/ media, physical protection, chemical protection, immunological protection were carried out quickly. the last four indicators -importance of target, number of people in the target, distribution of people in the target, and location of target -were without signi fi cance and any military/terrorist logic in a ba scenario. an accidental release scenario was possible, especially accounting for the circumstances (military compound dealing with production of a biological warfare agent close to the city). conclusion. "system elimination" clearly eliminates a natural epidemic and a spontaneous outbreak of a new or re-emerging disease through the perpetrator/sources/reservoirs component. large amounts of the agent in the sverdlovsk anthrax outbreak were not possible from natural sources/reservoirs of infection in such a short time and would not likely have been otherwise undetected before human cases occurred. the total score supports a scenario of a spontaneous outbreak by an accidental release of a pathogen as the most likely scenario. also, if a country rejects foreign help and experts and hides the circumstances of an epidemic it could raise suspicion for an accidental release epidemic scenario. the author has developed a new scoring method of outbreak analysis: for subtle and detailed differentiation. the method was applied to four uees: (1) an intentional attack by a deliberate use of a biological agent (amerithrax), (2) a spontaneous outbreak of a new or re-emerging disease ("swine fl u"), (3) a spontaneous outbreak by an accidental release of a pathogen (sverdlovsk anthrax), and (4) a spontaneous natural outbreak of a known endemic disease that may mimic bioterrorism or biowarfare (kosovo tularemia). it was found that "agent" was the most important and the most informative uee component of the new scoring method. this method might be helpful in the analysis of unusual epidemic events and a quick way to differentiate between biological attacks and other epidemics. survival of in fl uenza viruses on environmental surfaces the occurrence of in fl uenza a virus on house hold and day care center fomites serum cross-reactive antibody response to a novel in fl uenza a (h1n1) virus after vaccination with seasonal in fl uenza vaccine endemic, noti fi able bioterrorism-related diseases, united states, 1992-1999 emergence of a novel swine-origin in fl uenza a (h1n1) virus in humans discernment between deliberate and natural infectious disease outbreaks lessons from the west nile viral encephalitis outbreak in new york city, 1999: implications for bioterrorism preparedness pandemic potential of a strain of in fl uenza a (h1n1): early fi ndings the 2009 h1n1 in fl uenza outbreak in its historical context emerging natural threats and the deliberate use of biological agents a procedure for differentiating between the intentional release of biological warfare agents and natural outbreaks of disease: its use in analyzing the tularemia outbreak in kosovo in community transmission of in fl uenza a (h1n1) virus at a rock festival in belgium the role of risk analysis in understanding bioterrorism transmission of avian in fl uenza viruses to and between humans a dictionary of epidemiology an outbreak of west nile virus in a new york city captive wildlife population environmental health and bioterrorism a new model of bioterrorism risk assessment unusual epidemiological event -new model for early orientation and differentation between natural and deliberate outbreak bioterrorism -types of epidemics, new epidemiological paradigm and levels of prevention epidemics of panic during a bioterrorist attack -a mathematical model chinese curses, anthrax, and the risk of bioterrorism pathology of fatal west nile virus infections in native and exotic birds during the 1999 outbreak biosecurity: limiting terrorist access to deadly pathogens. peaceworks no. 52, united states institute of peace united states dual-use exports to iraq and their impact on the health of the persian gulf war veterans. 103rd congress 2nd session, : u.s. government printing of fi ce risk factors for human anthrax among contacts of anthrax-infected livestock in kazakhstan a discussion of fi ndings and their possible implications from a workshop on bioterrorism threat assessment and risk management key: cord-255514-wvjw8h4m authors: ma, yong; jiang, hao; xiao, weilin title: tax evasion, audits with memory, and portfolio choice date: 2020-10-19 journal: nan doi: 10.1016/j.iref.2020.10.010 sha: doc_id: 255514 cord_uid: wvjw8h4m in this study, we consider the memory property of tax audits to investigate the tax evasion problem from the perspective of portfolio choice. we explore the implications of the memory property for tax evasion, consumption, and asset allocation. assuming that tax audits and jumps in the risky asset both follow self-exciting hawkes processes, we provide a semi-analytical solution to this problem for an agent with constant relative risk aversion (crra) utility. we find that the memory feature does not change the agent's effective holding in the risky asset, and its effects on tax evasion and consumption are determined by the agent's risk aversion. it is suggested that government should treat agents differentially by their risk preferences and set audit-related parameters carefully to avoid unnecessary public expenditure. tax evasion is one of the most intractable and serious issues for government. rampant tax evasion means insufficient government revenues for debt repayments (chan et al., 2013) . furthermore, tax evasion may lead to a rise in inequality, since the wealthy can more readily evade their taxes (alstadsaeter et al., 2019 ). yet, tax evasion remains an inevitable global problem (slemrod, 2007) . according to data from murphy (2011) , tax evasion in 145 countries cost no less than usd 3.1 trillion in lost revenue in 2010, which exceeds 5% of their gdp and 54.9% of their expenditure on health. although recent empirical evidence shows a decrease in tax evasion, the average size of the shadow economy, as a proxy for tax evasion, still exceeds 30% of gdp (e.g., schneider et al., 2010; buehn and schneider, 2012; schneider, 2012; medina and schneider, 2018) . consequently, combatting tax evasion efficiently and rapidly has become an urgent task for government and has drawn extensive attention and interest from researchers. in the literature, allingham and sandmo (1972) and srinivasan (1973) originally provided some plausible reasons for and answers to the tax evasion issue. since then, many others have explored this phenomenon from the perspectives of government (roubini and sala-i martin, 1995; varvarigos, 2017; akhtar et al., 2019) , financial markets (blackburn et al., 2012; hanlon et al., 2015) and firms (crocker and slemrod, 2005; lópez, 2017; abdixhiku et al., 2017; gokalp et al., 2017) . some studies have sought to shed light on the mechanism of tax evasion (e.g., lin and yang, 2001; richardson, 2006; levaggi and menoncin, 2012) . however, the literature seldom explores the relationship between tax evasion, tax audit, and agents' behavior. levaggi and menoncin (2016) , as an exception, offer a portfolio approach to study the effects of taxation and fines on agents' investment and consumption behavior in the presence of tax evasion; however, they pay little attention to the interplay between tax audit strategy and agents' behavior. in this study, we focus on the relationship between agents' investment and consumption 2 j o u r n a l p r e -p r o o f decision and tax audit strategy. in terms of tax audit strategy, considering past audit results is a natural idea to improve the efficiency of a tax audit. for instance, the experiment in kleven et al. (2011) indicates that tax evasion responds negatively to prior audits, which implies that people tend to reinforce their beliefs about detection probability following a tax audit. on the other hand, alm et al. (1993) show that the memory property of tax audits, which means the audit rate depends on the past audit results, could increase taxpayer compliance; it thus has an important influence on taxpayers' decisions. note that, for simplicity, we hereinafter refer to the memory property of tax audits as audit memory. however, to the best of our knowledge, no one has incorporated this property into the theoretical study of tax evasion. we follow the same approach as was used in levaggi and menoncin (2016) to establish the linkage between tax audits with memory, tax evasion, and agents' investment and consumption choice. in addition, a clustering of downward jumps of financial asset prices has been observed during some periods, such as the recent 2007-08 financial crisis and the covid-19 pandemic, which is a type of contagion and has been empirically verified (e.g., see, lee and mykland (2007) and ). to coincide with these extreme market circumstances, we adopt a hawkes jump-diffusion process to model the risky asset prices, in which jumps in a risky asset price are described by a self-exciting hawkes process. considering audit memory, we investigate how an agent makes decisions about consumption, investment, and tax evasion in a financial market allowing for jump contagion. for agents with a crra utility function, we obtain the semi-analytical expressions for optimal consumption, investment, and tax evasion by adopting the method initiated by aït-sahalia and hurd (2015) . we find that, in the case of logarithm-utility, the level of taxation is positively correlated with tax evasion, and an increase in the penalty for tax evasion and audit intensity can effectively lower tax evasion. in the power utility case, we obtain some interesting findings. first, the effects of audit memory on tax evasion and consumption vary with the agent's risk aversion. more specifically, in the presence of audit memory, a high risk-averse agent whose relative risk aversion coefficient is more than 1 will evade more tax and reduce their consumption expenditure, while a low risk-averse agent whose relative risk aversion coefficient is less than 1 will be more compliant in taxation and promote their consumption 1 . this suggests that the audit fashion should be varied from agent to agent, in that agents with different risk aversion might behave contrarily in the same audit fashion. second, audit memory will not affect the effective holding in the risky asset. the strategy of tax audits only influences an agent's decision on tax evasion and will not influence the agent's portfolio choice with one risky asset and one riskless asset. third, tax audits with memory will enhance the positive effect of tax reduction and penalty for evasion on reducing tax evasion if an agent has high risk aversion and mitigate this effect if the agent has low risk aversion. finally, when audit memory is considered, policymakers ought to be cautious about the choice of audit memory-related parameters since these parameters will determine the effects of the audit memory. the remainder of this paper is organized as follows. section 2 sets up the model. in section 3, we discuss the optimal tax evasion, consumption, and asset allocation problem for the agent with crra utility. section 4 explores the effects of audit memory on tax evasion, consumption, and investment. section 5 concludes this paper. we assume that there are two assets in the economy. one is a risk-free asset paying interest at a constant rate r. the other is a risky asset paying no dividends, with price s t at time t. in the literature, the jump-diffusion process is extensively applied to the modeling of the risky asset price, where the jump arrivals mostly follow the lévy process. however, the 1 hereinafter, we refer to an agent as the high (low) risk-averse agent if he is more (less) risk-averse than the log-utility 1. lévy process cannot capture the empirically detected jump clustering effect during financial market turmoil. in this study, we adopt the self-exciting hawkes process to model the jump arrivals, and we assume the risky asset price is driven by the following jump-diffusion process: where µ is the expected return, σ is the volatility of the risky return, and b t is a standard are independent and identically distributed jump sizes with mean φ and support on (−1, ∞). n 1,t is a hawkes process with the jump arrival intensity λ 1,t , given by where α 1 , β 1 andλ 1 are all positive constants. for the hawkes process n 1 (t), once a jump occurs, the intensity will instantly increase by α 1 and then the increment decays exponentially at a rate β 1 over time. note that the hawkes process n 1 (t) can describe the jump contagion risk of the risky asset, and it will degenerate to a poisson process when α 1 = β 1 = 0. in this study, we consider an agent subject to capital gains taxation. as in levaggi and menoncin (2016) , we assume that (i) capital gains are taxed continuously and in a symmetric manner, that is, the agent pays tax if the change in asset price is positive and receives a refund if it is negative 2 ; and (ii) consumption will not be taxed. denote by τ g and τ the tax rates levied on the payoff of the riskless asset and risky asset, respectively. accordingly, we define r g = r(1 − τ g ) as the after-tax or tax-adjusted return of the riskless asset. in this economy, the agent might conceal part of their investment in the risky asset in order to evade tax, while evasion is subject to a fine when it is detected, and the fine is a proportion of the total amount of revenue evaded, denoted by θ(τ ) ∈ [0, 1]. as illustrated in levaggi and menoncin 2 the reasoning for this assumption is given in levaggi and menoncin (2016) , and this assumption is also made in cai et al. (2017) . (2016), this fine function θ(τ ) includes two important cases: a fine proportional to the value of the evaded asset and a fine proportional to the tax evaded. in terms of tax audits, in the existing literature, it is generally assumed that audits are conducted with constant intensity (state-independent), and thus the poisson process is often applied to the modeling of audits. in contrast, it is more practical that, once the evasion is discovered by the audit authority, the agent will be faced with a larger audit intensity; nonetheless, the audit authority will gradually decrease the audit intensity to the usual level until tax evasion is found again. in other words, the audit intensity should be dependent on the past evasion. to describe this feature of tax audits, we use another hawkes process n 2,t to model the arrival of audits. the arrival intensity λ 2,t of audits is driven by where 1(·) is the indicator function, and α 2 , β 2 andλ 2 are all positive constants. it means that, when evasion is discovered, the audit intensity will instantly increase by α 2 and then the increment of intensity decays at an exponential rate β 2 . meanwhile, we assume that b t , n 1,t , n 2,t and y t are mutually independent. let w t be the agent's wealth at time t. we assume that the agent consumes c t at time t and there is no endowment stream. therefore, their wealth follows the dynamics: (1 − τ )π t + π e t can be regarded as the effective holding weight in the risky asset hereinafter denoted by κ t . by combining (1) and (4), we obtain in this section, we study the infinite-horizon consumption and investment problem for an agent who has utility function u (x). with a positive initial wealth w 0 , the agent's objective is to maximize the discounted expected utility of consumption by determining their consumption, and the fractions of their wealth invested in the non-concealed and concealed risky asset. specifically, the optimal problem for the agent is to achieve max {πs,π e s ,cs:s≥0} where the wealth process w t is governed by (5). we now turn to solve the optimal consumption and asset allocation problem by using the stochastic control approach. following the standard procedure, the indirect utility function or the value function is defined by where e t,w,λ [·] is the expectation with initial conditions w t = w and (λ 1,t , λ 2,t ) = λ = (λ 1 , λ 2 ). according to the principle of optimal stochastic control, the following hamilton-jacobi-bellman (hjb) equation for the indirect utility function j is achieved : where j t , j w , j λ i denote the partial derivatives of j(t, w, λ) with respective to t, w, λ i , and j ww denotes the second-order partial derivative. in addition, α 1 = (α 1 , 0) and α 2 = (0, α 2 ). due to the time-homogeneity of the value functions for general infinite-horizon problems, the value function j is of the form subsequently, the hjb equation reduces to according to (9), the first order conditions for the optimal consumption c * and investment where κ * = (1 − τ )π * + π e * . subtracting (11) from (12), we obtain next, the optimal consumption, tax evasion and asset allocation for the agent with the crra utility will be derived. for the log-utility agent with utility function u (x) = log x, the optimal policy is summarized in the following proposition. proposition 3.1. let c * , π * and π e * denote the optimal consumption, the optimal fractions of wealth invested in the unconcealed risky asset and the concealed risky asset, respectively. let κ * = (1 − τ )π * + π e * . then the optimal policy for the log-utility agent is given by proof. see appendix a. corollary 3.2. for the log-utility agent, if short selling is not allowed in the market, the optimal weight in the unconcealed risky asset decreases with the jump intensity and increases with the audit intensity. the optimal weight in the concealed risky asset is independent of the jump intensity, while it decreases with the audit intensity. proof. see appendix b. according to proposition 3.1 and corollary 3.2, it is intuitive that tax evasion decreases with the fine θ(τ ), which means penalty for tax evasion is helpful in curtailing tax evasion. if the fine function θ(τ ) is proportional to the value of the evaded asset, as given in allingham and sandmo (1972) , clearly tax evasion will be increasing in the tax rate. if the fine function θ(τ ) is proportional to the tax evaded, say θ(τ ) = ατ with α > 1, the relationship between tax evasion π e * and tax rate τ depends on fiscal parameters (θ(τ ), λ 2 , τ, τ g ) and the interest rate r, which is consistent with the finding in levaggi and menoncin (2016) . specifically, it follows from (14) that ∂π e * ∂τ = αλ 2 −r g αr g 1 τ 2 . therefore, the effect of tax rate on tax evasion is positive when αλ 2 > r g . however, the effect is negative when αλ 2 < r g , i.e., an increase in tax rate leads to a decline in tax evasion, which is counter-intuitive, as also found by yitzhaki (1974) and levaggi and menoncin (2016) . on the other hand, the log-utility agent will reduce tax evasion when the audit intensity increases. it is intuitive; their holding in the unconcealed risky asset is increasing in audit intensity. these results support the view that increasing the enforcement of tax audits is an effective way to reduce tax evasion and can improve investment activity in formal financial markets. it has also been observed that, for the log-utility agent, jumps in the risky asset merely affect their unconcealed holding, which means that their decision on tax evasion has nothing to do with the market's jump risk. we now specialize this problem for an agent with power utility the optimal policy for the power-utility agent is concluded in the following proposition. j o u r n a l p r e -p r o o f proposition 3.3. let c * , π * and π e * denote the optimal consumption, the optimal fractions of wealth invested in the unconcealed risky asset and the concealed risky asset, respectively. then the optimal policy for the agent with power utility is where given by (c.8). proof. see appendix c. according to proposition 3.3, it seems that the optimal consumption and portfolio choice are "analytical". however, they are not 3 . for ease of computation, we consider a special case in which the asset price jumps follow a poisson process, i.e., and the agent's wealth process follows: corollary 3.4. suppose that the jumps of the risky asset are driven by a poisson process with constant intensity λ 1 ≡λ 1 , which is equivalent to the hawkes process n 1,t with α 1 = β 1 = 0. then, the optimal policy is where p(x) is the solution to the following ordinary differential equation (ode): proof. let h(λ 1 , λ 2 ) = p(λ 2 ). then, substituting λ 1 =λ 1 and α 1 = β 1 = 0 into (c.6) and (18) easily yields the desired result. it is found from (18) that κ * is independent of tax audits and naturally has nothing to do with the memory feature of tax audits. the economic explanation for this result is simple. recall that κ * can be regarded as the effective investment weight in the risky asset. since tax audits do not influence the intrinsic risk and expected return of the risky asset, the agent with a power utility function will not adjust their effective holding in the risky asset. on the other hand, due to κ * = (1 − τ )π * + π e * and 0 < 1 − τ < 1, the rise in tax evasion, caused by the decline in audit intensity, is associated with a decrease in the total investment weight in the risky asset and an increase in the weight in the safe asset, notwithstanding the unchanged expected return and risk of the risky asset. thus, when a representative agent has a power utility function, tax evasion could have a negative impact on investment activity in the capital market; this provides an alternative explanation for the inactivity in brazil's stock market discussed in kenyon (2008) . in what follows, we provide a method to numerically solve the optimal policy for the power-utility agent in the special case. according to the third equation in (18), the numerical solution for κ * can easily be obtained given the law of y . then we need only to yield p(x) which is determined by the ode (19) . intuitively, if there is no restriction imposed on investment, π e * will be a monotonically decreasing continuous function of λ 2 , taking on values from the positive to the negative as λ 2 increases. it implies that there exists a threshold, denoted by λ thold , such that π e * (λ thold ) = 0 and π e * (λ 2 ) < 0 for λ 2 > λ thold . on the other hand, short selling in the concealed account is impractical, so π e * must be positive or zero. accordingly, π e * = 0 for all λ 2 ≥ λ thold , i.e., the agent will not conceal any of their investment as the audit intensity passes some threshold. because tax evasion will vanish when the audit intensity equals the threshold, we have λ thold = r g τ θ(τ )(1−τ ) , which is deduced from the expression of π e * in (18). let y(x) = ln p(λ thold − x), then the ode (19) can be transformed into the following delay differential equation (dde): where since π e * (λ 2 ) = 0 and p(λ 2 ) =p(λ thold ) for all λ 2 > λ thold , we have y(x) = y(0) for all x < 0. it gives rise to the following boundary condition: where κ * is the solution to the third equation in (18). once the dde (20) with boundary condition (21) is solved 4 , p(x), c * and π e * can in turn be derived. in this section, to demonstrate more clearly the implications of audits with memory for the power-utility agent's tax evasion, consumption and investment decisions, we consider the poisson jump-diffusion model (16) for the risky asset price; accordingly our analysis is based on the results in corollary 3.4. for comparison, we take the constant audit intensity (and constant jump intensity) case as the benchmark. in the benchmark model, tax audits are memoryless and the wealth process is given by whereñ i,t are poisson processes with constant intensitiesλ i , i = 1, 2. the optimal policy for the benchmark model can be directly derived by setting α 2 = β 2 = 0 in corollary 3.4. in the following analyses, we assume each jump size y i is log-normally distributed, say, log(y i + 1) ∼ n (µ j , σ j ), and then the mean of y i is φ = exp(µ j + 1 2 σ 2 j ) − 1. under this table 1 . note that, to avoid yitzhaki's (1974) counterintuitive result, we assume the fine is proportional to the value of the evaded asset, i.e., θ(τ ) ≡ θ. in the following comparative static analyses, the parameters will take on the values as given in table 1 unless otherwise specified 5 . in this subsection, we mainly examine how agents with different risk aversion make decisions on tax evasion when audit memory is present and the effect of audit memory on tax evasion. 5 we have also considered a case where the tax rates for the risky and riskless assets are the same. we find that the results are quite similar with those in the case where the tax rates are different. due to the limited space, the results in this special case are not presented in this paper; they are available upon request. the results for a low risk-averse agent (γ = 0.8); right graphs: the results for a high risk-averse agent (γ = 4). as shown in figure 1 , in the presence of audit memory, the high risk-averse agent has a lower level of tax evasion than the low risk-averse one. this result is intuitive, for, the risk of being audited and being punished for the concealed account can be regarded as a source of risk; compared to the low risk-averse agent, the high risk-averse one will be less inclined to tax evasion if all other factors are the same. this result is in line with the empirical finding in alabede et al. (2011) , which shows a positive relationship between a taxpayer's risk preference and their tax compliance behavior. in addition, as expected, the negative relationship between audit intensity λ 2 and the concealed investment fraction π e is confirmed for both the high risk-averse and the low risk-averse agents, regardless of whether tax audits have memory. intuitively, the agent, regardless of their risk preference and the tax audit fashion, should cut down on their concealed investment when tax audits occur more frequently. this finding is also supported by many other studies in the literature (e.g., fischer et al., 1992; levaggi and menoncin, 2016; ma et al., 2019) . in what follows, we continue to examine the effect of audit memory on tax evasion. first, compared to the benchmark, the high (low) risk-averse agent tends to conceal more (less) of their risky investment when tax audits have memory property. it means that, although the negative relationship between tax evasion and audit intensity is always true whatever audit fashion is applied, the effect of audit memory on tax evasion is uncertain and depends on the agent's degree of risk aversion. this is consistent with the empirical finding in mohdali et al. (2014) . there is a plausible economic explanation for this result. in the circumstance that tax audits have the memory property, once tax evasion is detected, audit intensity will increase sharply and the rational action for the agent will be to close their concealed account until audit intensity returns to an acceptable level. as a result, the agent will suffer a substantial "loss" in the future because of the absence of tax evasion, which implies that audit memory is an indirect punishment on tax evasion (penalty is a direct one). this punishment will change the agent's investment opportunity set temporally. thus, he is supposed to take this change into consideration and react appropriately. as kim and omberg (1996) , , , and branger et al. (2009) have explained, the reaction of the agent depends on whether he is more or less risk-averse than the log-utility agent. for the agent who is more risk-averse than the log-utility agent, since their utility is unbounded from below, he is more careful about hedging and, therefore, will conceal more risky investment in the case of audits with memory, when tax evasion is still profitable at present, to pursue extra profit to compensate for the potential "loss" in the future. this tax evasion behavior is a version of "enjoying pleasure in good time". in contrast, for the agent who is less risk averse than the log-utility agent, their utility is bounded from below but unbounded from above, and hence hedging is a less attractive choice for him. additionally, the indirect punishment renders concealing their investment less attractive. therefore, the low risk-averse agent is prone to hold less concealed asset when audits have the memory property. from figure 1 , in the presence of audit memory, the sensitivity of tax evasion to audit intensity for the high risk-averse agent increases as audit intensity decreases, and an opposite result obtains for the low risk-averse agent. the rationale for this is, because the concealed investment decreases with audit intensity, the potential reduction in investment return resulting from indirect punishment tends to increase, and the hedging demand increases as audit intensity decreases. second, as the decay rate β 2 rises, the agent with low (high) risk aversion will boost (reduce) their concealed position. this result is obvious since the speed of audit intensity reverting to the normal level increases with the decay rate. therefore, for any specific agent, the effects of the decay rate and audit intensity on tax evasion are opposite. in addition, since the increment of audit intensity as a punishment for tax evasion is set to make the agent refrain from concealing any investment at least for a while, it is likely that the agent will not react to a larger increment of audit intensity. our findings in this subsection reveal that the effect of audit memory on tax evasion 18 j o u r n a l p r e -p r o o f depends on agent's risk aversion. therefore, as mohdali et al. (2014) say, the effect of the threat of punishment on tax evasion varies from group to group, and thus policymakers should consider taxpayers' risk preference when formulating their tax audit strategy. more specifically, policymakers should consider the history of the taxpayer's tax evasion if he is low risk-averse, while they should choose a memoryless tax audit strategy for the taxpayer with a high risk aversion. it is worth noting that, if policymakers adopt a tax audit strategy with the memory property, it is unnecessary for policymakers to set an extremely high punishment in terms of audit intensity. instead, they should pay more attention to the decay rate of audit intensity since it partly determines the effect of audit memory. from figure 2 , tax evasion can be reduced by increasing the fine or decreasing the tax rate, but the extent depends on agents' risk aversion in the presence of audit memory. specifically, for the high (low) risk-averse agent, the impact of the tax rate and fine on tax evasion will be strengthened (weakened) by adding memory to tax audits. the potential loss made from being caught and fined is the main explanation for this result. intuitively, when the tax rate decreases or the fine increases, tax evasion will decrease and the hedge demand for the potential loss will decline, which makes the high risk-averse agent reduce their current investment in the concealed asset. therefore, for the high risk-averse agent, audit memory enhances the effects of the tax rate and fine on tax evasion. on the other hand, the lower the concealed investment, the lower the potential loss of revenue caused by indirect punishment. thus, for the low risk-averse agent who is more sensitive to the potential loss of revenue, audits with memory make their concealed investment less sensitive to the tax rate and fine than memoryless audits. these results indicate that audit memory reduces evasion more (less) efficiently by reducing the tax and increasing the fine for the high (low) risk-averse agent. accordingly, if the authority decides to adopt a tax audit fashion which possesses the 19 j o u r n a l p r e -p r o o f memory property, they should consider taxpayers' degree of risk aversion before using taxes and fines as a means reduce tax evasion. figure 2 : the sensitivity of optimal tax evasion (π e ) to the tax rate and the fine (τ and θ). left graphs: the results for a low risk-averse agent (γ = 0.8); right graphs: the results for a high risk-averse agent (γ = 4). as illustrated in figure 3 , first, in comparison to the case of memoryless audits, the high (low) risk-averse agent will consume less (more) when audits have memory property. this implies that the optimal consumption also depends on the degree of risk aversion of tax audits with memory figure 3 : the sensitivity of optimal consumption-wealth ratio (c/w) to the audit-related parameters (λ 2 , β 2 and α 2 ). left graphs: the results for a low risk-averse agent (γ = 0.8); right graphs: the results for a high risk-averse agent (γ = 4). by an audit's memory property outweighs the corresponding income effect. in the presence of audits with memory, a high risk-averse agent temporarily takes higher audit risk, which increases the relative price of consumption and generates a substitution effect. although, on the other hand, the higher concealed investment also increases the agent's net expected return and buying power, the high risk-averse agent usually cares more about risk. as a result, the substitution effect is stronger than the income effect. for the low risk-averse agent, compared with the weaker income effect generated by the decreased income, the substitution effect produced by the decreased income and the increased punishment from tax audits will more sharply lower their interest in earning money and will promote their interest in consumption and enjoyment of leisure time. thus, in the presence of audits with memory, the low riskaverse agent is more willing to enjoy consumption, whereas the high risk-averse agent has less interest therein. second, for the high (low) risk-averse agent, their consumption is positively (negatively) related to audit frequency when tax audits have the memory property, which is in total contrast to their consumption behavior when audits are memoryless. when audit intensity rises, so does the likelihood of being caught in evasion. therefore, the net expected return on concealed investment, for all agents, would decline and thereby weaken the income effect. however, the substitution effect declines more rapidly than the income effect. on the one hand, the relative price of consumption increases with the intensity. on the other hand, the indirect punishment has less impact on the agent's decision about the concealed account and consumption. as a consequence, the substitution effect will diminish steeply, and the high risk-averse agent will consume more goods while the low risk-averse agent moderates their consumption when audit intensity increases. in this study, we develop a theoretical model of tax evasion to investigate the relationship between tax audits with memory and agents' tax evasion, investment, and consumption decisions. we find that tax evasion will curtail the total investment in the risky asset and increase investment in the riskless asset. however, the memory feature of audits does not affect an agent's effective holding in the risky asset. this finding provides theoretical support for the negative relationship between tax evasion and prosperity in the capital markets. we also find that the impact of audit memory on tax evasion is closely related to the degree of agents' risk aversion. specifically, in comparison to the case of memoryless audits, when audits have the memory property, the low risk-averse agent will reduce tax evasion and the high risk-averse agent behaves contrarily. the negative relationship between audit intensity and tax evasion will be enhanced (reduced) for the low (high) risk-averse agent when the audit intensity rises. although the audit intensity decay rate plays an important role in the battle against tax evasion, increasing the punishment through audit intensity is only effective up to a certain threshold. likewise, the effects of a tax reduction and an increase in penalty on tax evasion also vary with an agent's degree of risk aversion. finally, an agent's consumption behavior in the case of an audit with memory is opposite to that in the absence of the memory property, depending also on their risk aversion as well as audit-related parameters. from a policy point of view, policymakers should treat agents differentially in terms of tax audits. for the low risk-averse agent, tax audits should be dependent on their history of evasion since this audit fashion will reduce their evasion and increase their consumption expenditure, while for the high risk-averse agent tax audits should be memoryless. in addition, an increase in audit intensity as punishment for tax evasion should be modest since it is costly and its effect neutral at and beyond a certain threshold level. where policymakers adopt the audit fashion with the memory feature, they should be careful when using tax reduction and increased fines to reduce tax evasion, because the effectiveness of these measures depend on an agent's degree of risk aversion. lastly, policymakers need to attain a balance between audit cost and policy effect when deciding on an appropriate audit intensity decay rate. and the optimal weight in the unconcealed risky asset is π e * = 1 θ(τ ) 1 − λ 2 θ(τ )(1 − τ ) r g τ h(λ + α 2 1(π e * > 0)) h(λ) 1 γ (c.4) in addition, (11) can be rewritten as the optimal consumption and investment are given by (c.3), (c.4) and (c.5). however, they are the functionals of h(λ). now we proceed to derive the function h(λ). substituting the optimal consumption and investment into the hjb equation (9), we have + λ 2 [(1 − θ(τ )π e * ) 1−γ h(λ + α 2 1(π e * > 0)) − h(λ)]. (c.6) multiplying 1 γ h 1 γ −1 (λ) on both sides of (c.6), we have where k(λ) = 1 γ − ρ + (1 − γ)[r g (1 − π * − π e * ) + κ * (µ − φλ 1 ) − 1 2 κ * 2 σ 2 γ] let g(λ) = h 1 γ (λ). then [ag](λ) + k(λ)g(λ) = −1. (c.9) where a is the markov generator given by firm-level determinants of tax evasion in transition economies modeling financial contagion using mutually exciting jump processes portfolio choice in markets with contagion multinationals' tax evasion: a financial and governance perspective individual taxpayers attitude and compliance behaviour in nigeria: the moderating role of financial condition and risk preference income tax evasion: a theoretical analysis tax compliance with endogenous audit selection rules tax evasion and inequality an empirical investigation of continuous-time equity return models tax evasion, the underground economy and financial development what is the impact of stock market contagion on an investor's portfolio choice? shadow economies around the world: novel insights, accepted knowledge, and new estimates portfolio selection with capital gains tax, recursive utility, and regime switching tax audit and investigation in china and hong kong corporate tax evasion with agency costs detection probability and taxpayer compliance: a review of the literature competition and corporate tax evasion: an institution-based view taking the long way home: us tax evasion and offshore investments in us equity and debt markets tax evasion, disclosure, and participation in financial markets: evidence from brazilian firms dynamic nonmyopic portfolio behavior unwilling or unable to cheat? evidence from a tax audit experiment in denmark jumps in financial markets: a new nonparametric test and jump dynamics tax audits, fines and optimal tax evasion in a dynamic context optimal dynamic tax evasion: a portfolio approach a dynamic portfolio choice model of tax evasion: comparative statics of tax rates and its implication for economic growth dynamic asset allocation with event risk dynamic derivative strategies a quantitative theory of tax evasion optimal investment and consumption in the market with jump risk and capital gains tax shadow economies around the world: what did we learn over the last 20 years? imf working paper no the impact of threat of punishment on tax compliance and non-compliance attitudes in malaysia collect the evaded tax, avoid the cuts. the guardian 25 determinants of tax evasion: a cross-country investigation a growth model of inflation, tax evasion, and financial repression the shadow economy and work in the shadow: what do we (not) know? new estimates for the shadow economies all over the world cheating ourselves: the economics of tax evasion tax evasion: a model cultural norms, the persistence of tax evasion, and economic growth optimal consumption and savings with stochastic income and recursive utility a note on income tax evasion: a theoretical analysis writing-original draft, supervision, funding acquisition weilin xiao: methodology, investigation, formal analysis the authors are grateful to the editor, carl r. chen, and two anonymous reviewers for their insightful comments and suggestions. we are grateful to the editor, carl r. chen, and two anonymous reviewers for their insightful comments and suggestions. ma gratefully acknowledges the financial support of the national natural science foundation of china (no. 71971077), and the hunan provincial natural science foundation of china (no. 2019jj30001). xiao gratefully acknowledges the financial support of the national natural science foundation of china (no. 71871202). proof of proposition 3.1. for the log-utility agent, we conjuncture thatit follows from (10) that the optimal consumption process iscombining (a.2) and (13), we obtainthus, (11) can be rewritten asproof of corollary 3.2. in (a.4), taking the partial derivatives of π e * with respect to λ 1 andin (a.5), taking the partial derivative of κ * with λ 1 and λ 2 , we obtainit implies that the after-tax weighted proportion invested in the risky asset has nothing to do with the audit intensity. if short selling is not allowed in the market, then 0 < κ * ≤ 1appendix c.proof of proposition 3.2. for the power-utility agent, a candidate value function would befor some function h. obviously,it follows from (10) and (13) that the optimal consumption is the authors declared that they do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted. key: cord-021887-22lop0pk authors: artenstein, andrew w. title: biological attack date: 2015-10-23 journal: ciottone's disaster medicine doi: 10.1016/b978-0-323-28665-7.00079-0 sha: doc_id: 21887 cord_uid: 22lop0pk nan biological attack bioterrorism can be broadly defined as the deliberate use of microbial agents or their toxins as weapons. the broad scope and mounting boldness of worldwide terrorism exemplified by the massive attacks on new york city and washington, dc, on september 11, 2001 , coupled with the apparent willingness of terrorist organizations to acquire and deploy biological weapons, constitute ample evidence that the specter of bioterrorism will continue to pose a global threat. as in other aspects of daily life and the practice of medicine, in particular, the concept of "risk" is germane to considerations regarding an attack using biological agents. risk, broadly defined as the probability that exposure to a hazard will lead to a negative consequence, can be accurately calculated for a variety of conditions of public health importance (table 79-1) . however, the quantification of risk as it pertains to bioterrorism is imprecise because accurate assessment of exposure depends on the whims of terrorists, by nature, an unpredictable variable. although the probability of exposure to a biological attack is statistically low, it is not zero. because the negative consequences of an attack are potentially catastrophic, an understanding of biological threat agents and a cogent biodefense strategy are important components of disaster medicine. biological weapons have been used against both military and civilian targets throughout history, perhaps as early as 600 bc. 1 in the fourteenth century, tatars attempted to use epidemic disease against the defenders of kaffa, by catapulting plague-infected corpses into the city. british forces gave native americans blankets from a smallpox hospital in an attempt to affect the balance of power in the ohio river valley in the eighteenth century. 2 in addition to their well-described use of chemical weapons, axis forces purportedly infected livestock with anthrax and glanders to weaken allied supply initiatives during world war i. perhaps the most egregious example of biological warfare involved the japanese program in occupied manchuria from 1932 to 1945. based on survivor accounts and confessions of japanese participants, thousands of prisoners were murdered in experiments using a variety of virulent pathogens at unit 731, the code name for a notorious japanese biological weapons facility. 3 the united states maintained an active program for the development and testing of offensive biological weapons from the early 1940s until 1969, when the program was terminated by executive order of then president nixon. current efforts continue as countermeasures against biological weapons. the convention on the prohibition of the development, production, and stockpiling of biological and toxin weapons and on their destruction (bwc) was ratified in 1972, formally banning the development or use of biological weapons, and assigning enforcement responsibility to the united nations. 2 unfortunately, the bwc has not been effective in its stated goals; multiple signatories have violated the terms and spirit of the agreement. the accidental release of aerosolized anthrax spores from a biological weapons plant in the soviet union in 1979, with at least 68 human deaths from inhalational anthrax reported downwind, was proven years later to have occurred in the context of offensive weapons production. events within the past 30 years have established bioterrorism as a credible and ubiquitous threat: for example, the 1984 incident in the dalles, oregon, involving the intentional contamination of restaurant salad bars with salmonella, by a religious cult attempting to influence a local election. 4 public fears were additionally heightened by the international events following the japanese aum shinrikyo cult's sarin attack in tokyo in 1995, especially after investigations revealed that the group had been experimenting with aerosolized anthrax release from rooftops for several months prior. more recently, un weaponsinspector findings of significant quantities of weaponized biological compounds in iraq during the gulf war and the subsequent aftermath has served as sentinel warnings of a shift in terrorism trends. this trend culminated with the october 2001 anthrax attacks in the united states, which elevated bioterrorism to the forefront of international dialogue and heightened public concerns regarding systemic health care preparation against the threat of biological attacks. biological agents are considered weapons of mass destruction (wmds) because, as with certain conventional, chemical, and nuclear weapons, their use may result in large-scale morbidity and mortality. a world health organization (who) model based on the hypothetical effects of the intentional release of 50 kg of aerosolized anthrax spores upwind from a population center of 500,000 (analogous to that of metropolitan providence, ri) estimated that the agent would disseminate in excess of 20 km downwind and that nearly 200,000 people would be killed or injured by the event. 5 biological weapons possess unique properties among wmds. by definition, biological agents are associated with a clinical latency period of days to weeks, in most cases, during which time early detection is quite difficult with currently available technology. yet, early detection is critical because specific antimicrobial therapy and vaccines are available for the treatment and prevention of illness caused by certain biological weapons. casualties from other forms of wmds can generally only be treated by decontamination (with antidotes available for only some types), trauma mitigation, and supportive care. additionally, the possibility of a biological attack provokes fear and anxiety-"terror"-disproportionate to that seen with other threats, given their often invisible nature. the goals of bioterrorism are those of terrorism in general: morbidity and mortality among civilian populations, disruption of the societal fabric, and exhaustion or diversion of resources. a successful outcome from a terrorist standpoint may be achieved without furthering all of these aims but instead disrupting daily life. the anthrax attacks in the united states in 2001 evoked significant anxiety and diverted resources from other critical public health activities despite the limited number of casualties. in many cases, the surge capacity of our public health system has been inadequate to deal with the emergency needs, resulting in reform and additional planning after the event. to be used in large-scale bioterrorism, biological agents must undergo complex processes of production, cultivation, chemical modification, and weaponization. for these reasons, state sponsorship or direct support from governments or organizations with significant resources, contacts, and infrastructure would predictably be required in large-scale events. however, revelations have suggested that some agents may be available on the worldwide black market and in other illicit settings, thus obviating the need for the extensive production process. 6 although traditionally thought to require an efficient delivery mode, recent events, including the 2001 united states anthrax attacks, demonstrated the devastating results that can be achieved with relatively primitive delivery methods (e.g., high-speed mail-sorting equipment and mailed letters). numerous attributes contribute to the selection of a pathogen as a biological weapon: availability or ease of large-scale production, ease of dissemination (usually by the aerosol route), stability of the product in storage, cost, and clinical virulence. the last of these refers to the reliability with which the pathogen causes high mortality, morbidity, or social disruption. the centers for disease control and prevention (cdc) has prioritized biological-agent threats based on the aforementioned characteristics, and this has influenced current preparation strategies (table 79 -2). 7 category a agents, considered the highest priority, are associated with high mortality and the greatest potential for major effects on the public health. category b agents are considered "incapacitating" because of their potential for moderate morbidity but relatively low mortality. most of the category a and b agents have been experimentally weaponized in the past and thus have proven feasibility. category c agents include emerging threats and pathogens that may be available for development and weaponization. another factor that must be addressed in assessing future bioterrorism risk is the historical record of experimentation with specific pathogens, informed by the corroborated claims of various high-level soviet defectors and data released from the former offensive weapons programs of the united states and united kingdom. 2, 7, 8 information from these sources, combined with the burgeoning fields of molecular biology and genomics, demonstrates that future risk scenarios will likely have to contend with genetically altered and "designer" pathogens intended to bypass current known medical countermeasures or defenses. to this end, a miscellaneous grouping of potential threat agents is added to the extant cdc categories in table 79 bioterrorist attacks are often insidious. absent of advance warning or specific intelligence information, clinical illness will likely manifest before the circumstances of a release event are known. for this reason, health care providers are likely to be the first responders and reporting agents of this form of terrorism. this is in contrast to the more familiar scenarios in which police, firefighters, paramedics, and other emergency services personnel are deployed to the scene of an attack with conventional weaponry or a natural disaster. physicians and other health care workers must therefore maintain a high index of suspicion of bioterrorism, and recognize suggestive epidemiologic clues and clinical features to enhance early recognition and guide initial management of casualties. early recognition and rapid deployment of specific therapy remains the most effective way to minimize the deleterious effects of bioterrorism on both exposed individuals and public health. unfortunately, early recognition is hampered for multiple reasons. as previously discussed, it is likely that the circumstances of any event will only be known in retrospect. therefore responders may be unable to discern the extent of exposure immediately. also, terrorists have a nearly unlimited number of targets in most open democratic societies, and it is unrealistic to expect any governing body without detailed intelligence of an impending attack to secure an entire population at all times. certain sites, such as government institutions, historic landmarks, or large public gatherings, may be predictable targets; however, other facilities may fall victim to bioterrorism. in fact, government data support that businesses and other economic concerns were the main targets of global terrorism during the period from 1996 to 2002. 9 metropolitan areas are traditionally considered especially vulnerable given the dense populations and already existing public gathering areas such as subways and office buildings. because of the expansion of suburbs and the commuter lifestyle, as well as the clinical latency period between exposure and symptoms, casualties of bioterrorism are likely to present for medical attention in diverse locations and at varying times after a common exposure. an event in new york city on a wednesday morning may result in clinically ill persons presenting over the ensuing weekend to a variety of emergency departments within a 60-mile radius. finally, current modes of transportation ensure that there will be affected persons thousands of miles away, at both national and international locations, related to a single common exposure. this adds layers of complexity to an already complicated management strategy and illustrates the critical importance of surveillance and real-time communication in the response to suspected bioterrorism. further hindering the early recognition of bioterrorism is that initial symptoms of a biological weapon may be nonspecific and nondiagnostic. in the absence of a known exposure, many symptomatic persons may not seek medical attention early, or if they do, they may be misdiagnosed as having a viral or flu-like illness. if allowed to progress beyond the early stages, many of these illnesses deteriorate quite rapidly, and treatment may be significantly more difficult. most of the diseases caused by agents of bioterrorism are rarely, if ever, seen in modern first-world clinical practice. physicians are likely to be inexperienced with their clinical presentation and be less aware of alarming symptomatic constellations. additionally, these agents by definition will have been manipulated in a laboratory and may not present with the classic clinical features of naturally occurring infection. this was dramatically illustrated by some of the inhalational anthrax cases in the united states in october 2001. 10 early recognition of bioterrorism is facilitated by the recognition of epidemiologic and clinical clues. clustering of patients with common signs and symptoms-especially if regionally unusual or otherwise characteristic of bioterrorism agents-is suggestive of an intentional exposure and should prompt expeditious notification of local public health authorities. this approach will also lead to the recognition of outbreaks of naturally occurring disease or emerging pathogens. the recognition of a single case of a rare or nonendemic infection, in the absence of a travel history or other potential natural exposure, should raise the suspicion of bioterrorism. finally, unusual patterns of disease, such as concurrent illness in human and animal populations should raise suspicions of bioterrorism or another form of emerging infection. an effective response to bioterrorism requires coordination of the medical system at all levels, from the community physician to the tertiary care center, with rapid activation of public health, emergency management, and law enforcement infrastructures. this section provides a broad overview of the biological threat agents thought to be of major current concern-largely, the cdc category a agents. extensive coverage of specific pathogens can be found in related chapters in this text and in other sources. 11 these agents can possess rapid person-to-person transmission or the potential for rapid dissemination if weaponized, with high-mortality potential, small infective doses, and significant environmental stability. 12, 13 data concerning clinical incubation periods, transmission characteristics, and infection-control procedures for agents of bioterrorism are provided in table 79 -3. syndromic differential diagnoses for select clinical presentations are detailed in table 79 -4. anthrax results from infection with bacillus anthracis, a gram-positive, spore-forming, rod-shaped organism that exists in its host as a vegetative bacillus and in the environment as a spore. details of the microbiology and pathogenesis of anthrax are found in chapter 124. in nature, anthrax is a zoonotic disease of herbivores that is prevalent in many geographic regions; sporadic human disease results from environmental or occupational contact with endospore-contaminated animal products. 14 the cutaneous form of anthrax is the most common presentation; gastrointestinal and inhalational forms are exceedingly rare in naturally acquired disease. an additional form, injectional anthrax, represents a potentially lethal, deep soft-tissue infection that has been well described in injection heroin users in several western european countries. 14a cutaneous anthrax occurred regularly in the first half of the twentieth century in association with contaminated hides and wools used in the garment industry, but it is uncommonly seen in current-day industrialized countries because of importation restrictions. the last-known fatal case of naturally occurring inhalational anthrax in the united states occurred in 1976, when an individual was exposed to imported wool from pakistan. 15 case reports of naturally occurring anthrax do occur within the united states, although they are rare. 16 it has been previously hypothesized that large-scale bioterrorism with anthrax would involve aerosolized endospores with resultant inhalational disease, but the 2001 attacks in the united states illustrate the difficulties in predicting modes and outcomes in bioterrorism. these attacks were on a relatively small scale, and nearly 40% of the confirmed cases were of the cutaneous variety. 17 the serious morbidity and mortality of anthrax is instead related to inhalational disease, as was the case in the sverdlovsk outbreak in 1979. as a result, planning for larger-scale events with aerosolized agent is warranted given the high-mortality cost of an exposure to this more weaponized form of anthrax. the clinical presentations and differential diagnoses of cutaneous and inhalational anthrax are described in table 79 of cutaneous anthrax may be similar in appearance to other lesions, including cutaneous forms of other agents of bioterrorism; however, it may be distinguished by epidemiologic, as well as certain clinical, features. anthrax is traditionally a painless lesion, unless secondarily infected, and is associated with significant local edema. the bite of loxosceles reclusa, the brown recluse spider, shares many of the local and systemic features of anthrax but is typically painful from the outset and lacks such significant edema. 18 cutaneous anthrax is associated with systemic disease, and it carries an associated mortality in up to 20% of untreated cases, although with appropriate antimicrobial therapy mortality is less than 1%. 14 once the inhaled endospores reach the terminal alveoli of the lungs-generally requiring particle sizes of 1 to 5 î¼m-they are phagocytosed by macrophages and transported to regional lymph nodes. here the endospores germinate into vegetative bacteria and subsequently disseminate hematogenously. 13 spores may remain latent for extended periods in the host, up to 100 days in experimental animal exposures. 15 this translates to prolonged clinical incubation periods after respiratory exposure to endospores. cases of inhalational anthrax occurred up to 43 days after exposure in the sverdlovsk accident, although the average incubation period is thought to be 2 to 10 days, perhaps influenced by exposure dose. 13, 15 before the u.s. anthrax attacks in october 2001, most of the clinical data concerning inhalational anthrax derived from sverdlovsk, the largest outbreak recorded. although there is much overlap between the clinical manifestations noted previously and those observed during the recent outbreak, data that are more detailed are available from the recent u.s. experience. there were 11 confirmed persons with inhalational anthrax, 5 (45%) of whom died. this contrasts with a casefatality rate of greater than 85% reported from sverdlovsk with an estimated 100 deaths. the reliability of reported data from this outbreak is questionable, given soviet documentation, but a majority of victims were located downwind of the ill-fated weapons plant. 15, 18 patients almost on average present of 3.3 days after symptom onset with fevers, chills, malaise, myalgias, nonproductive cough, chest discomfort, dyspnea, nausea or vomiting, tachycardia, peripheral neutrophilia, and liver enzyme elevations. 11, 19, 20 many of these findings are nondiagnostic, and they overlap considerably with those of influenza and other common viral respiratory tract infections. recently compiled data suggest that shortness of breath, nausea, and vomiting are significantly more common in anthrax, whereas rhinorrhea is uncommonly seen in anthrax but noted in the majority of viral respiratory infections, an important clinical distinction. 21 other common clinical manifestations of inhalational anthrax include abdominal pain, headache, mental status abnormalities, and hypoxemia. abnormalities on chest radiography appear to be universally present, although these may only be identified retrospectively in some cases. pleural effusions are the most common abnormality, although radiographs may demonstrate patchy infiltrates, consolidation, and/or mediastinal adenopathy. the latter is thought to be an early indicator of disease, but computed tomography appears to provide greater sensitivity compared with chest radiographs for this finding. the clinical manifestations of inhalational anthrax generally evolve to a fulminant presentation with progressive respiratory failure and shock. b. anthracis is routinely isolated in blood cultures if obtained before the initiation of antimicrobials. pleural fluid is typically hemorrhagic; the bacteria can either be isolated in culture or documented by antigen-specific immunohistochemical stains of this material in the majority of patients. 11 in the five fatalities in the u.s. series, the average time from hospitalization until death was 3 days (range, 1 to 5 days), which is consistent with other reports of the clinical virulence of this infection. autopsy data typically reveal hemorrhagic mediastinal lymphadenitis and disseminated, metastatic infection. pathology data from the sverdlovsk outbreak confirm meningeal involvement, typically hemorrhagic meningitis, in 50% of disseminated cases. 22 the diagnosis of inhalational anthrax should be entertained in the setting of a consistent clinical presentation in the context of a known exposure, a possible exposure, or epidemiologic factors suggesting bioterrorism (e.g., clustered cases of a rapidly progressive illness). the diagnosis should also be considered in a single individual with a clinical illness consistent with anthrax exposure in the absence of another etiology. the early recognition and prompt treatment of inhalational anthrax is likely associated with a survival advantage. 11 therefore the emergency physician should promptly initiate empiric antimicrobial therapy if infection is clinically suspected. combination parenteral therapy is appropriate in the ill person for a number of reasons: to cover the possibility of antimicrobial resistance, to target specific bacterial functions (e.g., the theoretical effect of clindamycin on toxin production), to ensure adequate drug penetration into the central nervous system, and perhaps to favorably affect survival. 11 drainage of pleural effusions is indicated to reduce toxin burden. detailed therapeutic and post-exposure prophylaxis recommendations have been recently reviewed elsewhere. 22a a monoclonal antibody targeted at the protective antigen component of anthrax toxin, raxibacumab, is available for the adjunctive treatment of systemic anthrax. 14a in the future, it is likely that novel therapies such as toxin inhibitors or cell-specific receptor antagonists will be available to treat anthrax post exposure. 23 detailed therapeutic and postexposure prophylaxis recommendations for adults, children, and special groups have been recently reviewed elsewhere. 15 with regard to postexposure prophylaxis, the anthrax vaccine adsorbed is effective for prevention of cutaneous anthrax in human clinical trials, as well as preventing inhalational disease after aerosol challenge in nonhuman primates. 21 current studies are investigating the efficacy of this vaccine when paired with antibiotics in the postexposure period. for preexposure prophylaxis, the vaccine is generally very safe, but it requires five doses over 18 months, with the need for annual boosting for ongoing preventative immunity. 24 preexposure use of the vaccine is currently limited to individuals at high risk for anthrax exposure, such as military personnel and specific laboratory workers. although not currently available, additional research into second-generation anthrax vaccines is aimed to generate a more easily distributed means of mass prophylaxis following an anthrax exposure. 25 the last-known naturally acquired case of smallpox occurred in somalia in 1977. in one of the greatest triumphs of modern medicine, smallpox was officially certified as having been eradicated in 1980, the culmination of a 12-year intensive campaign undertaken by the who. 26 however, because of concerns that variola-virus stocks may have either been removed from or sequestered outside of their officially designated repositories, smallpox is considered a potential and certainly dangerous agent of bioterrorism. multiple features make smallpox an attractive biological weapon and ensure that any reintroduction into human populations would be a global public health catastrophe: it is stable in aerosol form, has a low infective dose, is associated with up to a 30% case-fatality rate, and has a large vulnerable target population because civilian vaccination was terminated in 1972. smallpox is also especially dangerous because secondary attack rates among unvaccinated close contacts are estimated at 37% to 88% and are only further amplified by the lack of vaccine-induced immunity and a lack of naturally circulating virus to induce low-level booster exposures. 27 because of the successful eradication, preexposure vaccination is currently limited to specific military and laboratory professionals. there are currently no antiviral therapies of proven effectiveness against this pathogen. after an incubation period of 7 to 17 days (average 10 to 12 days), patients will develop a prodrome of fever, rigors, headache, and backache that may last 2 to 3 days. this is followed by a centrifugally distributed eruption that generalizes as it evolves through macular, papular, vesicular, and pustular stages in synchronous fashion over approximately 8 days, with umbilication in the latter stages. enanthem in the oropharynx typically precedes the exanthem by 24 to 48 hours. the rash typically involves the palms and soles early in the course of the disease. the pustules begin crusting during the second week of the eruption; separation of scabs is usually complete by the end of the third week. the differential diagnosis of smallpox is delineated in table 79-4 . historically, varicella and drug reactions have posed the greatest diagnostic dilemmas; this would likely be further complicated by the absence of this clinical disease and therefore experience in its diagnosis for the past 40 years. 21 smallpox is transmitted person to person by respiratory droplet nuclei and (although less commonly) by contact with lesions or contaminated fomites. airborne transmission by fine-particle aerosols has also been documented under certain conditions. 21 the virus is communicable from the onset of the enanthem until all of the scabs have separated, although patients are thought to be most contagious during the first week of the rash because of high titers of replicating virus in the oropharynx. household members, other face-to-face contacts, and health care workers have traditionally been at highest risk for secondary transmission, given their proximity to infected individuals during the highly infectious period. as a result, patients with signs and symptoms concerning for smallpox should be placed in negative-pressure rooms with contact and airborne precautions to minimize this risk. those not requiring hospital-level care should remain isolated at home to avoid infecting others in public places. the suspicion of a single smallpox case should prompt immediate notification of local public health authorities and the hospital epidemiologist. containment of smallpox is predicated on the "ring vaccination" strategy, which was successfully deployed in the who global eradication campaign. this strategy mandates the identification and immunization of all directly exposed persons, including close contacts, health care workers, and laboratory personnel. vaccination, if deployed within 4 days of infection during the early incubation period, can significantly attenuate or prevent disease and may favorably affect secondary transmission. 21 because the occurrence of even a single case of smallpox would be tantamount to bioterrorism, an immediate epidemiologic investigation is necessary to establish a biological perimeter and trace initially exposed individuals for ring vaccination purposes. botulism is an acute neurologic disease caused by clostridium botulinum, which occurs both sporadically and in focal outbreaks throughout the world related to wound contamination by the bacterium or the ingestion of the foodborne toxin. a detailed discussion of botulism is found in chapter 154. aerosolized forms of the toxin are fortunately a rare mode of acquisition in nature, but they have been weaponized for use in bioterrorism. 5 botulinum toxin is considered the most toxic molecule known; it is lethal to humans in very minute quantities. it is estimated that a single gram of concentrated clostridium botulinum neurotoxin could kill up to 1 million otherwise healthy individuals. 29 the toxin functions by blocking the release of the neurotransmitter acetylcholine from presynaptic vesicles, thereby inhibiting muscle contraction. 30 botulism presents as an acute, afebrile, symmetric, descending, and flaccid paralysis. the disease manifests initially in the bulbar musculature and is unassociated with mental status or sensory changes. fatigue, dizziness, dysphagia, dysarthria, diplopia, dry mouth, dyspnea, ptosis, ophthalmoplegia, tongue weakness, and facial muscle paresis are early findings seen in more than 75% of cases. progressive muscular involvement leading to respiratory failure ensues. the clinical presentations of foodborne and inhalational botulism are indistinguishable in experimental animals. 24 fortunately, outside of the toxin itself being utilized for bioterrorism, botulism is not spread directly from person to person. typically, these patients will recover with supportive care in weeks to months. the diagnosis of botulism is largely based on epidemiologic and clinical features and the exclusion of other possibilities (table 79-4) . clinicians should recognize that any single case of botulism could be the result of bioterrorism or could herald a larger-scale "natural" outbreak. a large number of epidemiologically unrelated, multifocal cases should be clues to an intentional release of the agent, either in food sources, water supplies, or as an aerosol. the mortality from foodborne botulism has declined from 60% to 6% over the last four decades, likely because of improvements in supportive care and mechanical ventilation. because the need for the latter may be prolonged, limited resources (e.g., mechanical ventilators) would likely be exceeded in the event of a large-scale bioterrorism event. treatment with an equine antitoxin, available in limited supply from the cdc, may ameliorate disease if given early. there is no currently available vaccine. plague, a disease responsible for multiple epidemics throughout human history, is caused by the gram-negative pathogen yersinia pestis. this pathogen is found in a variety of forms in the natural world. it is extensively covered in chapter 125. plague is endemic in parts of southeast asia, africa, and the western united states. aerosolized preparations of the agent, the expected vehicle in bioterrorism, would be predicted to result in cases of primary pneumonic plague outside of endemic areas. additional forms of the disease, such as bubonic and septicemic plague, are also concerning from a bioterrorism perspective. primary pneumonic plague classically presents as an acute, febrile, pneumonic illness with prominent respiratory and systemic symptoms. patients will often endorse gastrointestinal symptoms and purulent sputum production, with variable levels of reported hemoptysis. 31 chest x-rays will typically show patchy, bilateral, multilobar infiltrates or consolidations. unlike other forms of community-acquired pneumonia, in the absence of appropriate treatment, there may be rapid progression to respiratory failure, vascular collapse, purpuric skin lesions, necrotic digits, and death. the differential diagnosis for these symptoms including rapidly progressive pneumonia is very broad as noted in table 79 -4. plague is suggested by the characteristic small gramnegative coccobacillary forms found in stained sputum specimens with the bipolar uptake ("safety pin") of giemsa or wright stain. 32 culture confirmation is necessary to establish the diagnosis; the microbiology laboratory should be notified in advance if plague is suspected because special techniques and precautions must be employed. of note, initial gram staining of samples can often be negative despite positive culture in y. pestis detection. serologic testing is also possible if the aforementioned studies are persistently negative. treatment recommendations for plague have been reviewed elsewhere. 26 pneumonic plague can be transmitted from person to person by respiratory droplet nuclei, thus placing close contacts, other patients, and health care workers at risk for secondary infection. prompt recognition and treatment of this disease, appropriate deployment of postexposure prophylaxis, and early institution of droplet precautions will help to interrupt secondary transmission. both live and attenuated plague vaccines exist; however, these are not currently approved for commercial use in the united states. high-risk populations, including laboratory and military personnel, may receive a formaldehyde-killed version of the vaccine as prophylaxis in certain situations. 13 fortunately, new recombinant vaccines are currently in development, although some parts of the world continue to use live versions of the vaccine. 33 tularemia francisella tularensis, the causative agent of tularemia, is another small gram-negative coccobacillus with potential to cause a primary pneumonic presentation if delivered as an aerosol agent of bioterrorism. this bacterium is commonly found in smaller mammals, most classically hares and rabbits. humans serve as an accidental host; typically, natural infections occur via insect bites, consuming infected animal products, or direct contact with infected domesticated animals. 34 the causative bacteria can be transmitted between humans by close contact via mucous membrane contact, cutaneous inoculation, and inhalation if patients are exposed to aerosolized forms of the bacteria. 14 pulmonary tularemia presents with the abrupt onset of a febrile, systemic illness with prominent upper-respiratory symptoms of a highly variable nature. patients may exhibit inconsistent development of pneumonia, hilar adenopathy, hemoptysis, pulse-temperature dissociation, malaise, and progression toward respiratory failure and death in excess of 30% of those who do not receive appropriate therapy. 35 the diagnosis is generally based on clinical features after other agents are ruled out, but again it requires a high level of clinical suspicion. confirmatory serology using various immunologic assays is currently available. 13 laboratory personnel should be notified in advance if tularemia is suspected because the organism can be very infectious under culture conditions. this agent is discussed in depth in chapter 126. moreover, treatment typically consists of antibiotic therapy with streptomycin or gentamicin, with an estimated overall mortality after treatment of only 1%. 13 a live attenuated vaccine against tularemia exists; however, it is not currently available for human use in the united states. 36 tularemia remains a significant concern, given the lack of current vaccine, especially when coupled with the high infectivity and mortality of pulmonary tularemia. the agents of viral hemorrhagic fevers are members of four distinct families of ribonucleic acid viruses that cause clinical syndromes with overlapping features: fever, malaise, headache, myalgias, prostration, mucosal hemorrhage, and other signs of increased vascular permeability with circulatory dysregulation. unfortunately, they are all capable of leading to shock and multiorgan system failure in advanced cases. 37 specific agents are also associated with specific target organ effects, although each has a propensity to damage vascular endothelium. these pathogens, discussed in detail in chapters 142 to 145, include ebola, marburg, lassa fever, rift valley fever, and congo-crimean hemorrhagic fever. hemorrhagic fever viruses have been viewed as being emerging infections because of their sporadic occurrence in focal outbreaks throughout the world; the ongoing epidemic of ebola hemorrhagic fever in west africa has resulted in more than 25,000 cases and 10,000 deaths since 2014. 36a often in novel outbreak situations, these severe effects of these viruses on humankind are thought to be the results of human intrusion into a viral ecologic niche. they are concerning potential weapons of bioterrorism because they are highly infectious in aerosol form, are transmissible in health care settings, cause high morbidity and mortality, and are purported to have been successfully weaponized. 9 blood and other bodily fluids from infected patients are extremely infectious, and person-to-person airborne transmission may occur, as well. as a result, strict contact and airborne precautions should be instituted if viral hemorrhagic fevers are implicated in a terrorism event. 28 the diagnosis of viral hemorrhagic fevers is complicated, especially in a potential bioterrorist attack, which would lack a known exposure, or following recent travel to africa. microbiology studies and immunological testing are difficult to perform routinely, and often require evaluation by cdc laboratories. 38 treatment is largely supportive, and it includes the early use of vasopressors as needed. ribavirin is effective against some forms of viral hemorrhagic fevers but not those caused by ebola and marburg viruses. for a majority of these diseases, the treatment is largely supportive therapy. nonetheless, ribavirin should be initiated empirically in patients presenting with a syndrome consistent with viral hemorrhagic fever until the exact etiology is confirmed. even though there are vaccines available for similar diseases, such as yellow fever and argentine hemorrhagic fever, there are no current options for preexposure vaccination for viral hemorrhagic fevers. this paired with the highly infectious nature and significant mortality rates make this category of viruses worrisome potential agents of bioterrorism. the approach to the management of diseases of bioterrorism must be broadened to include children, pregnant women, and immunocompromised persons. specific recommendations for treatment and prophylaxis of these special patient groups for selected bioterrorism agents have been recently reviewed. 14,26,27 a general approach requires an assessment of the risk of certain drugs or products in select populations versus the potential risk of the infection in question, accounting for extent of exposure and the agent involved. the issue extends to immunization because certain vaccines, such as smallpox, pose higher risk to these special groups than to others. this will affect mass vaccination strategies and will likely warrant case-by-case decisions. of note, the prevalence of antivaccine sentiments has implications with regard to global biosecurity. a decline in herd immunity against a vaccine-preventable communicable disease could leave even a medically prepared society vulnerable to a terrorist-introduced agent previously well controlled with prophylactic vaccinations. this will be yet another special population to consider in the event of a mass casualty bioterrorist attack. an often overlooked but vitally important issue in bioterrorism is that of psychosocial sequelae. these may take the form of acute anxiety reactions and exacerbations of chronic psychiatric illness during the stress of the event, or posttraumatic stress disorder (ptsd) in its aftermath. nearly half of the emergency department visits during the gulf war missile attacks in israel in 1991 were related to acute psychological illness or exacerbations of underlying problems. 39 data from recent acts of terrorism in the united states suggest that ptsd may develop in as many as 35% of those affected by the events. 40 in the early period after the 9/11 attacks in new york, ptsd and depression were nearly twice as prevalent as in historical control subjects. 41 although close proximity to the events and personal loss were directly correlated with ptsd and depression, respectively, there was a substantial burden of morbidity among those indirectly involved. among individuals working on capitol hill following the 2001 anthrax scare, 27% were diagnosed with ptsd, with up to 55% diagnosed with any variety of psychiatric disorder. moreover, a majority of these patients were not adherent with antibiotics prescribed, perhaps because of a newfound lack of trust in the health care system. 42 although not always clinically apparent, the psychological effect of a bioterrorism event is certainly a significant and important consideration for ongoing public health management strategies following any biological threat or terrorist attack. the response to bioterrorism is unique among wmds because it necessitates consequence management that is common to all disasters, as well as the application of basic infectious diseases principles. disease surveillance, diagnosis, infection control, antimicrobial therapy, postexposure prophylaxis, and mass preventative vaccinations are all important considerations when managing a bioterrorism event. for these reasons, physicians are likely first responders to bioterrorism and will be expected to be reliable sources of information for their patients, colleagues, and public health authorities. 43 a remaining number of potential pitfalls regarding disasters involving a biological attack must be identified and managed to optimize the public health response. as alluded to above, the clinical latency period between exposure to an agent and the manifestation of signs and symptoms is approximately days to weeks with most of the cdc category a, b, or c agents. thus, early diagnoses of the first cases are likely to prove problematic and require heightened clinical vigilance, a difficult task considering a majority of these agents are rarely observed in the developed world. 44 even after initial victims have been diagnosed, communications among hospitals and other health care institutions on a local, regional, national, and international level will be essential to help define the epidemiology and identify possible exposure sources. given the extent and ease of rapid individual movement within our globalized world, clinical presentations from a point-source biological attack could occur in widely disparate geographic locations. additionally it is possible that a terrorist attack would be multifocal in any case, with components of wmds paired with biological weapons for maximum effect. a fundamental and consistent epidemiologic approach using case definitions, case identification, surveillance, and real-time communications is necessary, whether the event is a malicious attack, emergent from nature, or of unknown etiology. 45 other potential bioterrorism management pitfalls reside in the arena of diagnostic techniques, treatment, and prevention of disease related to biological agents. although an active area of research, the development of field-ready and highly predictive rapid screening tests for many agents of bioterrorism has not yet progressed to the point at which such assays are approved by the u.s. food and drug administration and available in a "point-of-care" format. treatment and prevention issues such as the absence of effective therapies for many forms of viral hemorrhagic fevers, shortages in the availability of multivalent antitoxin for botulism, projected shortages in the availability of mechanical ventilators to manage a large-scale botulism attack, lack of human data regarding the use of antiviral agents in smallpox, and the unfavorable toxicity profiles of some currently available smallpox vaccines remain unresolved but active areas of research. emerging molecular biology techniques capable of producing genetically altered pathogens with "designer" phenotypes including antimicrobial or vaccine resistance add additional layers of complexity to an already multifaceted problem. as was vividly illustrated in the 2003 severe acute respiratory syndrome epidemic and previously well recognized when smallpox occurred with regularity, transmission of infection of potential bioterrorism agents within hospitals is common and difficult to control. 46, 21 health care workers, our first line of defense against an attack using biological agents, remain at significant occupational risk. research in the field of bioterrorism recognition has demonstrated a perceived weakness among clinicians in recognition of category-a infectious agents. 47 as pathogens of bioterrorism are not frequently encountered in daily practice, they often fall low on the differential without clinician knowledge of an insidious local mass casualty event. 48 clearly, awareness of a recent local event heightens clinical suspicion, but it is imperative for the front-line clinician to recognize, report, and initiate treatment of affected patients. this will only serve to facilitate the initial containment and facilitate rapid disaster-protocol activation. early recognition and initiation of a prompt, unified response will remain the primary challenge for all health care providers in the current era of bioterrorism. biological warfare and bioterrorism: a historical review biological warfare: a historical perspective factories of death: japanese biological warfare, 1932-45, and the american cover-up a large community outbreak of salmonellosis caused by intentional contamination of restaurant salad bars world health organization. health aspects of chemical and biological weapons: report of a who group of consultants. geneva: world health organization germs: biological weapons and america's secret war biological and chemical terrorism: strategic plan for preparedness and response united states department of state. patterns of global terrorism bioterrorism-related inhalational anthrax: the first 10 cases reported in the united states medical aspects of chemical and biological warfare. textbook of military medicine series. part i, warfare, weaponry and the casualty medical aspects of bio-terrorism biowarfare and bioterrorism novel approaches to the treatment of systemic anthrax anthrax as a biological weapon: medical and public health management investigation of inhalation anthrax case, united states anthrax as a biological weapon, 2002: updated recommendations for management cutaneous anthrax associated with microangiopathic hemolytic anemia and coagulopathy in a 7-monthold infant fatal inhalational anthrax in a 94-year-old connecticut woman considerations for distinguishing influenza-like illness from inhalational anthrax anthrax vaccine: evidence for safety and efficacy against inhalational anthrax tackling anthrax centers for disease control and prevention expert panel meetings on prevention and treatment of anthrax in adults pathology of inhalational anthrax in 42 cases from the sverdlovsk outbreak of 1979 use of anthrax vaccine in the united states: recommendations of the advisory committee on immunization practices (acip) phase 3 trial evaluating the immunogenicity and safety of a three-dose biothrax( â® ) regimen for postexposure prophylaxis in healthy adults smallpox and its eradication. geneva: world health organization diagnosis and management of smallpox smallpox response plan and guidelines (version 3.0) botulinum toxin: bioweapon & magic drug botulinum toxin as a biological weapon: medical and public health management occupational plague plague as a biological weapon: medical and public health management prophylaxis and therapy of plague tularemia: the disease and the weapon tularemia as a biological weapon: medical and public health management immunotherapy for tularemia ebola outbreak in west africa-case counts hemorrhagic fever viruses as biological weapons: medical and public health management polio outbreak leads to calls for a "vaccination ceasefire" in syria medical aspects of the iraqi missile attacks on israel post-traumatic stress disorder psychological sequelae of the september 11 terrorist attacks in new york city exposure to bioterrorism and mental health response among staff on capitol hill bioterrorism and physicians infectious diseases bioterrorism and physicians public health measures to control the spread of the severe acute respiratory syndrome during the outbreak in toronto critical challenges ahead in bioterrorism preparedness training for clinicians recognition of community-acquired anthrax: has anything changed since key: cord-013470-1obua17m authors: saylan, yeşeren; akgönüllü, semra; denizli, adil title: plasmonic sensors for monitoring biological and chemical threat agents date: 2020-10-15 journal: biosensors (basel) doi: 10.3390/bios10100142 sha: doc_id: 13470 cord_uid: 1obua17m sensors are excellent options owing to their ability to figure out a large number of problems and challenges in several areas, including homeland security, defense, medicine, pharmacology, industry, environment, agriculture, food safety, and so on. plasmonic sensors are used as detection devices that have important properties, such as rapid recognition, real-time analysis, no need labels, sensitive and selective sensing, portability, and, more importantly, simplicity in identifying target analytes. this review summarizes the state-of-art molecular recognition of biological and chemical threat agents. for this purpose, the principle of the plasmonic sensor is briefly explained and then the use of plasmonic sensors in the monitoring of a broad range of biological and chemical threat agents is extensively discussed with different types of threats according to the latest literature. a conclusion and future perspectives are added at the end of the review. two major threat agents are utilized in attacks-biological and chemical threat agents-and the detection of these agents is a modern subject of increasing impact and interest [1] . the nobel peace prize awarded to the organization for the prohibition of chemical weapons in 2013 "for its extensive efforts to eliminate chemical weapons" sheds light on modern-age issues such as this [2] . bioterrorism is a deliberate use of harmful biological threats agents that provoke disease or death in plants, animals, and humans [3] . unfortunately, these dangerous agents can spread through the air, water, or soil [4] . moreover, they can be modified to increase their disease-causing properties that make them resistant to existing drugs and improve their ability to spread to the environment [5, 6] . compared to chemical threat agents, biological threat agent's production is much cheaper and also the danger area and expected loss of life is more effective than chemical threat agents in a military attack [7] . the infection dose-the amount of organism required for an infection outbreak-is different for each threat agent [8] . the risk ratio of each threat agent is given not only by infection dose, but also by the natural propagation pathway, aerosol or water stability, and the possibility of spore formation in the case of bacteria [9] . in particular, biological threat agents obtain toxic substances that are relatively easy and inexpensive, and they can easily spread and cause fear and panic beyond real physical damage [10, 11] . several analytical techniques have been used to detect the various biological and chemical threat agents over the years [12] [13] [14] [15] [16] [17] [18] [19] [20] . researchers have been using analytical technique combinations including extraction, mass spectroscopy, liquid chromatography, and sensors in numerous situations due to the different needs that are hard to detect into a single suitable monitoring system and methodology [21] [22] [23] [24] [25] [26] . there is a need for the development of novel sensing methods with smart detection capabilities in defense and homeland security applications for the early detection of biological biosensors 2020, 10, 142 2 of 15 and chemical threat agents. a useful and commercially attractive sensor systems for accurate and rapid detection of biological and chemical threat agents has to satisfy some important properties [27] . various recent studies indicate that plasmonic sensors can be a key platform for monitoring biological and chemical threat agents owing to their combination of different charming properties such as sensitivity, rapid, unlabeled, low cost, real time, and portability [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] . on the other hand, despite a fast-growing quantity of articles, there is no up-to-date review of plasmonic sensors for monitoring biological and chemical threat agents. this review purposes to fill this gap by summing up the most recent advances in this subject. following an introduction to the principle of sensors and plasmonic sensors (section 2), the importance of the monitoring of these dangerous agents (section 3). then, a wide overview is discussed about the performance and characteristics of published plasmonic sensors for the detection of biological (section 3.1) and chemical (section 3.2) threat agents. on the whole, it was hoped that this review article will conduct and inspire all researchers for further plasmonic sensors developments and applications. a sensor is an analytical tool that has three main modules: a sensing receptor, a transducer, and a detector [39] . as demonstrated in figure 1 , a target analyte principally binds to the recognition element (receptor), and the sensing component specifically recognizes the analyte by a reaction, specific adsorption, or another process as physical and/or chemical and then the transducer translates changes to a quantifiable signal measured by the digital detector module [40] . biosensors 2020, 10, x for peer review 2 of 15 detection capabilities in defense and homeland security applications for the early detection of biological and chemical threat agents. a useful and commercially attractive sensor systems for accurate and rapid detection of biological and chemical threat agents has to satisfy some important properties [27] . various recent studies indicate that plasmonic sensors can be a key platform for monitoring biological and chemical threat agents owing to their combination of different charming properties such as sensitivity, rapid, unlabeled, low cost, real time, and portability [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] . on the other hand, despite a fast-growing quantity of articles, there is no up-to-date review of plasmonic sensors for monitoring biological and chemical threat agents. this review purposes to fill this gap by summing up the most recent advances in this subject. following an introduction to the principle of sensors and plasmonic sensors (section 2), the importance of the monitoring of these dangerous agents (section 3). then, a wide overview is discussed about the performance and characteristics of published plasmonic sensors for the detection of biological (section 3.1) and chemical (section 3.2) threat agents. on the whole, it was hoped that this review article will conduct and inspire all researchers for further plasmonic sensors developments and applications. a sensor is an analytical tool that has three main modules: a sensing receptor, a transducer, and a detector [39] . as demonstrated in figure 1 , a target analyte principally binds to the recognition element (receptor), and the sensing component specifically recognizes the analyte by a reaction, specific adsorption, or another process as physical and/or chemical and then the transducer translates changes to a quantifiable signal measured by the digital detector module [40] . sensors have numerous advantages, including user-friendly operation, exceptional performance, rapid response, minimum sample preparation and processing, portability, high sensitivity and specificity, relatively compact size, and real-time analysis that provide many advantages compared to standard analytical methods [41] [42] [43] [44] [45] . furthermore, they have a great impact on the environment, medical, safety, and many other applications [46] [47] [48] . detection elements are very diverse. they include conductivity, density, an electromagnetic radiation phase, electric current, mass, viscosity, electrical potential, temperature, and impedance [49] . in recent years, plasmonic sensors have attracted broad attention due to their perfect electromagnetic control ability. owing to the physical and optical properties of the plasmonic sensors, they have become the most famous device for working [50] . these unique properties make them an indispensable real-time and labelfree platform for target analyte detection. the current focus of the field is to further be growing the sensitivity to detect low concentrations or even single molecules in diluted solutions [51] . so, they are promising for ultrasensitive detection and have received increasing applications in biological and chemical analysis, food control, clinical diagnostics, biomedical research, etc. [52] . plasmonic sensors sensors have numerous advantages, including user-friendly operation, exceptional performance, rapid response, minimum sample preparation and processing, portability, high sensitivity and specificity, relatively compact size, and real-time analysis that provide many advantages compared to standard analytical methods [41] [42] [43] [44] [45] . furthermore, they have a great impact on the environment, medical, safety, and many other applications [46] [47] [48] . detection elements are very diverse. they include conductivity, density, an electromagnetic radiation phase, electric current, mass, viscosity, electrical potential, temperature, and impedance [49] . in recent years, plasmonic sensors have attracted broad attention due to their perfect electromagnetic control ability. owing to the physical and optical properties of the plasmonic sensors, they have become the most famous device for working [50] . these unique properties make them an indispensable real-time and label-free platform for target analyte detection. the current focus of the field is to further be growing the sensitivity to detect low concentrations or even single molecules in diluted solutions [51] . so, they are promising for ultrasensitive detection and have received increasing applications in biological and chemical analysis, food control, clinical diagnostics, biomedical research, etc. [52] . plasmonic sensors also focus on the analysis of a change of intensity in the optical characteristics of the transducer metal surface when the target analyte is captured by the recognition element ( figure 2 ). they are divided into many subclasses such as resonance, fluorescence, refraction, reflection, dispersion, phosphorescence, infrared absorption, raman scattering, and chemiluminescence [53] [54] [55] . biosensors 2020, 10, x for peer review 3 of 15 also focus on the analysis of a change of intensity in the optical characteristics of the transducer metal surface when the target analyte is captured by the recognition element ( figure 2 ). they are divided into many subclasses such as resonance, fluorescence, refraction, reflection, dispersion, phosphorescence, infrared absorption, raman scattering, and chemiluminescence [53] [54] [55] . plasmonic sensors belong to the group of optical-based affinity sensors. when the target analyte is captured by the ligand, it causes a measurable signal. various sensor surfaces with immobilized ligands are commercially available and/or can be custom designed and fabricated [56] . in general, plasmonic sensors are created of sensing elements that contain metal or metal-dielectric nanostructures supporting surface plasmons and recognition elements that can selectively bind a target analyte. the enlightenment of the sensing element by light produces surface plasmons on the structures generating the electromagnetic field that is highly concentrated at the surface of the structures. when a sample containing target analyte is brought into contact with the sensor, the capture of the target analyte by recognition element immobilized on the surface of the sensing element gives rise to a change in the refractive index in the region close to the surface [57] . the surface plasmons are electron oscillations that take place at the limit of a noble metal (plasmonic materials) and different media as an aqueous solution (dielectric material). changes in the dielectric constant, or refractive index, near the surface of the plasmonic sensor, are measured as a change in the angle or wavelength of p-polarized light absorbed by the surface plasmon. the metal film surface, usually gold or silver, is modified with surface chemistry steps to immobilize a series of molecular recognition elements. the molecular recognition element is used to capture the specific binding partners [58] [59] [60] . many researchers, including chemists, biologists, physicists, engineers, and medical doctors, have used the sensor platforms as an original application in different areas for the development of sensors [61] [62] [63] [64] [65] . today, the early detection of a biological and chemical attack of the threats can only be analyzed with commercial methods. thus, real-time sensors are needed for the security of the community [66] . the rapid and specific detection of threat agents is a critical aspect of defense. monitoring systems must be fast, responsive, portable, and also specific to threat agents [67, 68] . interest in the quality of life is increasing all over the world. it is maintained by various factors, such as quality of life, disease control, drug development, environmental cleaning, food safety, and homeland security [69, 70] . biological threat agents are classified into three categories as depicted in table 1 . in category a, the agents can easily spread or spread from person to person. this results in high death rates and has plasmonic sensors belong to the group of optical-based affinity sensors. when the target analyte is captured by the ligand, it causes a measurable signal. various sensor surfaces with immobilized ligands are commercially available and/or can be custom designed and fabricated [56] . in general, plasmonic sensors are created of sensing elements that contain metal or metal-dielectric nanostructures supporting surface plasmons and recognition elements that can selectively bind a target analyte. the enlightenment of the sensing element by light produces surface plasmons on the structures generating the electromagnetic field that is highly concentrated at the surface of the structures. when a sample containing target analyte is brought into contact with the sensor, the capture of the target analyte by recognition element immobilized on the surface of the sensing element gives rise to a change in the refractive index in the region close to the surface [57] . the surface plasmons are electron oscillations that take place at the limit of a noble metal (plasmonic materials) and different media as an aqueous solution (dielectric material). changes in the dielectric constant, or refractive index, near the surface of the plasmonic sensor, are measured as a change in the angle or wavelength of p-polarized light absorbed by the surface plasmon. the metal film surface, usually gold or silver, is modified with surface chemistry steps to immobilize a series of molecular recognition elements. the molecular recognition element is used to capture the specific binding partners [58] [59] [60] . many researchers, including chemists, biologists, physicists, engineers, and medical doctors, have used the sensor platforms as an original application in different areas for the development of sensors [61] [62] [63] [64] [65] . today, the early detection of a biological and chemical attack of the threats can only be analyzed with commercial methods. thus, real-time sensors are needed for the security of the community [66] . the rapid and specific detection of threat agents is a critical aspect of defense. monitoring systems must be fast, responsive, portable, and also specific to threat agents [67, 68] . interest in the quality of life is increasing all over the world. it is maintained by various factors, such as quality of life, disease control, drug development, environmental cleaning, food safety, and homeland security [69, 70] . biological threat agents are classified into three categories as depicted in table 1 . in category a, the agents can easily spread or spread from person to person. this results in high death rates and has a large impact on common health. it can also cause social disruption and public anxiety and need a special solution for health preparation. in category b, the agents are partially easy to biosensors 2020, 10, 142 4 of 15 spread. it results in a low disease rate and death and requires specific improved diagnostic capacity and disease surveillance. in category c, the agent can be designed for mass spreading due to their future availability. it is easy to produce and spread. these are potentially linked to high disease and death rates and high health effects [71] [72] [73] [74] [75] [76] [77] [78] . sharma et al. screened three monoclonal antibodies (mab1, mab2, and mab3) of ebola virus employing a surface plasmon resonance (spr) sensor to pick a suitable antibody for against a biological threat agent [79] . they first modified a gold chip surface with 4-mercaptobenzoic acid and then immobilized with the recombinant nucleoprotein of the ebola virus. they calculated the affinity constants as 809 nm, 350 pm, and 52 pm for mab1, mab2, and mab3 of the ebola virus interaction, respectively. they concluded the high affinity of mab3 with the ebola virus and confirmed this result with elisa. finally, the limit of the detection value of the plasmonic sensor is calculated as 0.5 pg/ml for mab3. sikarwar et al. prepared an spr sensor using 4-mercaptobenzoic acid-modified gold for brucella melitensis detection using its complementary dna targets with two different probes of the is711 gene [80] . they performed kinetics and thermodynamic analysis, and the results reflected that complementary dna targets and probe 1 are a more effective interaction than probe 2 ( figure 3) . furthermore, they carried out the real serum samples' analysis and reported the applicability of this plasmonic sensor for brucella melitensis detection in less than 10 min. furthermore, they carried out the real serum samples' analysis and reported the applicability of this plasmonic sensor for brucella melitensis detection in less than 10 min. patel et al. provided an spr sensor for botulinum neurotoxin type a light chain detection to show the feasibility of the newton photonics-based sensor [81] . the limit of detection was calculated as 6.76 pg/ml and they established that the detection sensitivity of the plasmonic sensor is comparable to the traditional mouse ld50 bioassay. versiani et al. developed a gold nanorod functionalized plasmonic sensor based on localized surface plasmon resonance (lspr) for dengue virus protein detection [82] . they first linked the gold nanorods to α-lipoic acid (α-la) by interaction with the thiol group and then 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (edc) and n-hydroxysuccinimide (nhs) are mixed to stabilize substitution reactions owing to recombinant proteins covalently bound to α-la on the gold nanorod surfaces. they used this protein-decorated gold nanorod as a sensor that shifts the emitted spectrum ( figure 4 ). they reported that this plasmonic sensor can detect one picogram of anti-dengue virus monoclonal antibodies in dengue virus-positive human sera. they also performed the crossselectivity experiments with zika-virus-infected patients and showed that this sensor can distinguish dengue virus serotype infected individual patients. patel et al. provided an spr sensor for botulinum neurotoxin type a light chain detection to show the feasibility of the newton photonics-based sensor [81] . the limit of detection was calculated as 6.76 pg/ml and they established that the detection sensitivity of the plasmonic sensor is comparable to the traditional mouse ld50 bioassay. versiani et al. developed a gold nanorod functionalized plasmonic sensor based on localized surface plasmon resonance (lspr) for dengue virus protein detection [82] . they first linked the gold nanorods to α-lipoic acid (α-la) by interaction with the thiol group and then 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (edc) and n-hydroxysuccinimide (nhs) are mixed to stabilize substitution reactions owing to recombinant proteins covalently bound to α-la on the gold nanorod surfaces. they used this protein-decorated gold nanorod as a sensor that shifts the emitted spectrum ( figure 4 ). they reported that this plasmonic sensor can detect one picogram of anti-dengue virus monoclonal antibodies in dengue virus-positive human sera. they also performed the cross-selectivity experiments with zika-virus-infected patients and showed that this sensor can distinguish dengue virus serotype infected individual patients. furthermore, they carried out the real serum samples' analysis and reported the applicability of this plasmonic sensor for brucella melitensis detection in less than 10 min. patel et al. provided an spr sensor for botulinum neurotoxin type a light chain detection to show the feasibility of the newton photonics-based sensor [81] . the limit of detection was calculated as 6.76 pg/ml and they established that the detection sensitivity of the plasmonic sensor is comparable to the traditional mouse ld50 bioassay. versiani et al. developed a gold nanorod functionalized plasmonic sensor based on localized surface plasmon resonance (lspr) for dengue virus protein detection [82] . they first linked the gold nanorods to α-lipoic acid (α-la) by interaction with the thiol group and then 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (edc) and n-hydroxysuccinimide (nhs) are mixed to stabilize substitution reactions owing to recombinant proteins covalently bound to α-la on the gold nanorod surfaces. they used this protein-decorated gold nanorod as a sensor that shifts the emitted spectrum ( figure 4 ). they reported that this plasmonic sensor can detect one picogram of anti-dengue virus monoclonal antibodies in dengue virus-positive human sera. they also performed the crossselectivity experiments with zika-virus-infected patients and showed that this sensor can distinguish dengue virus serotype infected individual patients. surface-enhanced raman spectroscopy (sers) depends on plasmonic platforms and nano-antenna, which have developed into an effective area of study [83] . sers is one of the leading techniques for label-free ultrasensitive vibrational fingerprinting of a variety of compounds. sers has been identified as key platform thanks to distinctive features such as: ultrahigh sensitivity, detection from a wide variety of matrices and quantification of multiple species in a single measurement, allowing for real-time detection in the field [84] . prakash et al. demonstrated positively charged silver/gold bimetallic nanoparticles for escherichia coli, salmonella typhimurium, and bacillus subtilis detection using surface-enhanced raman scattering (sers)-based sensor [85] . this plasmonic sensor has important advantages, such as high sensitivity, short detection time at low power, and an easy operating process. they obtained a reusable and specific cell wall fingerprint and intracellular components of these bacteria by sers that allows for the differentiation and classification of these bacteria employing multivariate analyses ( figure 5 ). biosensors 2020, 10, x for peer review 6 of 15 surface-enhanced raman spectroscopy (sers) depends on plasmonic platforms and nanoantenna, which have developed into an effective area of study [83] . sers is one of the leading techniques for label-free ultrasensitive vibrational fingerprinting of a variety of compounds. sers has been identified as key platform thanks to distinctive features such as: ultrahigh sensitivity, detection from a wide variety of matrices and quantification of multiple species in a single measurement, allowing for real-time detection in the field [84] . prakash et al. demonstrated positively charged silver/gold bimetallic nanoparticles for escherichia coli, salmonella typhimurium, and bacillus subtilis detection using surface-enhanced raman scattering (sers)-based sensor [85] . this plasmonic sensor has important advantages, such as high sensitivity, short detection time at low power, and an easy operating process. they obtained a reusable and specific cell wall fingerprint and intracellular components of these bacteria by sers that allows for the differentiation and classification of these bacteria employing multivariate analyses ( figure 5 ). wang et al. also built up an sers-based sensor for bacterial pathogen (francisella tularensis, bacillus anthracis, and yersinia pestis) detection [86] . they first dipped the lateral flow assay (lfa) strips into well-plates containing mixtures of sers nanotags (raman reporter-labeled gold nanoparticles) and broad concentrations of bacteria in buffer solution and then formed complexes that relocated by capillary action to measure and analyze raman signals ( figure 6 ). they needed a short time (15 min) and a low volume (40 μl) to obtain this highly selective detection. wang et al. also built up an sers-based sensor for bacterial pathogen (francisella tularensis, bacillus anthracis, and yersinia pestis) detection [86] . they first dipped the lateral flow assay (lfa) strips into well-plates containing mixtures of sers nanotags (raman reporter-labeled gold nanoparticles) and broad concentrations of bacteria in buffer solution and then formed complexes that relocated by capillary action to measure and analyze raman signals ( figure 6 ). they needed a short time (15 min) and a low volume (40 µl) to obtain this highly selective detection. chemical threat agents are inactive molecules that are classified into several categories in terms of the effects on the human body. even though their usage is forbidden, they are still employed for dirty aims [87] . the chemical threat agents are generally separated as the following classes: nerve agents, blood and suffocating agents, vesicant and blister agents, cytotoxic proteins, pulmonary agents, incapacitating agents, and lachrymatory agent. the first three are best known according to their high toxicity, and the researcher also utilizes their mimics for investigation (table 2) . so, the rapid monitoring of the chemical threat agents is highly necessary and urgent [88] [89] [90] [91] [92] . inactivates the enzyme acetylcholinesterase, preventing the breakdown of the neurotransmitter acetylcholine in the victim's synapses and causing both muscarinic and nicotinic effects. agents are acid-forming compounds that damage skin and respiratory system, resulting in burns and respiratory problems. cyanogen chloride hydrogen cyanide cyanide directly prevents cells from using oxygen. the cells then use anaerobic respiration, creating excess lactic acid and metabolic acidosis. song et al. investigated an lspr-based signal enhancement mechanism using photonic crystals sensing material that was modified with silica microspheres and gold nanoparticles for atrazine detection [93] . they first modified the surface of silica microspheres with gold nanoparticles and then immobilized with atrazine aptamer (containing 5′-terminal mercapto group) for specific atrazine chemical threat agents are inactive molecules that are classified into several categories in terms of the effects on the human body. even though their usage is forbidden, they are still employed for dirty aims [87] . the chemical threat agents are generally separated as the following classes: nerve agents, blood and suffocating agents, vesicant and blister agents, cytotoxic proteins, pulmonary agents, incapacitating agents, and lachrymatory agent. the first three are best known according to their high toxicity, and the researcher also utilizes their mimics for investigation (table 2) . so, the rapid monitoring of the chemical threat agents is highly necessary and urgent [88] [89] [90] [91] [92] . inactivates the enzyme acetylcholinesterase, preventing the breakdown of the neurotransmitter acetylcholine in the victim's synapses and causing both muscarinic and nicotinic effects. sulfur mustard nitrogen mustard agents are acid-forming compounds that damage skin and respiratory system, resulting in burns and respiratory problems. cyanogen chloride hydrogen cyanide cyanide directly prevents cells from using oxygen. the cells then use anaerobic respiration, creating excess lactic acid and metabolic acidosis. song et al. investigated an lspr-based signal enhancement mechanism using photonic crystals sensing material that was modified with silica microspheres and gold nanoparticles for atrazine detection [93] . they first modified the surface of silica microspheres with gold nanoparticles and then immobilized with atrazine aptamer (containing 5 -terminal mercapto group) for specific atrazine detection with the intensity change of the photonic crystals (figure 7) . they easily prepared this lspr-based photonic crystal and reached ultrasensitive detection (10 −12 g/ml) in the 0.0001 ng/ml to 500 ng/ml range without any label. detection with the intensity change of the photonic crystals (figure 7) . they easily prepared this lspr-based photonic crystal and reached ultrasensitive detection (10 −12 g/ml) in the 0.0001 ng/ml to 500 ng/ml range without any label. verma and chandra presented a nonlinear plasmonic sensor that depended on nonlinear optics combination with amalgamation chemistry and plasmonic properties of gold nanorods for mercury detection [94] . they prospered an improved limit of detection value of mercury as compared to lsprbased sensing by operating with ultra-high sensitivity in the nonlinear optics process against mercury-induced change in the plasmonic nanorod's local electric field ( figure 8 ) and achieved as low as (58 pm) the limit of detection value with high selectivity. they also reported that the mercury amount utilized in the simulation corresponded to a concentration of 10 −11 m. amirjani and haghshenas summed up recent studies about silver nanoparticle-based plasmonic sensors for heavy metal ion (co 2+ , hg 2+ , cd 2+ , pb 2+ , and cu 2+ ) detection [95] . as depicted in figure 9 , they dared to shed more light on the monitoring of lspr properties of silver nanoparticles that depended on the changes of aggregation, anti-aggregation, oxidation reduction, and dimensionalverma and chandra presented a nonlinear plasmonic sensor that depended on nonlinear optics combination with amalgamation chemistry and plasmonic properties of gold nanorods for mercury detection [94] . they prospered an improved limit of detection value of mercury as compared to lspr-based sensing by operating with ultra-high sensitivity in the nonlinear optics process against mercury-induced change in the plasmonic nanorod's local electric field ( figure 8 ) and achieved as low as (58 pm) the limit of detection value with high selectivity. they also reported that the mercury amount utilized in the simulation corresponded to a concentration of 10 −11 m. biosensors 2020, 10, x for peer review 8 of 15 detection with the intensity change of the photonic crystals (figure 7) . they easily prepared this lspr-based photonic crystal and reached ultrasensitive detection (10 −12 g/ml) in the 0.0001 ng/ml to 500 ng/ml range without any label. verma and chandra presented a nonlinear plasmonic sensor that depended on nonlinear optics combination with amalgamation chemistry and plasmonic properties of gold nanorods for mercury detection [94] . they prospered an improved limit of detection value of mercury as compared to lsprbased sensing by operating with ultra-high sensitivity in the nonlinear optics process against mercury-induced change in the plasmonic nanorod's local electric field ( figure 8 ) and achieved as low as (58 pm) the limit of detection value with high selectivity. they also reported that the mercury amount utilized in the simulation corresponded to a concentration of 10 −11 m. amirjani and haghshenas summed up recent studies about silver nanoparticle-based plasmonic sensors for heavy metal ion (co 2+ , hg 2+ , cd 2+ , pb 2+ , and cu 2+ ) detection [95] . as depicted in figure 9 , they dared to shed more light on the monitoring of lspr properties of silver nanoparticles that depended on the changes of aggregation, anti-aggregation, oxidation reduction, and dimensionalamirjani and haghshenas summed up recent studies about silver nanoparticle-based plasmonic sensors for heavy metal ion (co 2+ , hg 2+ , cd 2+ , pb 2+ , and cu 2+ ) detection [95] . as depicted in figure 9 , they dared to shed more light on the monitoring of lspr properties of silver nanoparticles that depended on the changes of aggregation, anti-aggregation, oxidation reduction, and dimensional-morphological. they mentioned that the cost-effective and excellent plasmonic feature of silver compared to gold stems from its imaginary part of the dielectric function which is close to zero in a wide range of wavelengths. morphological. they mentioned that the cost-effective and excellent plasmonic feature of silver compared to gold stems from its imaginary part of the dielectric function which is close to zero in a wide range of wavelengths. thenmozhi et al. advanced an spr sensor using the finite element method for 1, 4-dioxane, and diethanolamine detection [96] . they first coated the sensor's surface with indium titanium oxide and placed it over the analyte, which increased the plasmon excitation on the surface. as shown in figure 10 , the coupling of silica, surface plasmon polariton (spp), and the imaginary part of silica modes of maximum wavelengths have the energy distribution of silica and spp modes, respectively. then, they succeeded in the maximal spectral sensitivity (50,000 nm/riu) participated with high-resolution (4 × 10 −4 riu) and also maximum amplitude sensitivity (1266.67 riu −1 ). heleg-shabtai et al. developed the sers-based sensor made by hand-held raman spectrometers for gas-phase nerve (methylphosphonothioic acid: vx) and blister (sulfur mustard: hd) agents monitoring [97] . they comprised the sensor with gold nanoparticles that were modified onto quartz fibers and performed gas-phase experiments utilizing a homemade flow system. they also optimized different sers methods for vx and hd detection in solution and calculated the limit of detection values as 1.8 × 10 −3 μg/ml and 2.5 × 10 −3 μg/ml for hd and vx, respectively. thenmozhi et al. advanced an spr sensor using the finite element method for 1, 4-dioxane, and diethanolamine detection [96] . they first coated the sensor's surface with indium titanium oxide and placed it over the analyte, which increased the plasmon excitation on the surface. as shown in figure 10 , the coupling of silica, surface plasmon polariton (spp), and the imaginary part of silica modes of maximum wavelengths have the energy distribution of silica and spp modes, respectively. then, they succeeded in the maximal spectral sensitivity (50,000 nm/riu) participated with high-resolution (4 × 10 −4 riu) and also maximum amplitude sensitivity (1266.67 riu −1 ). biosensors 2020, 10, x for peer review 9 of 15 morphological. they mentioned that the cost-effective and excellent plasmonic feature of silver compared to gold stems from its imaginary part of the dielectric function which is close to zero in a wide range of wavelengths. thenmozhi et al. advanced an spr sensor using the finite element method for 1, 4-dioxane, and diethanolamine detection [96] . they first coated the sensor's surface with indium titanium oxide and placed it over the analyte, which increased the plasmon excitation on the surface. as shown in figure 10 , the coupling of silica, surface plasmon polariton (spp), and the imaginary part of silica modes of maximum wavelengths have the energy distribution of silica and spp modes, respectively. then, they succeeded in the maximal spectral sensitivity (50,000 nm/riu) participated with high-resolution (4 × 10 −4 riu) and also maximum amplitude sensitivity (1266.67 riu −1 ). heleg-shabtai et al. developed the sers-based sensor made by hand-held raman spectrometers for gas-phase nerve (methylphosphonothioic acid: vx) and blister (sulfur mustard: hd) agents monitoring [97] . they comprised the sensor with gold nanoparticles that were modified onto quartz fibers and performed gas-phase experiments utilizing a homemade flow system. they also optimized different sers methods for vx and hd detection in solution and calculated the limit of detection values as 1.8 × 10 −3 μg/ml and 2.5 × 10 −3 μg/ml for hd and vx, respectively. heleg-shabtai et al. developed the sers-based sensor made by hand-held raman spectrometers for gas-phase nerve (methylphosphonothioic acid: vx) and blister (sulfur mustard: hd) agents monitoring [97] . they comprised the sensor with gold nanoparticles that were modified onto quartz fibers and performed gas-phase experiments utilizing a homemade flow system. they also optimized different sers methods for vx and hd detection in solution and calculated the limit of detection values as 1.8 × 10 −3 µg/ml and 2.5 × 10 −3 µg/ml for hd and vx, respectively. lafuente et al. also prepared the sers-based sensor for gas-phase monitoring of dimethyl methyl phosphonate (dmmp) that is a surrogate molecule of nerve agents [98] . they designed the plasmonic sensor using gold nanoparticles that coated with a citrate layer that reacts as a powerful trap of the dmmp by hydrogen bonding interactions under optimum conditions ( figure 11 ). they reported that this plasmonic sensor can monitor low concentrations of dmmp (130 parts-per-billion) in the gas phase. biosensors 2020, 10, x for peer review 10 of 15 lafuente et al. also prepared the sers-based sensor for gas-phase monitoring of dimethyl methyl phosphonate (dmmp) that is a surrogate molecule of nerve agents [98] . they designed the plasmonic sensor using gold nanoparticles that coated with a citrate layer that reacts as a powerful trap of the dmmp by hydrogen bonding interactions under optimum conditions ( figure 11 ). they reported that this plasmonic sensor can monitor low concentrations of dmmp (130 parts-per-billion) in the gas phase. figure 11 . interaction of dimethyl methyl phosphonate (dmmp) with an sers-based sensor. republished with permission from [98] ; permission conveyed through copyright clearance center, inc. defense against the threats is based on the early monitoring of the biological and chemical agents, the separation of infected individuals, and the assessment of the contaminated area. therefore, fast, sensitive, and portable platforms are required for the real-time detection of these threats. plasmonic sensor-based platforms have become an important part of laboratories. these sensors provide similar sensitivity compared to other conventional platforms. owing to their small size and low cost, plasmonic sensors are suitable not only in the laboratory routine but also for mobile laboratories and field-portable systems. therefore, there is a clear need to develop new technologies to act immediately and detect long-term threat agents in the event of the release of both intentional and unintentional agents. in this review, the recent developments of plasmonic sensors are overviewed for biological and chemical threat agents' detection. compared with conventional platforms that often need labor-intensive sample preparation and advanced instrument, these platforms demonstrate more encouraging applications for human health and life improvement. even though there are many solutions to detecting biological and chemical threat agents, there is still considerable room for improvement. understanding the mechanisms of how the biological and chemical threat agents attack and destroy the living is obligatory to create new strategies for monitoring, discriminating, and destroying these agents. the growth of plasmonic sensors for many of these agents would be a step towards capabilities. defense against the threats is based on the early monitoring of the biological and chemical agents, the separation of infected individuals, and the assessment of the contaminated area. therefore, fast, sensitive, and portable platforms are required for the real-time detection of these threats. plasmonic sensor-based platforms have become an important part of laboratories. these sensors provide similar sensitivity compared to other conventional platforms. owing to their small size and low cost, plasmonic sensors are suitable not only in the laboratory routine but also for mobile laboratories and field-portable systems. therefore, there is a clear need to develop new technologies to act immediately and detect long-term threat agents in the event of the release of both intentional and unintentional agents. in this review, the recent developments of plasmonic sensors are overviewed for biological and chemical threat agents' detection. compared with conventional platforms that often need labor-intensive sample preparation and advanced instrument, these platforms demonstrate more encouraging applications for human health and life improvement. even though there are many solutions to detecting biological and chemical threat agents, there is still considerable room for improvement. understanding the mechanisms of how the biological and chemical threat agents attack and destroy the living is obligatory to create new strategies for monitoring, discriminating, and destroying these agents. the growth of plasmonic sensors for many of these agents would be a step towards capabilities. explosive and chemical threat detection by surface-enhanced raman scattering: a review interaction of dimethyl methyl phosphonate (dmmp) with an sers-based sensor explosive and chemical threat detection by surface-enhanced raman scattering: a review confronting the threat of bioterrorism: realities, challenges, and defensive strategies current status of persistent organic pesticides residues in air, water, and soil, and their possible effect on neighboring countries: a comprehensive review of india biosensor technology: recent advances in threat agent detection and medicine biosensors for the detection of op nerve agents biosensors for biological warfare agent detection detecting biothreat agents: from current diagnostics to developing sensor technologies sensors for detecting biological agents chapter 6: microfluidics application for detection of biological warfare agents regulatory underpinnings of global health security: fda's roles in preventing, detecting, and responding to global health threats pyrene substituted amphiphilic romp polymers as nano-sized fluorescence sensors for detection of tnt in water porphyrin functionalized graphene for sensitive electrochemical detection of ultratrace explosives desorption and transformation of nitroaromatic (tnt) and nitramine (rdx and hmx) explosive residues on detonated pure mineral phases a portable and autonomous multichannel fluorescence detector for on-line and in situ explosive detection in aqueous phase voltammetric platform for detection of 2,4,6-trinitrotoluene based on a molecularly imprinted polymer near real time detection of hazardous airborne substances a carbazole-fluorene molecular hybrid for quantitative detection of tnt using a combined fluorescence and quartz crystal microbalance method novel and selective detection of tabun mimics fluorescent discrimination between traces of chemical warfare agents and their mimics paper-based rna extraction, in situ isothermal amplification, and lateral flow detection for low cost, rapid diagnosis of influenza a (h1n1) from clinical specimens identification of pathogens by mass spectrometry identification of type a, b, e, and f botulinum neurotoxin genes and of botulinum neurotoxigenic clostridia by denaturing high-performance liquid chromatography aptasensors for biosecurity applications chapter 10: surface plasmon resonance sensors for medical diagnosis micro-and nanostructure based oligonucleotide sensors an integrated portable multiplex microchip device for fingerprinting chemical warfare agents a facile and sensitive detection of pathogenic bacteria using magnetic nanoparticles and optical nanocrystal probes a fiber optic biosensor for specific identification of dead escherichia coli o157: h7 rapid detection of mycobacterium tuberculosis biomarkers in a sandwich immunoassay format using a waveguide-based optical biosensor multiplex fiber optic biosensor for detection of listeria monocytogenes, escherichia coli o157: h7 and salmonella enterica from ready-to-eat meat samples molecularly imprinted nanoparticles based plasmonic sensors for real-time enterococcus faecalis detection hydrous ferric oxide-magnetite-reduced graphene oxide nanocomposite for optical detection of arsenic using surface plasmon resonance graphene oxide coupled with gold nanoparticles for localized surface plasmon resonance based gas sensor inkjet-printed paper-based sers dipsticks and swabs for trace chemical detection biomimetic nanoparticles based surface plasmon resonance biosensors for histamine detection in foods plastic antibody based surface plasmon resonance nanosensors for selective atrazine detection fabrication of an "ion-imprinting" dual-emission quantum dot nanohybrid for selective fluorescence turn-on and ratiometric detection of cadmium ions an alternative medical diagnosis method: biosensors for virus detection molecularly imprinted polymer based sensors for medical applications opportunities and challenges for biosensors and nanoscale analytical tools for pandemics: covid-19 enhancing the nanoplasmonic signal by a nanoparticle sandwiching strategy to detect viruses biosensors and sensing systems for rapid analysis of phenolic compounds from plants: a comprehensive review molecularly imprinted polymer integrated plasmonic nanosensor for cocaine detection molecularly imprinted polymers for capturing and sensing proteins: current progress and future implications advances in biomimetic systems for molecular recognition and biosensing chapter 33: virus detection using nanosensors chapter 3: sensor application for environmental pollutants nanoparticle-based immunochemical biosensors and assays: recent advances and challenges high quality factor, high sensitivity metamaterial graphene-perfect absorber based on critical coupling theory and ımpedance matching portable and field-deployed surface plasmon resonance and plasmonic sensor determining the interfacial refractive index via ultrasensitive plasmonic sensors rapid sensing of cu +2 in water and biological samples by sensitive molecularly imprinted based plasmonic biosensor molecular engineering of organic semiconductors enables noble metal-comparable sers enhancement and sensitivity molecular imprinting based hybrid ratiometric fluorescence sensor for the visual determination of bovine hemoglobin surface plasmon resonance instruments. in handbook of surface plasmon resonance surface plasmon resonance sensors for detection of chemical and biological species modern surface plasmon resonance for bioanalytics and biophysics localized surface plasmon resonance sensors a disposable microfluidic-integrated hand-held plasmonic platform for protein detection survey of analytical techniques for noroviruses detecting fingerprints of waterborne bacteria on a sensor visual simultaneous detection of hepatitis a and b viruses based on a multifunctional molecularly imprinted fluorescence sensor overview of recent advances in the design of plasmonic fiber-optic biosensors recent trends in rapid detection of influenza infections by bio and nanobiosensor graphene-based nanocomposites for sensitivity enhancement of surface plasmon resonance sensor for biological and chemical sensing: a review nanomaterial-based optical chemical sensors for the detection of heavy metals in water: recent advances and challenges functionalized gold nanoparticle supported sensory mechanisms applied in detection of chemical and biological threat agents: a review the recent advances of fluorescent sensors based on molecularly imprinted fluorescent nanoparticles for pharmaceutical analysis plasmonic janus microspheres created from pickering emulsion drops raman spectroscopy of optically trapped single biological micro-particles portable biosensors for the rapid detection of biochemical weapons of terrorism epidemiology of biowarfare and bioterrorism. in medical aspects of biological warfare; office of the surgeon general borden institute us army medical department center and school health readiness center of excellence fort sam houston detection of biological agents sensing of biowarfare agents the history of biological weapons use: what we know and what we don't. health secur emerging viral diseases: confronting threats with new technologies centers for diseases control and prevention (cdc) surface plasmon resonance sensing of ebola virus: a biological threat dna-probe-target interaction based detection of brucella melitensis by using surface plasmon resonance a novel surface plasmon resonance biosensor for the rapid detection of botulinum neurotoxins nanosensors based on lspr are able to serologically differentiate dengue from zika infections design of a plasmonic platform to improve the sers sensitivity for molecular detection. photonic sens review of sers substrates for chemical sensing direct detection of bacteria using positively charged ag/au bimetallic nanoparticles: a label-free surface-enhanced raman scattering study coupled with multivariate analysis highly sensitive detection of high-risk bacterial pathogens using sers-based lateral flow assay strips gold nanoparticles as sensors in the colorimetric and fluorescence detection of chemical warfare agents a review of biosensors and biologically inspired systems for explosives detection nanomaterial based biosensors for detection of biomarkers of exposure to op pesticides and nerve agents: a review recent advances in the development of chromophore-based chemosensors for nerve agents and phosgene development of surface plasmon resonance sensors based on molecularly imprinted nanofilms for sensitive and selective detection of pesticides spr nanosensor based on molecularly imprinted polymer film with gold nanoparticles for sensitive detection of aflatoxin b1 lspr-enhanced photonic crystal allows ultrasensitive and label-free detection of hazardous chemicals nonlinear plasmonic sensing for label-free and selective detection of mercury at picomolar level ag nanostructures as the surface plasmon resonance (spr)-based sensors: a mechanistic study with an emphasis on heavy metallic ions detection d-shaped pcf sensor based on spr for the detection of carcinogenic agents in food and cosmetics surface-enhanced raman spectroscopy (sers) for detection of vx and hd in the gas phase using a hand-held raman spectrometer highly sensitive sers quantification of organophosphorous chemical warfare agents: a major step towards the real time sensing in the gas phase key: cord-020766-0gacqii4 authors: murthy, sreekant; papazoglou, elisabeth; kanagarajan, nandhakumar; murthy, narasim s. title: nanotechnology: towards the detection and treatment of inflammatory diseases date: 2006 journal: in vivo models of inflammation doi: 10.1007/978-3-7643-7520-1_8 sha: doc_id: 20766 cord_uid: 0gacqii4 biological systems operate at the nanoscale. nanomedicine is the application of nanotechnology to monitor and treat biological systems in health and disease. this is accomplished by real time monitoring of molecular signaling at the cellular and tissue level. during the past decade, there has been an explosion in this field, resulting in revolutionary advances in determining the microstructure and function of living systems. these discoveries have led to the development of powerful tools for fundamental biological and medical research. nanotechnology has been applied to targeted drug delivery to minimize side effects, creating implantable materials as scaffolds for tissue engineering, creating implantable devices, surgical aids and nanorobotics, as well as throughput drug screening and medical diagnostic imaging. the nanoinitiatives are funded by governments and private sources throughout the world to develop or further refine the technology to provide the beyond-imaginable, most sophisticated tools to a physician and scientists to inflammatory diseases. no doubt, there will be many technical, regulatory and legal challenges in the deployment of these technologies. unquestionably, there is enough desire and commitment to meet these challenges for the good of society and betterment of the quality of life. accomplishments, combining them with available new technologies to identify the earliest signatures of inflammation and cancer. such developments will allow us to provide immediate and specific intervention and monitor the progress before it cascades into chronic inflammation and malignancy. to fulfill this objective, it requires the development of technologies of 1-100 nm size, which display unique mechanical, electrical, chemical, and optical properties and assist in visualizing or interacting with receptors, cytoskeleton, specific organelles and nuclear components within the cells. it will be very rewarding when many of these technologies can migrate into monitoring the disease condition through non-invasive methods in vivo in a physically undisturbed state, thus minimizing the influence of artifacts induced by physical methods while securing biological samples. the integration of nanotechnology with biology and medicine has given birth to a new field of science called "nanomedicine". the ultimate goal of nanomedicine is to develop well-engineered nanotools for the prevention, diagnosis and treatment of many diseases. in the past decade, extraordinary growth in nanotechnology has brought us closer to being able to vividly visualize molecular and cellular structures. these technologies are beginning to assist us in our ability to differentiate between normal and abnormal cells and to detect and quantify minute amounts of signature molecules produced by these cells. most of these represent real time measurements, relating to the dynamic relationship among structures in the damaged area and also to repair of damaged tissues. novel pharmaceutical preparations have been developed to fabricate nanovehicles to deliver drugs, proteins and genes, contrast enhancement agents for imaging, and hyperthermia agents to kill cancer cells. several of these inventions have already transitioned into basic medical research and clinical applications. because of this, some social, ethical, legal and environmental issues have emerged. thus, regulatory and educational strategy needs to be developed for the society to gain benefit from these discoveries. the focus of this chapter is to provide an overview of the state-of-the-art in nanotechnology with particular reference to detecting and treating inflammation and cancer at the earliest settings. nanotechnology encompasses multiple scientific disciplines, which exploit materials and devices with functional co-assembled molecules or sensors that has been engineered at the nanometer scale typically ranging from 0.1 to 100 nm [2] . in medical fields, it offers a wide range of tools that can be used as drug delivery platforms [3] , better contrast agents in imaging [4] , chip-based bio-laboratories [5] and nanoscale probes [6] that are able to track cell movements and manipulate molecules. multifunctional nanostructures can combine diagnostic and therapeutic modalities and target cellular events. the concept of molecular medicine to develop personalized treatments can be made possible with the information available with the developed nanotools. these devices, systems and functionalized structures contain unique properties as a result of their nanosize. for example, gold nanoparticles and carbon nanotubes possess different properties [7] at the micron scale. semiconductor particles exemplified by quantum dots exhibit quantum confinement effects, hence they fluoresce at various wavelengths compared to the semiconductor particles at micron size which do not exhibit the same optical or magnetic properties [8] . macromolecular structures such as dendrimers and liposomes at the nanoscale are also considered valid nanotools [9] , while biological molecules of nanometer size in their native state such as dna and monoclonal antibodies are not examples of "nanotools". thus, nanometer size is of critical importance to the cell and the living organisms. interestingly, fabricated nanoscale devices are of the same size as subcellular organelles. some nanoscale structures are of the order of enzymes, receptors and key molecules within the cell membrane or cytoplasm. for example, the lipid bilayer surrounding cells is 6 nm and hemoglobin is 5 nm in size. by modifying the surface chemistry of these nanostructures, which permits covalent or ligand-receptor (lock-key-type) or electrostatic interaction with key molecules, we can identify biomolecules on the cell surface and in the cytoplasm. an advantage of this will be our ability to map the transport of those molecules from the cytoplasm across the cell membrane so we can understand the cellular behavior in health and disease. nanoparticles smaller than 20 nm can pass through blood vessel walls [10] , which opens opportunities of diagnostic imaging and targeted delivery of drugs when non-toxic nanoparticles are used. what is critical for scientists engaged in inflammation science and engineering is that these technologies should be applicable to detect or monitor: -host biochemical and immune responses -bacterial and viral pathogens interacting within the local immune system -effect of noxious chemicals and pharmaceutical agents -thrombosis -neurogenic inflammation -wound healing and remodeling -imaging -diagnosing and treating vulnerable plaques -drug delivery and therapeutics for local delivery and retention nanotechnology in the medical field offers a wide range of tools that can be used as drug delivery platforms, better contrast agents in imaging, chip-based biolabs and nanoscale probes able to track cell movements and manipulate molecules [10] . combination of these multifunctional nanostructures through cross-disciplinary interactions may further enhance our diagnostic and therapeutic capabilities and to monitor intra-and extracellular cellular events in inflammation and cancer. existing and emerging technologies, which may impact on early detection of inflammation, prevention and early detection of cancer, include several diverse technological innovations. they are bio-mimicry self-assembling peptide systems, which serve as building blocks to produce nucleotides, peptides and phospholipids, which support cell proliferation and differentiation and give insights into protein-protein interactions [11] . microchip drug release systems, micromachining hollow needles and two-dimensional needle arrays from single crystal silicon for painless drug infusion, intracellular injections, microsurgeries and needle-stick blood diagnosis form another group of tools [12, 13] . all of these inventions could one day lead to develop personalized treatments [14] . the creation, control and use of structures, devices and systems with a length scale of 1-100 nm is the domain of nanotechnology. macromolecular structures such as dendrimers and liposomes at the nanoscale are also considered valid nanotools [9, 15] . the application of various nanotools in various areas of medicine is depicted in figure 1 . this list in the figure is by no means exhaustive as nanotechnology is continuing to grow with new technologies emerging each day. biomolecular nanopore detector technology was first developed to rapidly discriminate between nearly identical strands of dna thereby replacing the tedious process of running billions of copies of dna through sequencing machines and minimizing errors and saving time [16] . in this technology single molecules of dna are drawn through pores that are 1-2 nm in size and serve as a sensitive detector. the detection system through its electronic signature process can sequence more than one base pair per millisecond. this technology has the potential to detect dna polyploidy, and dna mutations. this field includes biomimetics encouraged to mimic nature and create biomolecular nanomachines to handle various biological problems. many biological systems use self assembly to assemble various complex molecules and structures [17] . numerous man-made self-assembling systems that mimic natural self assembly of molecules are created to snap together fundamental building blocks of complex polymer molecules structured easily, and inexpensively, on beads, tubes, wires, and flat supports, and in suspensions and liposomes. these assemblies can have geneti-cally introduced bifunctionality so that nonspecific molecules are repelled from fusing with the cell membrane fusion layers. dna, lipid bilayers, atp synthase, peptides and protein foldings are target candidates for self assembly. liposomes are an example of a human-made supramolecular structure. nanoscale cantilevers are about 50-mm-wide flexible diving board-like beams that can be coated with antibodies and dna complementary to a specific protein or a gene. when molecules come in contact with these substrates coated on the surface of cantilevers, they bind to the substrate and make the cantilevers resonate or bend as a result of this binding event [18] . this bending deflection is proportional to the quantity of binding, thus making it a quantitative technique. multiple cantilevers can be used simultaneously to differentiate between bound and unbound molecules. likewise, multiple antibodies can be used in the same reaction set up to quantify several markers at a time. an important advantage of this technique is that there is no need to add fluorescent tags to detect and quantify the molecule. any biological sample containing biomolecules of interest can be tested. nanoscale cantilevers, constructed as part of a larger diagnostic device, can provide rapid and sensitive detection of inflammation and cancer-related molecules and to evaluate how various drugs bind to their targets at a concentration 20 times lower than clinical threshold. carbon nanotubes, also known as "bucky tube" and "buckyballs" are a member of fullerene structural family potentially useful in a number of biological applications. they could be cylindrical (nanotubes), spherical (buckyballs) or branched (fullerenes). nanotubes could be single-walled nanotubes (swnt) or multi-walled nanotubes (mwnt). the usefulness of nanotubes in drug delivery and cancer therapy is accomplished through the transporting capabilities of carbon nanotubes via suitable functionalization chemistry and their intrinsic optical properties. proper surface functionalization is necessary to make carbon nanotubes biocompatible. in their most recent applications, swnts have been used to transport dna inside living cells [19] . intracellular protein transport has also been accomplished [20] , although they are suspected to cause severe immune responses. most swnts have diameters close to 1 nm, with a tube length that can be many thousands of times larger. swnts with lengths up to the order of centimeters have been produced [21] . swnts are a very important class of carbon nanotubes because they exhibit important electrical properties not shared by the mwnt variants. on the other hand, mwnts are fabricated as multiple concentric nanotubes precisely nested within one another for perfect linear or rotational bearing. the technology has now advanced into merging these mwnts with magnetic nanomaterials like magnetite, which can be functionalized. gadofullerenes offer the ability to concentrate more gadolinium at the site of disease, than traditional gd-dtpa. this is the result of the shielding that the carbon structure provides and its ability to link more gadolinium per conjugate. gadofullerenes also take advantage of the gadolinium-water interactions as the gadolinium is brought along the periphery of the structure and can maintain its interaction with water, which is the basis of traditional proton density magnetic resonance imagine (mri). these properties lead to a greater signal, which can increase sensitivity to small lesions. recently, the technology has been further improved by developing smart bionanotubes that could be manipulated to produce open or closed end nanotubes to encapsulate drugs or genes to deliver them in a particular location [22] . thus, possibilities exist for using nanotubes to improve gene sensing, gene separation, drug delivery and detection of biomarkers to improve health care, protection against bioterrorism and other areas of molecular sensing. they are mostly spherical particles with specific properties that allow their detection, analysis and quantification. fundamentally, they are nanoprobes where these particles are complexed with biomolecules, drugs and other reagents. they exhibit various physical and optical properties. for example, iron oxide nanoparticles exhibit super paramagnetic properties [23] , and gold nanoparticles specific optical absorption properties depending on their size [24] . it is important to note that particles made from the same materials but of micron dimensions do not exhibit such unique optical or magnetic properties [8] . one can thus combine the immense surface to volume ratio of these nanostructures to deliver higher loads of compounds encapsulated or linked to their surface, while their presence can be measured due to their characteristic magnetic or optical properties. quantum dots (qds) are tiny light-emitting particles on the nanometer scale. they are emerging as a new class of biological probes that could replace traditional organic dyes and fluorescent proteins. the fundamental benefit of using qds is the high quantum yield and strong emission intensity. the emission spectrum of qds is a function of the particle size, and hence by varying particle size qds can emit from visible to infrared wavelengths [25] . they can be excited by uv light. their broad excitation spectrum and narrow emission spectrum with little or no spectral overlap makes them attractive for imaging and resolving multiple species at the same time without complex optics and data acquisition systems. qds offer higher signal to noise ratios compared to traditional fluorochromes. their high sensitivity allows accurate detection even in the presence of strong autofluorescence signals encountered during in vivo imaging [26] . their excellent resistance to photobleaching is particularly useful for long-term monitoring of biological phenomena, critical in live cell imaging and thick tissue specimens. already qds are finding increasing use in live cell imaging, by themselves or as fluorescence resonance energy transfer (fret) donors combined with traditional fluorchromes, and in in vitro assays and live animal imaging for cancer and tumor diagnostics. limitations of qds can arise from the stability of the core shell structure. most commercially available materials comprise a core of cdse and a shell of zns. to render this inorganic structure hydrophilic, amphiphilic polymers are used to cap the shell layer and provide reactive sites for further linking to proteins. it is, therefore, the stability of this layer that controls the aggregation of qds, as well as possible release of core materials (cd ions) to their surroundings, which may result in toxicity [27] . this has limited the immediate clinical use of qds, but has focused applications to animal testing and in vitro assay developments. another reported problem is the blinking property of the qds. qds tends to blink at the single dot level and hence present some limitation in absolute fluorescence quantification. however, when used for imaging of biomarkers, this property does not have much of an effect as there are hundreds or thousands of qds in a sample to allow proper averaging. paramagnetic iron oxide nanoparticles are a new class of contrast agents that are finding increasing applications in the field of diagnostics and molecular imaging based on magnetic resonance (mr) [23] . traditional mri agents rely on the interaction of the proton density, i.e., water molecules and the magnetic properties of the tissue. these paramagnetic agents accelerate the rate of relaxation of protons in the longitudinal direction, resulting in bright images, and hence are highly dependent on water molecules. however, the super paramagnetic iron oxide nanoparticles, by the virtue of their nanoscale properties, disturb the magnetic field independently of their environment, and hence are not dependent on presence of water molecules. they are also called negative enhancers as they act as negative contrast agents and appear dark where they are sequestered. the traditional mr agents such as gadolinium-diethylenetriamine penta-acetic acid (dtpa) enhance the signal from the vascular compartments and are nonspecific, whereas the nanoparticle-based contrast agents impact the mr signal from tissues and cells. iron oxide nanoparticles are classified into two types depending on their size: (1) superparamagnetic iron oxides (spios) (50-500 nm), and (2) ultra-small super paramagnetic iron oxide (< 50 nm). the advantage of these contrast agents lies in their ability to get sequestered anywhere within a support matrix and still generate a contrast, whereas the traditional mr agents need water in their vicinity of generate contrast. these nanoparticles can be used for both passive and active targeting. because of the small size of these particles, tissue macrophages readily take up these agents, and hence it is possible to image liver, spleen, lymph nodes, and lungs. in addition, it is also possible to functionalize these nanoparticles using a wide variety of ligands, antibodies, peptides, aptamers, drugs, etc., to achieve site-specific or biomarker-specific targeting. this is an added advantage since traditional paramagnetic formulations are difficult to conjugate to antibodies, and even when conjugated, owing to the small number of cellular receptors, the signal intensity is not sufficient for accu-rate imaging. thus, as a result of their superparamagnetic properties, iron oxide nanoparticles have been used as contrast agent for imaging of cancer, brain inflammation, arthritis, and atherosclerotic plaques. because of the small size, these iron oxide particles have been able to distinguish between the normal and tumor-bearing lymphatic nodes [28] . these nanoparticles may also distinguish very small metastases (less than 2 mm in diameter) within normal lymph nodes, a size well below the detection limit of the most sensitive imaging techniques such as positron-emission tomography (pet) available today. using cells loaded with iron oxide nanoparticles, it has been shown that these particles are non toxic and are cleared from the cell after five to eight divisions. lewin et al. [29] labeled stem cells with iron oxide particles using hiv tat peptide and injected them systemically. the labeled stem cells homed on to the bone marrow, and the labeled stem cells did not cause any impairment. however, due to the small size of these particles, a long time is required (up to 24 h) to clear them from the organs and blood to reduce background signals. thus, mri using spios may result in improved sensitivity and selectivity, and may assist diagnosis of tumors at the earliest stages of malignancy or metastasis. dendrimers are a new class of hyper-branched polymer macromolecules that radiate from a central core with structural symmetry. they can vary in shape, size, surface, flexibility and topography, enabling fabrication of functional nanoscale materials that would have unique properties [30, 31] . they may be useful for developing antiviral drugs, tissue repair scaffolds, and targeted carriers of chemotherapeutics. certain dendrimers are now being used commercially as immuno-diagnostic agents and gene transfection vectors. dendrimers complexed with gadolinium (iii) ions (gadomer-17) are being tested (phase i clinical trial) for mri angiography [32] . it is anticipated that many exciting developments will emerge from the use of dendrimers in the near future. these are made up of phospholipid bilayers exhibiting multifunctioning characteristics. they facilitate encapsulating of various classes of drugs and diagnostic agents for their controlled delivery into cellular and therapeutic targets. important drug delivery strategies utilizing these agents include polymersomes, hydrogel matrices, nanovesicles/nanofiber mats and biodegradables [33] . both small and large molecules can be used. biodegradeable polymersomes based on polyethylene oxide have been synthesized, and they may be used as a surface to anchor antibodies or other targeting molecules. quite recently fluorescent materials have been embedded into 163 these cell-like vesicles [34] to produce near-infrared emissive polymersomes that could be used to locate areas of inflammation and deliver a load of drug to inflammation sites. interestingly, inflammation sites deeper than 1 cm could be imaged with this technique. efforts are being made to target dna complexes into hepatocytes and macrophages with the idea of enabling gene therapy and delivering genetically derived vaccines in a safe and efficacious manner. polymeric micelles are useful as developing agents for -scintigraphy, mri, and computed tomography (ct) [35] . liposomes can be injected intravenously and when they are modified with lipids that render their surface more hydrophilic, their circulation time in the bloodstream can be increased significantly. another class of polymersomes, called polymer nanotubes, has been synthesized by directly pulling on the membrane of polymersomes using either optical tweezers or a micropipette [36] . these polymersomes are composed of amphiphilic diblock copolymers consisting of an aqueous core connected to the aqueous interior of the polymersome, which are less than 100 nm in diameter. they are unusually long (about 1 cm) and are stable enough to maintain their shape indefinitely. the pulled nanotubes are stabilized by subsequent chemical cross-linking. the aqueous core of the polymer nanotubes together with their robust character offer opportunities for nanofluidics and other applications in biotechnology, especially in the development of nanohyperdermic syringes [36] . development and use of analytical tools in diagnostic area possibly presents immediate benefits to the user. many diagnostic tools have been developed to improve human health. the diagnostic detection methods involve measuring antibody or antigen-based complexes, enzyme-based reaction rates, and polymerase chain reactions using micro-electro-mechanical systems (mems) [37] . other methods include whole-cell bacterial sensors and biosensors which utilize aptamers, which are biomimetic synthetic bioreceptors that can complex with proteins, nucleic acids and drugs. the signal processing in these systems may be optical, electrochemical, or mass-related, creating resonance and thermal detection. some diagnostic methods utilize nanoparticles as nanoprobes where nanoparticles are interfaced with biological molecules such as antigens, antibodies or chemicals. the nanoparticles used in diagnostics include qds, nanobarcodes, metallic nanobeads, silica, magnetic beads, carbon nanotubes, optical fibers and nanopores [37] . in antibody/antigen-based detection methods, for example, 1-2-nm-wide, boron-doped silicon wires laid down on a silicon grid can be coated with antigens to provide real time detection of antibodies. antibody binding to immobilized antigen gives a measurable conductance change at antibody concentrations less than 10 nm. detection of single copies of multiple viruses has been accomplished via antibody-conjugated nanowire field effect transistors [38] . the dream of optical biopsy is closer to reality with antibody-functionalized semiconductor nanoparticles (qds) detected by fluorescence microscopy. multiplexed assays can be developed since the fluorescence emission of qds is tunable by changing their size. outstanding detection sensitivity of antibodies in whole blood (picogram per ml) has been obtained using gold nanoparticle conjugates [39] . for detection purposes classical tools can be used as well as nanobased methods, such as atomic force microscopy (afm) and near-field scanning optical microscopy (nsom), methods where quantum tunneling plays a key role in amplifying detection capability. again, such phenomena are related to the nanometer distance between the instrument probe and the surface/specimen under examination. afm is used to elucidate structures of biomolecules under physiological conditions [40] , to determine antibody/antigen binding properties [41] , to image the topology of viruses [42] , and to image pathologies at the molecular scale [43] . researchers in academia and the pharmaceutical industries traditionally use bioassays, which are often cumbersome and riddled with errors. the recent explosive development in the field of microfluidics, biotechnology and functional genomics has resulted in the miniaturization of these bioanalytical assays to micron scales for routine and throughput screening [44] . these assays have been used for genomic and proteomics analysis, though their application to proteomics still requires refinement since replication of proteins as opposed to dna is yet to be fully realized. efforts are being made to improve miniature microarrays, which are still used for analyzing proteins. these include fabrication of afm-based dip-pen nanolithography (dpn), which can probe complex mixtures of proteins, reactions involving the protein features and antigens in complex solutions, and can aid the study of cellular adhesion at the submicrometer scale. protein nanoarrays generated by dip-pen nanolithography are emerging [45] . with further advances in miniaturization techniques like dpn, it will be possible to design nanoarrays that can detect biological entities on a single particle level in a time-and cost-efficient manner and also profiling of new diagnostic biomarkers at a detection level beyond our imagination. the pharmaceutical industry, physicians and patients have long desired better pharmaceutical formulations to improve and extend the economic life of proprietary 165 drugs, to reduce the costs of preparation and treatment, and reduce toxicity or even death. nanotechnology has already made significant inroads into the problems of improving delivery of injectibles, oral formulations, drug device implants, and topical and transdermal delivery of drugs. more is expected from nanotechnology in improving the delivery of drugs to the brain, as many of the formulations aimed at treating diseases of the brain fail to cross the blood-brain barrier. various methods have been tested for drug delivery. for example, carbon-based materials, nanostructures, silicone-based materials, polymers and liposomes, which are capable of delivering drug molecules directly into cells, tumors and sites of inflammation either actively or passively. there is no question that there are many limitations such as opsonization, problems with encapsulation and leakiness of drug that needs to be tackled. some of these obstacles have been overcome by the development of agents like "stealth liposomes", which escape attack by the immune system. thus, nanotechnology is expanding our capabilities through promising approaches for delivery of therapeutic agents. nanosystems and nanoparticles have opened up hitherto unforeseen avenues in diagnostics and therapeutics in medicine, especially in the fields of inflammation and cancer. the previous treatment strategies in the fields of autoimmune diseases and cancer involved non-targeted treatment options with extensive "collateral damage". nanodelivery of drugs is envisioned to reduce this collateral damage, extend a drug's availability and effectiveness at the site, and reduce toxicity, cost and storage. the focus of this section is to highlight several nanomedicine applications that have made an immediate major impact in these fields. biological nanostructures used in drug delivery systems include lipid-, silica-, polymer-, fullerene (carbonbased buckyballs, bucky tubes)-based nanostructures such as liposomes, micelles and nanoparticle systems. liposomes have been widely used as drug delivery systems, but current knowledge extends the use of any nanoparticle as an efficient carrier with necessary modifications. liposomes are vesicles with phospholipid membranes that contain hydrophilic substances in their core. the properties vary widely based on the size, lipid composition, surface charge and method of preparation [46] . liposomal formulations have been used as anti-cancer and anti-fungal drugs, and have helped reduce the adverse effects of these drugs, while improving the efficacy and pharmacokinetics. conventional liposomes are short-lived in vivo, and are rapidly cleared by the reticuloendothelial system (res). a novel liposomal formulation with a polyethylene glycol (peg) coating avoids res-mediated clearing, and is called a stealth liposome. these stealth liposomes have favorable properties like long circulation half-life and targeted accumulation in tumor tissues [47] . liposomes have been extensively used in cancer therapy. some of the major classes of anti-cancer drugs in liposomal formulations that are currently available or in late stages of development include anthracyclines, camptothecins, platinum derivatives, anti-metabolites and cell-cycle-specific drugs like vincristine and doxorubicin. liposomal formulations have been shown by clinical trials to decrease cardiotoxicity as compared to conventional formulations [48] . current liposomal formulations include pegylated liposomal doxorubicin (doxil â® orthobiotech, caelyx â® schering-plough) non-pegylated doxorubicin (myocet â® elan pharma) and liposomal daunorubicin (daunoxome â® , gilead sciences). this protective strategy to limit toxicity has aided in limiting the cumulative dose of anthracyclines and administration of dexrazoxane, a highly effective cardioprotective agent, prior to anthracycline administration [49] . liposomal platinum derivatives like cisplatin and carboplatin are used in the treatment of head and neck cancers, testicular cancer, lung cancer and many other malignancies. they have shown significantly reduced toxicity and better pharmacokinetic profiles compared to conventional formulations [50] . some formulations have not yielded the best results. for example, spi-077, a stealth liposomal cisplatin showed low clinical efficacy in phase i/ii clinical trials, possibly secondary to inadequate release of drug from liposomes [51, 52] . lipoplatin, which has shown lipid bilayer fusing properties [53] , has shown significant nephrotoxicity, but further clinical research is awaited to see if this translates into improved clinical efficacy. the use of liposomes has also been extended for enhancing immunotherapeutic effects. it is now known that liposomal targeting can be achieved by passive targeting or active targeting. passive targeting is achieved in both inflammatory and cancerous conditions taking advantage of the leakiness caused by many vascular factors that enhance permeability. this opens up a window of opportunity to increase the drug delivery, with accumulation of drug at higher concentration at the targeted site by extravasation, thereby reducing toxicity and collateral damage. on the other hand, active targeting depends on certain unique properties and molecular strategies involving monoclonal antibody-liposomal conjugates (immunoliposomes), which enables specific tumor cell targeting by antigen identification and drug delivery by internalization of the liposome by tumor cells [54] . promising examples are the enhanced anti-tumor activity of anti-her2 immunoliposomes containing doxorubicin [55] , and the use of anti-epidermal growth factor receptor (anti-egfr) immunoliposome, which showed increased cytotoxic effect in vitro against tumor cells overexpressing egfr and enhanced efficacy in vivo in xenograft models [56] . the recent advent of the widespread use of monoclonal antibodies in cancer therapy promises more such targeted therapeutic agents. liposomal preparations have also been studied in various autoimmune and chronic inflammatory diseases. targeted delivery of anti-inflammatory agents to 167 inflamed tissue is a promising approach limiting the adverse impact of these agents on healthy tissues. in animal colitis models, liposomal formulations of 5-asa achieved significantly higher local concentrations of 5-asa in inflamed colonic tissues compared to current treatment methods. on the other hand, liposomal preparation of mercaptopurine (6-mp) failed to improve local delivery [57] because the drug is metabolized before it reaches the inflammation site. when peg-liposomes containing glucocorticoids were injected in mouse collagen arthritis models, longlasting reduction in joint inflammation was achieved with a single dose, while regular steroids needed multiple injections [58] . weekly inhaled liposomal budesonide was as effective as daily inhaled budesonide in a mouse model of asthma [59] . if this result can be replicated in clinical trials, it can greatly enhance patient compliance. this is particularly important since treatment of chronic inflammatory diseases is hampered by patient non-compliance. similarly, it is reported that liposomal preparations of anti-oxidants can also be used in diseases like adult respiratory distress syndrome (ards), sepsis, radiation lung injury and emphysema [60] . liposomes have also been shown to be effective in diverse clinical applications such as enhanced drug delivery systems for analgesics [61] [62] [63] . the carbon nanostructures that have gained most attention have been fullerene nanotubes and the geodesic dome-shaped c60 fullerenes. they have wide ranging applications as drug carriers and can also be used as vaccine delivery tools enhancing the immune response [64] . they have demonstrated neuroprotective properties in cortical cell cultures and have potential therapeutic applications in neuronal-inflammation and neurodegenerative disorders like parkinson's disease, amyotropic lateral sclerosis (als) and cerebral ischemia [65] . carbon nanotubes can cross the cell membrane without causing damage and they can act as "nanoneedles" [66] . tectodendrimers are multicomponent dendrimers capable of multiple functions like identifying defective cells, delivering imaging and therapeutic agents to the cell and reporting the response to therapy. they can be individualized for each specific disease state and can be mass produced. baker and colleagues [67] designed dendrimers with folic acid, fluorescein and methotrexate, and showed a 100-fold increase in the cytotoxic response of cells to methotrexate. in some cases, nanoparticles also aided in avoiding harmful adverse effects of drug vehicles, as in the case of abraxane â® (american bioscience), nanospheres of albumin-bound paclitaxel thus avoiding the need for toxic solvents like cremophor [68] . recently, kriz et al. [69] described a new sensing technology platform integrating a magnetic permeability detection and a two-site heterogeneous immunoassay using monoclonal anti-crp-conjugated superparamagnetic nanoparticles and solid-phase polyclonal anti-crp-conjugated silica microparticles to assay crp in blood samples. the results were comparable to assays performed in a clinical laboratory. the methodology seems applicable for rapid screening of biomarkers and drugs in a rapid and cost-effective manner using whole blood samples. the field of molecular imaging has exploded in recent times. significant advances have been made in real-time cellular imaging and for detecting cellular pathophysiology. over the last few years, varieties of nanostructures containing novel contrast agents and nanomaterials such as qds, gold nanoparticles or nanoshells, supramagnetic nanoparticles complexed with biological agents that can specifically bind molecular signatures of inflammation and cancer have been described [70, 71] . qds have enabled in vivo live imaging, down to the level of a single qd inside a cell. qds provide several advantages over organic fluorochromes since they are photostable permitting imaging over extended periods of time, avoid interference with cellular autofluorescence, permit tracking of multiple processes simultaneously in the cells and are less toxic than organic dyes [72] . although there is a possibility of cellular toxicity from the metallic components of qds, no cellular toxicity was seen, even under selection pressure, when qds were used to track metastatic tumor cell extravasations in an animal model [73] . however, toxicity to humans is still being debated. qds have multimodal applications as contrast agents in bioimaging, microarrays, and facs analysis, in monitoring pharmacokinetics of therapeutic agents, and in multicolor optical coding for high throughput screening [74] . qds have been successfully been used for sentinel lymph node sampling in gastrointestinal tract in pig models. qds can have immediate applications in oncological surgery if the safety profile can be established for humans [75] . there are several potential pitfalls, including lack of convincing evidence for absence of cytotoxicity. further research is needed before we move forward towards widespread use of qds in biological systems [76] . qds can also potentially replace conventional fluorochromes in complex fluoroimaging techniques like fret and fluorescence lifetime imaging microscopy, but intensive research is needed before that can happen. spio crystal core nanoparticles have magnetic properties that can be used to enhance current mri techniques due to their selective activity during t2 relaxation times. they can also act as 'negative enhancers' [77] . utilizing the lymphotropic properties of these nanoparticles, weissleder and colleagues [78] showed that superparamagnetic nanoparticle enhanced high-resolution mri. it was far superior to conventional high-resolution mri in detecting clinically occult prostatic cancer metastasis to lymph nodes. the combined use of qds with superoxide paramag-netic crystals may provide additional information by targeting specific molecular targets for imaging [79] . a particularly interesting application is the use of spio particles for the study of nucleic acid sequences and surface topography of subcellular organelles. this is achieved by a modified afm with a nanoneedle mounted on a cantilever beam that deflects when it comes in contact with a paramagnetic nanoparticle. this response can be quantified and mapped [80] . metal nanoshells are nanoparticles that can serve as strong near infrared absorbers. this property has been exploited to provide targeted thermal therapy selective to tumor cells without damaging normal tissue using gold nanoshells [81] . gadolinium neutron capture therapy has several advantages, including more efficient tumor killing effects and the potential for simultaneous mri to assess response. fukumori and colleagues [82] utilized cationic polymer chitosan nanoparticles that incorporated gadolinium for efficient cellular uptake, and demonstrated significant in vitro tumoricidal effect at relatively low concentrations. recently, bankiewicz and colleagues [83] described an integrated strategy to delivery drugs to the brain. the combined technology involved conventionenhanced delivery (ced) to deliver liposomes containing gadoteridol, with dil-ds and mri to obtain detailed images of drugs moving through a living primate brain following ced for imaging, and to induce better clinical efficacy. molecular imaging has now crossed-over into medical imaging through the use of smart imaging agents for in vivo molecular imaging and imaging of animal models [84] [85] [86] . a recent study showed that magnetic nanoparticle conjugated with anti-vcam-1 antibodies can detect vcam-1 expression through fluorescence and magnetic resonance on endothelial cells in vivo and in vitro [87] . this is an important step, opening up opportunities to use many specific markers specific for inflammation and cancer to diagnose and monitor many inflammatory diseases and cancers. biological systems operate at the nanoscale. nanomedicine is the application of nanotechnology to monitor and treat biological systems in health and disease. this is accomplished by real time monitoring of molecular signaling at the cellular and tissue level. during the past decade, there has been an explosion in this field, resulting in revolutionary advances in determining the microstructure and function of living systems. these discoveries have led to the development of powerful tools for fundamental biological and medical research. nanotechnology has been applied to targeted drug delivery to minimize side effects, creating implantable materials as scaffolds for tissue engineering, creating implantable devices, surgical aids and nanorobotics, as well as throughput drug screening and medical diagnostic imaging. the nanoinitiatives are funded by governments and private sources throughout the world to develop or further refine the technology to provide the beyond-imag-170 inable, most sophisticated tools to a physician and scientists to inflammatory diseases. no doubt, there will be many technical, regulatory and legal challenges in the deployment of these technologies. unquestionably, there is enough desire and commitment to meet these challenges for the good of society and betterment of the quality of life. intestinal inflammation: a complex interplay of immune and nonimmune cell interactions nanoparticle and targeted systems for cancer therapy surfactant stabilized contrast agent on the nanoscale for diagnostic ultrasound imaging nano-scale proteomics approach using two-dimensional fibrin zymography combined with fluorescent sypro ruby dye plasmonics-based nanostructures for surfaceenhanced raman scattering bioanalysis engineered materials for biophotonics applications: improving sensing, and therapeutics biological applications of dendrimers nanomedicine: current status and future prospects emerging biological materials through molecular self-assembly micromachined needles and lancets with design adjustable bevel angles microfabricated microneedles for gene and drug delivery role of nanobiotechnology in developing personalized medicine for cancer genospheres: self-assembling nucleic acid-lipid nanoparticles suitable for targeted gene delivery rapid discrimination among individual dna hairpin molecules at singlenucleotide resolution using an ion channel molecular self-assembly bioassays based on molecular nanomechanics carbon nanotubes as intracellular transporters for proteins and dna: an investigation of the uptake mechanism and pathway carbon nanotubes as intracellular protein transporters: generality and biological functionality direct synthesis of long single-walled carbon nanotubes strands cationic liposome-microtubule complexes: pathways to the formation of two-state lipid-protein nanotubes with open or closed ends accumulation of ultra-small super paramagnetic particles of iron oxide in human atherosclerotic plaques can be detected by in vivo magnetic resonance imaging applications of nanotechnology to biotechnology quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules quantum dot bioconjugates for imaging, labeling and sensing differences in subcellular distribution and toxicity of green and red emitting cdte quantum dots superparamagnetic iron oxide nanoparticles: nodal metastases and beyond tat peptidederivatized magnetic nanoparticles allow in vivo tracking and recovery of progenitor cells dendrimers for pharmaceutical and biomedical applications designing dendrimers for biological applications synthesis and evaluation of gadolinium complexes based on pamam as mri contrast agents biodegradable polymersomes as a basis for artificial cells: encapsulation, release and targeting near-infrared-emissive polymersomes: self-assembled soft matter for in vivo optical imaging recent advances with liposomes as pharmaceutical carriers stable and robust polymer nanotubes stretched from polymersomes nanotechnology on duty in medical applications electrical detection of single viruses whole-blood immunoassay facilitated by gold nanoshell-conjugate antibodies biological cryo atomic force microscopy: a brief review adhesion forces between individual ligand-receptor pairs atomic force microscopy imaging of retroviruses: human immunodeficiency virus and murine leukemia virus adhesion at calcium oxalate crystal surfaces and the effect of urinary constituents functional protein nanoarrays for biomarker profiling protein nanoarrays generated by dip-pen nanolithography self-assembled lipid superstructures: beyond vesicles and liposomes from conventional to stealth liposomes: a new frontier in cancer chemotherapy reduced cardiotoxicity and comparable efficacy in a phase iii trial of pegylated liposomal doxorubicin hcl (caelyx/doxil) versus conventional doxorubicin for first-line treatment of metastatic breast cancer the effect of dexrazoxane on myocardial injury in doxorubicin-treated children with acute lymphoblastic leukemia comparative intravenous toxicity of cisplatin solution and cisplatin encapsulated in long-circulating, pegylated liposomes in cynomolgus monkeys phase i-ii study of pegylated liposomal cisplatin (spi-077) in patients with inoperable head and neck cancer phase i and pharmacokinetic study of spi-77, a liposomal encapsulated dosage form of cisplatin pharmacokinetics and adverse reactions of a new liposomal cisplatin (lipoplatin): phase i study immunoliposomes for the targeted delivery of antitumor drugs anti-her2 immunoliposomes: enhanced efficacy attributable to targeted delivery epidermal growth factor receptor (egfr)-targeted immunoliposomes mediate specific and efficient drug delivery to egfr-and egfrviii-overexpressing tumor cells liposomal formulations of inflammatory bowel disease drugs: local versus systemic drug delivery in a rat model liposomal targeting of glucocorticoids to synovial lining cells strongly increases therapeutic benefit in collagen type ii arthritis efficacy of liposomal budesonide in experimental asthma antioxidant strategies in respiratory medicine transdermal delivery of an analgesic agent using elastic liposomes: preparation, characterization and performance evaluation preparation oral liposome-encapsulated recombinant helicobacter pylori heat shock protein 60 vaccine for prevention of hp infection topical liposomal gel of idoxuridine for the treatment of herpes simplex: pharmaceutical and clinical implications immunization with peptide-functionalized carbon nanotubes enhances virus-specific neutralizing antibody responses fullerenebased antioxidants and neurodegenerative disorders carbon nanotubes for the delivery of therapeutic molecules design and function of a dendrimer-based therapeutic nanodevice targeted to tumor cells through the folate receptor phase i and pharmacokinetic study of abi-007, a cremophor-free, protein-stabilized, nanoparticle formulation of paclitaxel detection of c-reactive protein utilizing magnetic permeability detection based immunoassays nanocrystal imaging in vivo in vivo cancer targeting and imaging with semiconductor quantum dots turning the spotlight on cellular imaging tracking metastatic tumor cell extravasation with quantum dot nanocrystals and fluorescence emission-scanning microscopy quantum dots for live cells, in vivo imaging, and diagnostics sentinel lymph node mapping of the gastrointestinal tract by using invisible light potentials and pitfalls of fluorescent quantum dots for biological imaging use of magnetic nanoparticles as nanosensors to probe for molecular interactions noninvasive detection of clinically occult lymph-node metastases in prostate cancer oligomerization of paramagnetic substrates result in signal amplification and can be used for mr imaging of molecular targets novel nanosensors for rapid analysis of telomerase activity nanoshell-mediated near-infrared thermal therapy of tumors under magnetic resonance guidance in vitro cellular accumulation of gadolinium incorporated into chitosan nanoparticles designed for neutron-capture therapy of cancer gadolinium-loaded liposomes allow for real-time magnetic resonance imaging of convection-enhanced delivery in the primate brain pet scanners dedicated to molecular imaging of small animal models in vivo molecular and genomic imaging: new challenges for imaging physics current advances in molecular imaging: noninvasive in vivo bioluminescent and fluorescent optical imaging in cancer research in vivo imaging of activated endothelium using an anti-vcam-1 magnetooptical probe key: cord-004091-gex0zvoa authors: abdulkareem, shaheen a.; augustijn, ellen-wien; filatova, tatiana; musial, katarzyna; mustafa, yaseen t. title: risk perception and behavioral change during epidemics: comparing models of individual and collective learning date: 2020-01-06 journal: plos one doi: 10.1371/journal.pone.0226483 sha: doc_id: 4091 cord_uid: gex0zvoa modern societies are exposed to a myriad of risks ranging from disease to natural hazards and technological disruptions. exploring how the awareness of risk spreads and how it triggers a diffusion of coping strategies is prominent in the research agenda of various domains. it requires a deep understanding of how individuals perceive risks and communicate about the effectiveness of protective measures, highlighting learning and social interaction as the core mechanisms driving such processes. methodological approaches that range from purely physics-based diffusion models to data-driven environmental methods rely on agent-based modeling to accommodate context-dependent learning and social interactions in a diffusion process. mixing agent-based modeling with data-driven machine learning has become popularity. however, little attention has been paid to the role of intelligent learning in risk appraisal and protective decisions, whether used in an individual or a collective process. the differences between collective learning and individual learning have not been sufficiently explored in diffusion modeling in general and in agent-based models of socio-environmental systems in particular. to address this research gap, we explored the implications of intelligent learning on the gradient from individual to collective learning, using an agent-based model enhanced by machine learning. our simulation experiments showed that individual intelligent judgement about risks and the selection of coping strategies by groups with majority votes were outperformed by leader-based groups and even individuals deciding alone. social interactions appeared essential for both individual learning and group learning. the choice of how to represent social learning in an agent-based model could be driven by existing cultural and social norms prevalent in a modeled society. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 when facing risks, people go through a complex process of collecting information, deciding what to do, and communicating with others about the effectiveness of their actions. social influence may interfere with personal experiences, making peer groups and group interactions important factors. this is especially important in understanding disease diffusion and the emergence of epidemics, as these phenomena annually take thousands of lives worldwide [1] . hence, good responsive and preventive strategies at both the individual and government levels are vital for saving lives. a choice of strategy depends on behavioral aspects, complex interactions among people [2] , and the information available about a disease [3] . perceiving the risk of an infectious disease may trigger behavioral change, as during the 2003 sars epidemic [4] . gathering information and experience through multiple sources is essential for increasing disease risk awareness about the disease and taking protective measures [5] . to help prevent epidemics, we need advanced tools that identify the factors that help spread of information about life-threatening diseases and that change individual behavior to curbs the diffusion of disease. various scientific approaches have been developed to tackle this challenge. network science is prominent in studying how epidemics propagate and how different awareness mechanisms can help to prevent the outbreak of disease. some researchers propose a framework with different mechanisms for spreading awareness about a disease as an additional contagion process [6] . others model populations as multiplex networks where the disease spreads over one layer and awareness spreads over another [7] . the influence of the perception of risk on the probability of infection also has been studied [8] . several recent studies have shown how information spreads in complex networks [9, 10] . however, a different approach is needed to account for individual heterogeneity (such as income and education levels), the richness of the information on social and spatial distance or media influence. here, a combination of modeling with data-driven machine learning becomes particularly attractive. simulation tools are commonly used to assess the effects of policy impacts in the health domain [3, 11, 12] . among the models for policy-making, agent-based modeling (abm) is recommended as the most promising modeling approach [13] . abm studies the dynamics of complex systems by simulating an array of heterogeneous individuals that make decisions, interact with each other, and learn from their experiences and the environment. the method is widely used to analyze epidemics [14] [15] [16] [17] . its advantage is in analyzing the factors that influence the spread of infectious diseases and the actions of individual actors [18] . as a bottom-up method, abm integrates micro-macro relationships while accommodating agents' heterogeneity and their adaptive behavior. it ensures that the interaction between the spatial environment and the behavior agents can integrate a variety of data inputs including aggregated, disaggregated and qualitative data [19] [20] [21] [22] . two processes are essential in representing agents' health behavior and disease dynamics, the evolution of risk perception, and selection of a coping strategy. hence, the core of a disease abm lies in defining the learning methods that steer these two processes. sensing of information (global, from the environment, and social, i.e., from other agents), exchanging information (i.e., interactions between agents), and processing of information (i.e., decision making) are critical. machine learning (ml) techniques can support these three elements and offer a more realistic way to adjust agents' behavior in abm [23] [24] [25] [26] . as more data become available in the analysis of the spread of disease, supporting abm with data-driven approaches becomes a prominent research direction. ml has the potential to enhance abm performance, especially when the number of agents is large (e.g., pandemics) and the decision-making process is complex (e.g., depending on both past experience and new information from the environment and peers). ml approaches in abm can provide agents with the ability to learn by adapting their decision-making process in line with new information. people make decisions both as individuals and as members of a group who imitate the decisions taken by the group or its leader [27] . information about social networks is becoming increasingly available, e.g., through social media analysis. it may reveal collective behavior in various domains, including health [28] . for example, people are not entirely rational and imitate others in their views about vaccines [29] . many abms rely solely on the decisions of individuals, paying little attention to group behavior [30] . yet, mirroring emotions, beliefs, and intentions in an abm with the collective decision making of crowds affects social contagion in abms [31] . agents-individuals and groups-may learn in isolation or through interactions with others, such as their neighbors [32] . in isolated learning, agents learn independently, requiring no interaction with other agents. in interactive learning, several agents are engaged in sensing and processing information and communicating and cooperating to learn effectively. interactive learning can be done in multiple ways, i.e., based on different social learning strategies [33] . agents might be represented as members of local groups, learning together and mimicking behavior from other group members (i.e., collective learning) [34] . yet, the impact of different types of interactive learning in groups compared to learning by an individual is an under-explored domain in the development of abms of socio-environmental systems. this article examines the influence of individual vs group learning on a decision-making process in abms enhanced with ml. to illustrate the implications of individual and collective intelligence in abms, we used a spatially explicit disease model of cholera diffusion [35] as a case study. bayesian networks (bns) steer agents' behavior when judging on risk perception (rp) and coping appraisal (ca). we quantitatively tested the influence of agents' ability to learn-individually or in a group-on the dynamics of disease. the main goal is, therefore, methodological: to introduce ml into a spatial abm with a focus on comparing individual learning to collective learning. the added value of the analysis of alternative implementations of learning in abms goes beyond the domain of disease modeling. it illustrates the effects of individuals learning and collective learning on the field of abms of socio-environmental systems as a whole. therefore, our main objectives are to (1) simulate the learning processes of agents on a gradient of learning from individual to collective, and (2) understand how these learning processes reveal the dynamics of social interactions and their emergent features during an epidemic. to address these objectives, the article aims to answer the following research questions: (rq1) what is the impact of social interactions on the perceptions and decisions of intelligent individuals facing a risk? (rq2) how do different implementations of group learning-deciding by majority voting vs by leaders-impact the diffusion process? (rq3) what are the implications of implementing collective learning for risk assessment combined with individual coping strategies? by answering these methodological questions for our case study, we reveal whether individuals perform better than groups at perceiving risks and at coping during epidemics. to explore the implications of intelligent learning on the gradient from individual to collective, we advance the existing cholera abm (cabm) originally developed to study cholera diffusion [35] . in cabm, mls steer agents' behavior [23, 35, 36] , helping them to adjust risk perception and coping during an epidemic outbreak. for this study, we ran eight abms to test various combinations of individual and group learning, using different information sources-with or without interactions among agents-as factors in the bns. we investigate the extent to which the epidemic spreads, depending on these different learning approaches regarding risk perception and coping decisions. s1 appendix provides a technical description of the model and the mls. below we briefly outline the processes in cabm essential to understand the performed simulation experiments. nowadays, 69 countries worldwide are labeled as cholera-endemic, with 2.8 million cases each year leading to 91,000 deaths [37] . people in urban slums and refugee camps are at high risk of cholera because of limited or no access to clean water and adequate sanitation. cabm is an empirically and theoretically grounded model developed to study the 2005 cholera outbreak in kumasi, ghana [35] . the open-source code for the model code is available online. cabm is grounded in the protection motivation theory (pmt) in psychology [23, 38] . the empirically-driven bns model a two-stage decision process of people facing a disease risk: learning to update risk perceptions (threat appraisal, bn1 in fig 1) and making decisions about how to adapt their behavior during the epidemic (coping appraisal, bn2 in fig 1) . according to pmt, threat appraisal depends on individual perceptions of the severity of the disease (evaluating the state of the environment and observing what happens to others) and one's own susceptibility. the coping appraisal is driven by the perceived response efficacy (the belief that the recommended behavior will protect) and one's own self-efficacy (the ability to perform the recommended behavior). cabm simulates individuals who are spatially located in a city. these agents differ by income and education level. individual agents form households and neighborhood groups and are susceptible to cholera at the beginning of the simulation. cabm implements an adjusted seir model [39] as explained in fig 2 below . instead of going directly from susceptible to exposure, we introduced an awareness component in which agents can assess their risk. options included: no risk perception in which the agent will be exposed (arrow 1, fig 2) ; no risk perception yet no exposure (arrow 2, fig 2) ; and risk perception leading agents to the coping phase (arrow 3, fig 2) . exposure to cholera takes place through the use of unsafe river water. agents can influence their exposure by selecting alternative water sources. these alternative water sources can either reduce their exposure to zero (arrow 5, fig 2) or have no effect on their infection risk (arrow 4, fig 2) . their actions are contingent on income and education levels, as well as on the information that they retrieve from their own experience, information received from others, or observations of the environment. it is not possible to judge by sight whether surface water is infected with cholera, but the agents use other types of visual pollution, e.g., floating garbage, as a proxy. when household agents find the visual pollution level too high, they may decide on an alternative. household agents with high incomes do not take a risk and will buy safe water. in cabm, the risk perception was updated using bn1, which depends on the agent memory (me), the visual pollution at the water fetching point (vp), and the evidence of the severity of the epidemic based on communication from the media (m) and potentially with neighbor households (cnh). media broadcast news about cholera starting on day 21 onward (see: ghana news archive). during a simulation, household agents may also interact with their neighbors zero to seven times a day (applied randomly) [40] . when interactive learning was activated, social interactions among household agents helped to share information on cholera cases that occurred in their communities and on the effectiveness of coping decisions. if risk perception was positive (bn1 returns a value above 0.5), household agents activate bn2 to decide which action (d1 -d4, fig 1) to take given their income (i) and education (e) level, the experience of their own household with cholera (oe), and possibly their neighbors' experiences with cholera (ne) [22] . s1 appendix provides further details on how the bns are implemented, together with tables of the parameters. sensitivity analysis of the aggregated model dynamics on the bns inputs and training alternatives can be found in [23, 36] . a feeling of risk among individuals is fueled by the type of information, the amount of information communicated, and the attention to specific information that may trigger fear and stimulate a learning process regarding a new response strategy [41] . gained information helps individuals (i) to estimate the severity of the emerging event, (ii) to assess the probability of being exposed to infection, and (iii) to evaluate the efficiency of their coping responses. we used a complex network approach to illustrate the gradual processes from individual to collective learning in cabm (fig 3) . each stage is presented as a single network over which a given learning process spreads. each network in fig 3 had the same set of nodes and connections to show how different processes can lead to different outcomes in the same network structure when different information is used to make decisions. in individual learning (fig 3, process 1a and process 1b), agents depend on their prior knowledge (memory, experience, and/or the perceived risk of the environment, such as visual pollution). such learning is the process of gaining skills or knowledge, which an agent pursues individually to support a task [42] . group learning is the process of acquiring new skills or knowledge that is undertaken collectively in a group of several individual agents and driven by a common goal [32] . group learning can be realized by making all group members use their own ml algorithms to gather information to perform a specific sub-task (decentralized), and then pool their opinions collectively by making one decision for the entire group (fig 3, process 2a and process 2b). here, we adopt a "majority vote" as the resolution mechanism in the decentralized group decision-making. however, group learning can also be realized by introducing a single agent (leader) who uses ml to learn for the whole group to help it accomplish its group task (centralized). in the centralized group learning, agents in the group copy the decisions of their leader. in both cases, all agents that belong to a group share the same decision, but the information on which this decision is based on varies considerably (fig 3, process 3a and process 3b). both individuals and groups may learn by either by taking information from their social networks (i.e., have it as an additional source of information in their ml algorithms) or not. when individual agents are isolated learners (fig 3, process 1a) , they do not have a social network but use only their own information to make a decision in an isolated environment using the information they possess. when individuals learn in an interactively (fig 3, process 1b) , https://doi.org/10.1371/journal.pone.0226483.g003 they gain new skills or knowledge by perceiving information, experience, and the performance of other agents through their social network. like individual agents, groups can also learn in isolation or interactively. in isolated learning, agents learn independently within their groups, without exchanging any information with each other or with their neighbors (fig 3, process 2a and process 3a). in interactive learning, agents communicate with their neighbors to learn effectively within their groups (fig 3 process 2b and process 3b). neighbors could be members of the same group or belong to other income/education groups but live in the same community and share the water collection points. therefore, there might be communication across the groups (fig 3, process 2b and process 3b). groups can be defined in different ways and at different hierarchical levels. this model uses three levels of an organization, the individual agent, groups of agents and communities that comprise several groups. in cabm, household agents living in the same community are grouped based on their income and education level since their coping behavior depends on these factors. agents' behavior in the disease abm also is contingent on their geographic location. hence, all neighbors that share the same water fetching point may contact and exchange information between their groups in cabm. the size and compilation of the groups impact the results of the different learning strategies. when applying interactive learning, a group's decision can be influenced by information retrieved from neighbors inside the group and neighbors outside the group but inside the community. for interactive groups, process 2b (fig 3) shows a situation in which individual household agents make decisions that account for interactions in their social networks (as in process 1b). then each household conducts a majority vote, allowing it to proceed with the option chosen by the majority of its members. process 3b (fig 3) shows a situation in which the leaders of each group make decisions based on their interactions with others (nodes a, b, and c are leaders of groups g1, g2, and g3 respectively in fig 3) . the decisions of group leaders are adopted by the household agent of the group. we designed eight simulation scenarios to answer the research questions about the influence of isolated vs interactive individual learning (rq1); centralized vs decentralized learning in processes-during both the risk perception (rp, bn1) and coping appraisal (ca, bn2) processes (rq2); and collective learning about risk perception combined with individual coping appraisal (rq3) on the dynamics of the epidemic and the performance of the model (table 1) . we systematically vary cabm settings following the steps in fig 4 to change the gradient of intelligent learning (steps 2 and 3) in different cognitive stages corresponding to our decisions of interest: risk and coping appraisal (step 1). table 1 shows the setup of the eight scenarios that reflects the three stages shown in fig 4. the area of the case study captured in cabm is 19.2 km 2 and comprises of 21 communities. we assumed that high-income households bought water, so they were excluded from intelligent learning. communities can have up to four groups based on their income and education levels. ten to fifteen percent of the household agents in the case study area usually fetch water from the river. two communities in our dataset (#11 and #20) hosted only high-income households, so they were excluded from the intelligent learning. hence, we simulated 76 groups spread over 19 communities. each simulation was run for 90 days with a time step equal to one hour. given the inherent randomness of abms, we ran each model for 100 times, generating a new synthetic population every 10 runs. besides the extensive gis data and aggregated data on disease dynamics, we ran a survey via a massive open online course (mooc) geohealth in two rounds (2016 and 2017) to gain data on individual behavior. the participants-primarily students from developing countrieswere introduced to the problem of cholera disease, saw pictures of water, and were asked if they would use the water as it is (d1 in fig 1) , walk to a cleaner water point (d2), or use the water after boiling it (d3). the survey data were used to construct and train our bns [36] . we also used these data to evaluate the results of expert-driven bns in cabm [43] . table 2 shows that trust in boiled water was much higher than trust in un-boiled water. agents also changed their behavior and began boiling water in the model. to evaluate the impact of individual and social intelligence on agents' learning processes regarding risk perception and coping appraisal and the resulting patterns of disease spread, we used four output measures: disease diffusion, risk perception, spatial patterns, and model performance. these aspects are described in more detail in the odd protocol (s1 appendix). we also measured the performance of models m1 -m8 in terms of run time and the number of intelligent decision steps, i.e., when agents called their bn1 and/or bn2. given the stochastic nature of abms, we ran each of the eight models 100 times. the average and standard deviations of the results of these runs for each output measure were listed in table 3 . behavioral changes can lead to different duration times of the epidemic and reduce the number of infected cases [44] . this was shown by running cabm with the eight models. models m5, m6, m7, and m8 recorded a longer duration of active infection during the epidemic (75-79 days, table 3 ). these results are closer to the real duration of the epidemic in 2005 (75 days, table 3 ). m5, m6, and m8 applied centralized learning, while m7 applied decentralized learning, but only for the risk perception stage. m2, which is individual learning with social interactions, also recorded a shorter duration when compared to the real data of 2005 (68 days in m2). however, isolated learning and decentralized learning for both risk perception and coping appraisals recorded shorter epidemic duration, with an average difference of -25% compared to the empirical data (table 3) . all eight scenarios generated more infected cases than the empirical data. this was because infection with cholera bacteria leads to a clinical spectrum that ranges from asymptomatic cases to symptomatic cholera cases. asymptomatic cases are not reported, although they represent roughly half of all cases [45] . in our simulations, we did not differentiate between symptomatic and asymptomatic cases; we considered all infected cases are considered to be symptomatic cases. therefore, following [45] , in table 3 , we reported that 57% of the total infected cases occurred when running the eight models. m8, which uses centralized learning for risk perception and individual interactive learning for coping appraisal, reported the fewest infected cases (2,107 against 1,621 in reality). this was followed by m2 (individual social learning) with 2,279 cases and m1 (individual isolated learning) with 2,457 occurrences. these three values reflect the fact that when household agents learned to cope and make decisions individually, they were more efficient than when they were in groups. when these decisions were combined with social interactions, they lead to better protection (m2 and m8). in general, group behavior had a negative effect, although centralized groups had a less negative impact compared to decentralized ones. finally, in m7, where household agents learned risk perception in decentralized groups and learned to cope individually, 2,911 infected cases were recorded (table 3) . hence, cabm household agents' engagement in decentralized groups for appraising disease risk hindered the perception of risk, lowering agents' motivation to change their behavior to more protective alternatives. the spatial distribution of infected cases (spi) of m8 reported the closest spi over the communities (0.75) compared to 1 in the empirical data. this was followed by m5, with 0.7 ( table 3 ). the spatial patterns of the two collective learning models (m8 and m5) reflected their similarity to the spatial patterns in the empirical data. the correlation between the peak of the epidemic and the peak of risk perception reflects the responsiveness of the household agents' risk perception of the epidemic. scenarios m2, m5, m6, and m8 were more responsive. that is, the peak of risk perception in m2 came three days after its epidemic peak, and the peaks in m5, m6, and m8 came seven days after their epidemic peaks (table 3) . m1, m3, m4 and m7 showed peaks for risk perception near the end of the simulation time. individuals in m1 were isolated, along with individuals in m3; therefore, they kept following their usual behavior of fetching water and using it as it is. in m4 and m7, household agents depended on majority votes in their groups to make their decisions on risk and to change behavior. more explanations are represented visually in the next sections. table 4 shows the number of steps and the time required to run one simulation of each model. the number of agents that were supposed to go for risk perception daily was 15% of the total number of household agents (which totaled 8,500). this percentage was derived from national statistical data from ghana statistical services [35] . over the 90 days of the epidemic, 114,750 agents appraised their risk perception (use their bn1). table 4 also shows the number of steps, during which agents perceived the risk of disease (i.e., risk perception equals 1). notably, in m3 -m8, if a group at large assessed the risk perception as zero, then none of its members did the coping appraisal, i.e., the number of steps when bn2 was activated is zero. in such cases, only the total number of steps with activated bn1 assessing risk perceptions was included in table 4 . models with centralized learning required the shortest computation times (table 4) . for example, m5, where only the isolated leaders with the centralized learning consult their bns, had the best performance with the shortest runtime. moreover, m5 and m6 recorded the fewest steps across all models. although the average number of agents with risk perception per simulated day was high (410 agents), there were only 6,840 steps in risk perception and the same number of steps when coping appraisal mls were activated. that is, only leaders activated their bn1 and bn2. this is only 22% compared to what it would be if agents decided individually. without voting, only one agent per group assessed the situation and made decisions. this made m5 and m6 time-efficient. on the opposite end, m4 recorded the highest computational time because of the intensive calculations required in the individual agents' network and the decentralized group network. among all models, m4 recorded the longest process time. agents individually perceived risk (bn1) before going back to their groups to negotiate a final decision on risk perception and then repeating the same individual-group sequence for the coping appraisal. in models m5, m6, and m8 only one agent per group-a total of 76 leaders-assessed risk perception daily, leading to 6,840 steps over the 90-day epidemic. in m5 and m6 only the 76 leaders also went for coping appraisal, while in m8, the group members individually assessed the coping appraisal (26,370 steps in m8 vs 6,840 steps in m5 and m6). calibration of the original model was conducted in two steps: first the hydrological submodel was calibrated, followed by a calibration of the complete model [35] . after the calibration, a stability check was performed [35] . for the current work, the objective of this epidemic model is not to reproduce the real data. it is focusing on the impact of social interactions (present or not, on the level of individual or groups) on both risk perception and coping appraisal of the individual agent. to calibrate a scenario further, one would need risk perception data for that area for the duration of the epidemic. however, such data are very scarce, not only for kumasi but worldwide. hence, risk perception was randomized at initialization. therefore, the eight models cannot be calibrated individually because they need to be comparable at initialization. s2 appendix shows the statistical analysis that was performed on the output data of the eight models to show and analyze the distribution of the obtained results. when household agents evaluated the risks of getting cholera and made coping decisions individually (m1), they relied only on their own experience. that is, each had individual bn1 and bn2 and did not communicate with neighbors. scenario m2 extends this stylized isolated benchmark case by assuming that while agents continued to make decisions individually, they table 4 risk perception and behavioural change during epidemics did share information with neighbors about the perception of risk and protective behavior. that is, both bn1 and bn2 included neighbors' experiences among the information input nodes. fig 5 shows the epidemic curves and the dynamics of risk perception for all scenarios. in the absence of social interactions, more agents became infected with cholera. the peak of the epidemic curve in m1 (in-i) is higher than m2 (in-n), leading to 11% more cases of disease ( fig 5 and table 3 ). overlaying risk perception and epidemic curves suggests that when agents made decisions in isolation (m1: in-i), the dynamics of risk perception were hardly realistic (fig 5a) . namely, when the epidemic was at its peak, household agents in m1 responded very slowly, with bn1 delivering a wrong evaluation of risk perception (fig 5a) . they became aware of the risks very late, so when the epidemic vanished, the number of agents with risk perception = 1 kept increasing. in the absence of communication and experience sharing among peers (in-i), the information about disease spread slowly and there was a significant time-lag between the occurrence of the disease and people's awareness. the small stepwise increase, around day 21, was because the media started to broadcast information about the epidemic on that day. in m2, household agents behaved according to the expected pattern: risk perception became amplified by media coverage and social interactions and then vanished as disease cases became rare (fig 5b) . only those who experienced cholera infection in their households remained alert. household agents in m2 after day 21 had more responses to the media's news compared to isolated agents. media supported the agents' social interactions with their risk perception and behavioural change during epidemics neighbors, which led to more agents perceiving risk, especially when the number of infected cases reached their peak (fig 5b) . even in m2, there were limitations of making decisions about risk perceptions individually: risk perception fell too quickly, implying that people stopped worrying about the epidemics although they continued. since household agents in m1 did not have interactions with other agents, running this model required less time than m2 (creating a 10% increase in performance, table 3 ). the interaction between household agents required time to process the information exchanged between agents. in addition, (m1: in-i) and (m2: in-n) were approximately the same in terms of the realistic spatial distribution of infected cases over the communities, with values of 0.65 and 0.66, respectively (table 3) . fig 6 presents the spatial distribution of decision types over the study area in both m1 (in-i) and m2 (in-n). the household agents in isolated learning were not aware of the cholera-infected cases in their neighbors' household. household agents in m1 took an unsecured decision and trusted more in using the water fetched from the river as it is (d1 in fig 6a) . household agents in m2 were more rational and mostly boiled the water that they fetched from the river (d3 in fig 6b) . in decentralized learning, groups of household agents vote for risk perception and coping appraisal. the final decision of the group is the output of the majority votes. thus, all group members follow the final decision of the group. these groups represent the democratic system, which depends very much on the composition of the group. the decentralized groups with a majority vote can lead to a negative perception of risk. besides, a coping appraisal that depends on a majority vote can lead to inappropriate decisions regarding protection from cholera. when individuals are engaged in social groups, their behaviors are not independent anymore https://doi.org/10.1371/journal.pone.0226483.g006 [46] . this leads to an increase in the randomization of decentralized learning models (m3 and m4). these two models had higher standard deviations in all measures ( table 3) . the qualitative patterns of the three scenarios (m3, m4, and m7) were the same regardless of the social interactions that added new information to ml (fig 5) . for the development of the disease, the voting mechanisms seemed to overwrite individual judgments. the m3 scenario assumes that household agents were isolated when performing risk perception and coping appraisals. in contrast, m4 and m7 allowed household agents to communicate with neighbors during the process of risk perception and before making a coping decision. as a result, m4 and m7 generated greater risk perception than m3 (fig 5c, 5d and 5e ). this suggests that the social interactions still amplify both the awareness of risks and the diffusion of preventive actions. given approximately the same peak heights, the epidemic curves in the three majority voting scenarios reported more infected cases than the other models. among the majority votes, m7 reported the fewest infected cases, since household agents in their coping appraisal relied on themselves rather than their decentralized groups. overall, it seems that all three models-m3, m4, and m7 -got the process of disease risk evaluation wrong. in those cases, risk perception slowly grew in the days when the epidemic was peaking (fig 5c, 5d and 5e) and did not react to the peak in any way, which is unrealistic. moreover, risk perception in the three models continued to grow when the epidemics were almost over. risk perception peaked when there was no longer a risk, i.e., in the last days of the simulation, as shown in table 3 . hence, group voting on risk perception operated with a major time lag: household agents ignored early signals of disease that occurred in just a few households. then they increased their awareness about risk only when most of them were already infected, and they continue to be falsely alerted when the epidemic was over. in m3, the small stepwise increase in risk perception represents the response to media, and it is similar to m1 (in-i) in its development (fig 5c) . the household agents in their decentralized groups did not have contact with neighbors, therefore, no cases were reported to them from their neighborhoods. as such, they were disconnected from what is happening around them. in m4 and m7, which included social interactions, the development of risk perception seems more responsive, especially after the activation of media on day 21. nevertheless, their response time was still slow (fig 5d and 5e ). in these models, the group decisions were very much dependent on the composition of the group members' opinions. these varied from one another and had different information sources for the final decisions about risk perception (in both m4 and m7) and coping appraisal (in m4). thus, majority voting led to unsecured decisions. groups in these models were heterogeneous in that household agents had different levels of exposure to the group members with which they voted. decentralized groups with isolated input information (m3) led household agents to vote to use the water fetched from the river (d1) most of the time (fig 7, map a) . because of their lack of communication with neighbors, household agents missed the opportunity to get information about the infection in their neighborhoods. this explains the higher numbers of infected cases in the majority vote models. social interactions in both m4 and m7 helped agents make better decisions, although following the majority still biased their choices. for instance, in m4 high-income communities (upper communities in maps b and c, fig 7) , household agents mostly used the river water as it was even though they were rich enough to boil it before using it (d3) or to buy bottled water (d4). the opposite also occurred when a majority vote forced low-income households to buy bottled water, which is an expensive decision for them. the group voting on the coping appraisal in m4 might have made individual members uncomfortable when they followed the decisions of their groups even though they might not protect. in reality, household agents sought a balance between preventive behavior and their capability to implement it. moreover, there is always the possibility of routinely changing one's mind based on daily updates of information regarding the epidemic and updates from neighbors. as in m4, the household agents in m7 relied on their decentralized groups for risk perception. this, often led to risk ignorance (fig 5e) . however, since the agents in m7 decided on coping appraisals individually, more agents adopted d1 (fig 7c) . when they perceived risk during the last days of the epidemic, household agents in the middle-income level switched to boiling water or buying bottled water (d3, d4 in fig 7c) . those in the low-income level walked to another water fetching point (d2). in centralized groups, one household agent is randomly selected to be the group leader. the leader is responsible for risk perception and the coping appraisal of the group. group members copy the risk perception and disease preventive decisions of their leaders. it is argued that group leaders may improve their group's performance if they model the responses to the situation the group faces [47] . in this article, we considered two types of leaders: a dictator making top-down decision about risk perception and coping strategy (m5 and m6), and an opinion leader evaluating risk perception top-down but giving group members the freedom to pursue their own disease coping behavior (m8). the qualitative trends of all three models coincided with what is expected: peaks caused by amplification of risk perception followed by a gradual decrease when epidemics plateau (fig 5f, 5g and 5h) . the centralized group learning on average represented the processes well, as the leader alerted the group members about the disease. however, since no real data are available on risk perception dynamics or the actual coping behaviors that people pursued during the epidemic, we cannot determine which of the models m5, m6, and m8 is the best. the following subsections compare models with a leader-dictator (m5, and m6) to one with an opinion leader (m8). a dictator-leader decides on behalf of his group regarding disease risk and coping strategies, and both decisions are adopted top-down. a dictator leader learns either in isolation (m5) or in interaction with her/his neighbors (m6). isolated dictators in m5 are overestimated disease risks (fig 5f) . for example, if such a leader had his/ her own bad experience with cholera, s/he would keep warning the group. with social interactions (m6), there is less uncertainty in the process of updating the risk perception than in m5. for example, compare risk perception assessments around the epidemic peak (fig 5g) . fig 8 illustrates the impact of social interactions on the dictator's decisions regarding coping appraisal. isolated leaders guided their groups to various types of decisions (fig 8a) , which were sometimes less secure decisions (e.g., d1). with social interactions, leaders relied on their neighbors and decided more often to walk to a point along the river where the water was cleaner (d2). very few dictators directed their groups to boil the fetched water (d3) or buy bottled water (d4) (fig 8b) . this shows how centralized decisions making undermines heterogeneity in individual circumstances, such as disease exposure or coping capacity. in m8, the leaders in the centralized groups were responsible for evaluating disease risks for their groups, but they interacted with neighbors during the risk perception process. for the coping appraisal, the group members made their own decisions, using the information from their social networks. as a result of this combination of centralized speed alertness about risk perception and individual coping strategies, m8 generated the fewest infections. the shape of the epidemic curve (except for its height) is very close to the empirical data of 2005, (fig 5h) . as in m6, the uncertainty in the process of risk perception in m8, is lower than in m5 (fig 5h) . the risk perception curve developed around the epidemic peak followed the dynamics of the epidemic (fig 5g) . when group members relied on social interaction to learn about the effectiveness of various coping strategies but eventually chose one themselves (m8), there was a diversity of coping strategies. fig 8c shows the spatial distribution of different types of decisions during the simulation. more household agents went for d3 and d4, which were considered to be the most protective decisions. consequently, communities pursued at least three types of decisions, reflecting the disease coping diversity so important for resilience. the goal of this paper is to perform a systematic comparison of individual vs group learning. the methodological advancements showed that different implementations of individual and collective decision-making in agents' behavior led to different model outcomes. in particular, the stepwise approach of testing how learning (on a gradient from individual learning-without any interactions-to collective-with social networks) affects an abm's dynamics is generic and can be used for other models. to illustrate the subtle difference in implementing learning in abms, we used the example of the spatial empirical abm of cholera diffusion with intelligent agents that employ ml to assess disease risk and decide on protective strategies, which define the dynamics of the epidemic. interactive learning, which assumes that agents share information about risks and potential protective actions, outperformed isolated learning both for individuals and in groups. this underlines the fact that social learning in the decision-making process is very important in abms. while we used disease modeling as a case study, the results may be contingent on the endogenous dynamics of this particular cholera abm. notably, simulation results may differ for abms with other underlying dynamics. this calls for further scrutiny in testing and reporting cases of intelligent social and individual learning in other models. the results indicate that decentralized groups with majority votes are less successful than groups with leaders, whether dictators or opinion leaders. when evaluating current disease risks, majority voting appears to be the worst mechanism for group decisions, often arriving at a wrong decision because of time lags compared to the dynamics of objective disease risks. perceiving risk is a very personal decision-making process [48] . in contrast, when leaders develop risk perception and propose it to the group, such groups perform better in terms of risk appraisal. moreover, opinion leaders are very effective in helping their group members be alert about disease while giving them the freedom to make coping decisions that accommodate heterogeneity in their socio-economic status and geographical locations. in contrast, dictatorleaders and majority votes that impose a decision that all group members must follow are less effective in reducing the incidence of disease. in our simulation experiments, the structure of the groups is simple and is formed based on the spatial and socio-demographic characteristics of the agents. as grouping seems to have an impact on the spatio-temporal diffusion of the disease, the importance of disease modeling stresses the fact that for this type of model a careful evaluation of the social structures in the case study area should be conducted, to generate trustworthy results. future research should focus on constructing groups based on different variables (family ties, religion, tribes). also, in our abm the leaders had no particular knowledge but were randomly selected and assigned to groups. in reality, this may not be the case. leaders may have access to better information or have already earned the group's trust and respect. in addition, decentralized groups can be improved by giving greater weights to more trusted partners to make wise decisions. the model's performance can be a strong argument when the number of agents is massive, e.g., when simulating a pandemic or epidemics within a very large population is needed to detect a worldwide diffusion mechanism. in that case, social group learning, as described in model m5, is a very good alternative to individual interactive behavior. moreover, m5 shortens the computation time by 73% while maintaining a good quality model output. the number of contacts each household agent has when they are in their collective learning may impact the diffusion of cholera. however, running a fat tail distribution of the number of contacts would be an interesting topic for future study. different considerations steer the ultimate decision on which type of social behavior to use. besides the technical model performance metrics discussed here, the choice of a particular type of social behavior can also be based on the society that is being modeled. different political systems, the presence of tribes, and different ethnic groups or religious leaders require careful considerations of the social interactions in a model. one should make sure that the actual situation regarding social learning represents the cultural and social norms of the society being modeled. in this article, it was not possible to define, which implementation (m1 -m8) represented the situation in kumasi most closely. to validate the risk perception-behavior, one would need risk perception data for that area for the duration of the epidemic. however, such data are very scarce, not only for kumasi but worldwide. as we illustrated in this study, many different implementations of social behavior using ml are technically possible, but data are needed to validate alternative implementations. yet, research on risk perception during epidemics is often conducted too late (when the peak is over) or at distance (not in the area where the disease spreads). hence, researches provide little empirical proof of people's behavior and risk perception. more research on risk perception during epidemics, including other variables such as cultural aspects and group behavior, can be very helpful in generating a model that represents a specific society realistically. on a technical note, agent-based modeling software does not always include ml toolkits and libraries. this complicates the implementation of different types of social intelligence. hence, better integration of abm and ml in one software package or linkable libraries could eliminate this problem in the future. finally, an important direction of future research is to implement other ml techniques besides bns, such as decision trees and genetic algorithms. in addition, modeling groups with different ml algorithms may lead to different results since groups will be heterogeneous in terms of members' learning algorithms. several developments in health research drew our attention to the implementation of learning in disease models. one is the impact of fake news on the behavior of people. the other is the fact that human behavior toward vaccination can change radically based on (fake) news it. therefore, including these factors and testing their impact on the behavior of agents may lead to more conclusions for policymakers to consider in their efforts to control epidemics. managing epidemics: key facts about major deadly diseases. world heal organ learning from each other: where health promotion meets infectious diseases modeling infection spread and behavioral change using spatial games severe acute respiratory syndrome epidemic and change of people's health behavior in china the role of risk perception in reducing cholera vulnerability towards a characterization of behavior-disease models dynamical interplay between awareness and epidemic spreading in multiplex networks epidemic spreading and risk perception in multiplex networks: a self-organized percolation method spreading processes in multilayer networks interacting spreading processes in multilayer networks epidemic and intervention modelling-a scientific rationale for policy decisions? lessons from the 2009 influenza pandemic wrong, but useful": negotiating uncertainty in infectious disease modelling models for policy-making in sustainable development: the state of the art and perspectives for research using data-driven agent-based models for forecasting emerging infectious diseases out of the net: an agent-based model to study human movements influence on local-scale malaria transmission modelling the transmission and control strategies of varicella among school children in shenzhen a taxonomy for agent-based models in human infectious disease epidemiology an open-data-driven agent-based model to simulate infectious disease outbreaks modeling human decisions in coupled human and natural systems: review of agent-based models spatial agent-based models for socio-ecological systems: challenges and prospects agent-based models global sensitivity/uncertainty analysis for agent-based models intelligent judgements over health risks in a spatial agent-based model a multi-agent based approach for simulating the impact of human behaviours on air pollution simulating exposurerelated behaviors using agent-based models embedded with needs-based artificial intelligence agent-based modeling in supply chain management: a genetic algorithm and fuzzy logic approach evidence for a relation between executive function and pretense representation in preschool children scalable learning of collective behavior based on sparse social dimensions the impact of imitation on vaccination behavior in social contact networks agent based simulation of group emotions evolution and strategy intervention in extreme events modelling collective decision making in groups and crowds: integrating social contagion and interacting emotions, beliefs and intentions learning in multi-agent systems simulate this! an introduction to agent-based models and their power to improve your research practice do groups matter? an agent-based modeling approach to pedestrian egress agent-based modelling of cholera bayesian networks for spatial learning: a workflow on using limited survey data for intelligent learning in spatial agent-based models. geoinformatica the global burden of cholera a comprehensive review of the applications of protection motivation theory in health related behaviors seasonality and period-doubling bifurcations in an epidemic model social contact structures and time use patterns in the manicaland province of zimbabwe cognitive and physiological processes in fear appeals and attitute change: a revised theory of porotection motivation. social psychophysiology: a sourcebook artificial intelligence: a modern approach. third edit integrating spatial intelligence for risk perception in an agent based disease model resilience management during large-scale epidemic outbreaks susceptibility to vibrio cholerae infection in a cohort of household contacts of patients with cholera in behavioral modeling and simulation risk perception and human behaviors in epidemics risk perception it's personal. environmental health perspectives. national institute of environmental health science key: cord-021917-z9wpjr0d authors: stephens, r. scott; wiener, charles m.; rubinson, lewis title: bioterrorism and the intensive care unit date: 2009-05-15 journal: clinical critical care medicine doi: 10.1016/b978-0-323-02844-8.50069-x sha: doc_id: 21917 cord_uid: z9wpjr0d nan • no country is fully prepared to avert illness when large portions of the population are covertly exposed to a serious bioweapons agent; a moderate or large-scale intentional release of a serious pathogen will likely cause lifethreatening illness in a large portion of exposed people. • intensivists will play a key role in the medical response to a bioterrorism event due to the clinical conditions caused by serious bioweapons pathogens, such as severe sepsis, septic shock, hypoxemic respiratory failure, and ventilatory failure. • compared with conventional disasters, bioterrorist attacks may not be readily recognized; thus, accurate clinical diagnoses and management on the basis of clinical suspicion are critical not only for appropriate care of individual patients but also for instituting an epidemiological investigation. • victims of a bioterrorist attack who require intensive care unit-level care may be more contagious than those who are less sick. • health care workers, accustomed to putting the welfare of patients ahead of their own in emergency situations, must be prepared for the proper use of personal protective equipment and trained in specific plans for the response to an infective or bioterrorism event. tive prophylactic countermeasures exist, a large portion of exposed people will likely develop life-threatening illness. although intensivists working in developed countries generally have little experience treating specific illnesses caused by serious bioweapon pathogens, these diseases result in clinical conditions that commonly require treatment in intensive care units (icus) (e.g., severe sepsis and septic shock, hypoxemic respiratory failure, and ventilatory failure). therefore, intensivists will play a key role in the medical response to a bioterrorism event. usual critical care practices will likely require modification for any event resulting in more than a few critically ill victims, and critical care specialists should participate in planning for such situations. capabilities to provide medical care, especially critical care, services to large numbers of contagious patients are very limited in most countries. local hospitals will be expected to care for seriously ill victims of a bioterrorist (bt) attack, and the ability to care for large numbers of critically ill patients will likely be a major determinant of the medical impact of such events. although the current risk of a large-scale bt event is uncertain, a number of groups throughout the world during past decades have deliberately exposed civilians to biologic agents. fortunately, none of these prior events produced a large number of casualties because of the nonlethal nature of the pathogens released (e.g., salmonella typhimurium in oregon in 1984), the lack of technical expertise to successfully disseminate lethal pathogens (aum shinrikyo in japan in 1994), or relatively limited exposure (the anthrax cases of 2001 in the united states). these limited exposures should not result in predictions of the numbers of potential casualties being reduced; rather, they simply reveal that an increasing number of groups throughout the world are willing to use biologic agents. the scope and effect of prior events could have been much greater if a contagious agent were used or if a lethal pathogen were widely disseminated. despite the increased attention to biodefense, the risk of subsequent bt events may paradoxically be increasing. rapid advancements in science are making synthetic and novel biologic agents more accessible and technologies to disseminate agents may no longer be restricted to only a few nations. to reduce the medical effect of a bt event, the major determinants of morbidity and mortality must be understood ( fig. 64 .1). the number of deaths from a bt event depends in part on the lethality and infectivity of the released agent, in addition to how effectively and widespread it is delivered. many biologic agents could theoretically be used as weapons, but some are an intentional release of a biologic agent within a civilian population, exposing hundreds or thousands of people to a serious pathogen, is increasingly recognized as a plausible terrorism event. unlike most mass casualty incidents, releases of bioweapons agents may be covert, thus providing additional security and public health challenges to the medical response beyond the generic burden of scores of casualties. an optimal medical response to a bioweapon attack will require all or most of the following: early diagnosis, rapid case finding, large-scale distribution of countermeasures for postexposure prophylaxis or early treatment, immediate isolation of contagious victims, and enhanced capacity for providing medical care to seriously and critically ill victims. no country is fully prepared to avert illness when thousands of people are covertly exposed to a serious bioweapons agent. hence, after a moderate or large-scale intentional release of a serious pathogen, even one for which effec-more lethal, more available, and more easily disseminated (table 64 .1). the agent's characteristics alone will not, however, determine the overall impact of the bt event. population characteristics (i.e., vulnerabilities) will also affect the impact of a bt event ( fig. 64.1 ). many agencies, professions, and community members must be involved in preparing and responding to a bt event. to optimally respond, hospital and public health cooperation, planning, and preparedness need to occur before the disaster. the integration and coordination of all these responders is very important, and detailed operational descriptions are beyond the scope of this chapter (see box 64.1). instead, this chapter is intended to be an introduction to the medical response issues for a bt event, specifically the critical care medical response. the centers for disease control and prevention (cdc) has compiled a list of potential agents of bioterrorism and divided these into three categories, a-c (see table 64 .1). category a agents are those believed to pose the greatest threat in terms of potential lethality, ability for widespread dissemination, ability for subsequent human-to-human transmission, and disruptive impact on the community and the public health system. these agents include variola major (smallpox), bacillus anthracis (anthrax), yersinia pestis (plague), clostridium botulinum (botulism), francisella tularensis (tularemia), and viral hemorrhagic fevers (vhfs). this section provides a brief summary of the pathogenesis and diagnosis of each category a agent and also reviews current recommendations for treatment. bacillus anthracis is acquired from contact with infected animals or animal products and causes three forms of disease: cutaneous, inhalational, and gastrointestinal. approximately 2000 cases of cutaneous anthrax occur annually worldwide. inhalational anthrax has not occurred naturally within the united states since 1976; any case must therefore be considered a possible sentinel case of a bioterrorist event. gastrointestinal anthrax is uncommon in developed countries and is not discussed here. cutaneous and inhalational anthrax are the forms expected following an aerosol release of spores, with the latter being the most lethal. a 1993 estimate by the u.s. congressional office of technology assessment predicted that an aerosolized release of 100 kg of "weaponized" anthrax over a populated city would cause 130,000 to 3 million deaths, similar to the mortality of a thermonuclear detonation. inhalational anthrax results from spore particles 1 to 5 mm in diameter entering the alveolar spaces and being transported by macrophages to mediastinal lymph nodes. after an incubation period that ranges from 2 days to 7 www.acponline.org/bioterro www.upmc-biosecurity.org www.shea-online.org www.cdc.gov weeks (median of 4 days during the 2001 cases and up to 60 days after the release of spores in sverdlovsk in 1979), germination occurs, with the vegetative bacilli producing two toxinslethal toxin and edema toxin. initial symptoms of inhalational anthrax are nonspecific: fevers, chills, drenching sweats, nonproductive cough, dyspnea, nausea, vomiting, and fatigue. hemorrhagic thoracic lymphadenitis and mediastinitis develop, and hemorrhagic pleural effusions with compressive atelactasis are common. hemorrhagic meningitis may also occur. patients may rapidly develop hemodynamic collapse, which typically has been refractory to treatment if it develops prior to antimicrobial administration. diagnosis is predicated on a high index of suspicion. in the 2001 attacks, all patients with inhalational anthrax had an abnormal chest radiograph or thoracic computed tomography scan. mediastinal widening due to lymphadenopathy and large bilateral pleural effusions were the most common features (figs. 64.2 and 64.3). these findings in the setting of a previously healthy patient with the abrupt development of sepsis should raise suspicion of anthrax infection. sputum gram stain and culture are rarely positive. blood cultures may yield a diagnosis but require hours to days to grow the organism. as in many infections, blood cultures lose diagnostic utility if obtained after antibiotic administration. hemorrhagic meningitis was common (50% of patients) during the sverdlovsk incident but was only confirmed in 1 of 11 patients in 2001. patients suffering from inhalational anthrax are likely to require icu care but do not require respiratory isolation. antibiotics must be started as soon as possible without waiting for diagnostic confirmation. treatment recommendations are summarized in table 64 .2. for adults, combination antimicrobial therapy with intravenous ciprofloxacin and one or two other agents is recommended. given the potential for meningitis, agents with good central nervous system penetration, such as rifampin, penicillin, or chloramphenicol, are recommended. clindamycin has been administered for the theoretical benefit of reducing toxin production by the vegetative bacilli, although this has been done in too few instances to critically evaluate its effectiveness. a change to oral therapy is acceptable once the patient is stable. therapy should continue for 60 days. the concomitant use of an anthrax vaccine in a modified dosing regimen (three doses within the first month) may be limited by availability. there is anecdotal evidence that drainage of pleural effusions carries some benefit. whether the benefit is simply reduction of pleural fluid volume to improve oxygenation or actually helps to reduce toxin burden remains uncertain. during the 2001 outbreak, pleural drainage was accomplished via serial thoracenteses and tube thoracostomy. optimal management may require tube thoracostomy due to the hemorrhagic nature of the fluid. historically, inhalational anthrax had a mortality rate of approximately 90%. in the 2001 attacks, modern critical care interventions and use of multiple antimicrobial agents reduced mortality to 45%. cutaneous anthrax results after the inoculation of skin with anthrax spores. these patients are unlikely to require icu-level care if they are treated promptly. two exceptions are the possibilities of airway compromise due to a neck lesion with resulting edema or postoperative management of a compartment syndrome. local edema is the first feature of the condition, with the subsequent appearance of a macule or papule that rapidly ulcerates and develops into a painless black eschar ( fig. 64.4 ). systemic disease, including lymphadenopathy and lymphangitis, can follow. in the absence of antibiotic therapy, mortality has been reported to be as high as 20%; death is rare if adequate treatment is instituted. between 1347 and the early 1350s, y. pestis, the causative agent of plague, swept through europe, eventually killing 20 to 30 million people-one-third of the population. plague continues to occur naturally as an insect-borne illness, infecting approxi-bioterrorism and the intensive care unit mately 1700 people annually worldwide. in the united states, most cases occur in the rural states of the southwest. plague occurs in three forms: bubonic, septicemic, and pneumonic. naturally occurring bubonic plague occurs when infected fleas bite a human and typically results in enlarged lymph nodes (bubo) and severe sepsis. a smaller percentage of patients may develop sepsis without bubo formation, and this is termed primary septicemic plague. rarely, patients with bubonic or septicemic plague develop pneumonia, and this is termed secondary pneumonic plague. patients with pneumonic plague can transmit disease through respiratory droplets. those who contract pneumonic plague from person-to-person transmission are considered to have primary pneumonic plague and do not develop buboes. both primary and secondary pneumonic plague are transmissible from person to person. the intentional release of aerosolized plague would result in primary pneumonic plague, a condition that is rare in naturally occurring plague. world health organization (who) estimates from 1970 predict that 50 kg of y. pestis released over an urban area with 5 million inhabitants would cause pneumonic plague in 150,000, with 36,000 fatalities. exposure is followed 1 to 6 days later by fever, dyspnea, and cough with bloody, watery, or purulent sputum. gastrointestinal symptoms also occur. cervical buboes are rare. pneumonia progresses rapidly with unilateral or bilateral infiltrates or consolidation. severe sepsis and septic shock develop with leukocytosis, multisystem organ failure, and disseminated intravascular coagulation. in the absence of therapy, irreversible shock and death occur 2 to 6 days after exposure. a bioterrorist attack with aerosolized plague would likely present as an outbreak of severe pneumonia and sepsis. diagnosis depends on standard microbiologic studies, with confirmatory tests available only at select laboratories. hence, unless epidemiologic clues alert health care workers that these patients do not have community-acquired pneumonia, the first group of patients will likely be cared for with the hospital's usual infection control policies for this condition. unless droplet precautions are commonly used and strictly adhered to, a number of health care workers and additional patients may be exposed to plague early in the outbreak. therapeutic recommendations for pneumonic plague appear in table 64 .2. streptomycin and gentamicin are the first-line agents. doxycycline, ciprofloxacin, and chloramphenicol are alternative choices. in the event of a mass casualty situation, or for postexposure prophylaxis, doxycycline or ciprofloxacin are the preferred agents for adults. francisella tularensis is an extremely infectious pathogen; exposure to as few as 10 organisms can cause tularemia. naturally occurring throughout north america, europe, and asia, tularemia is transmitted to humans via arthropod bites, contact with small mammals or contaminated food, and inhalation. tularemia can take many clinical forms (box 64.2). disease manifestations depend on virulence, dose, and site of infection. the disease can begin in the skin, starting as a papule and resulting in an ulcer, and also involve regional lymph nodes (ulceroglandular). if contaminated water is ingested or contaminated droplets are inhaled, pharyngeal ulcers with cervical lymphadenitis can occur (oropharyngeal). the eyes can be the initial portal of entry leading to chemosis and lymphadenitis (oculoglandular). sometimes, lymphadenitis may occur without ulceration (glandular). inhalational tularemia may also occur naturally due to aerosolization of contaminated materials. this clinical scenario is the most likely after an intentional release, since aerosol release would be the most likely method of dissemination. who estimated that 50 kg of aerosolized f. tularensis in a city of 5 million would affect 250,000 people and cause 19,000 deaths. after an incubation period of 1 to 21 days, abrupt fever develops, accompanied by influenza-like symptoms (headache, chills, rigors, myalgias, coryza, and pharyngitis). bronchiolitis, pleuropneumonitis, and hilar lymphadenitis would be expected, although inhalational tularemia can often present as a systemic disease without respiratory features. progressive weakness, fever, chills, malaise, and anorexia rapidly incapacitate victims. hematogenous spread can lead to pleuropneumonia, sepsis, and meningitis. sepsis due to tularemia may manifest as severe sepsis or septic shock. mortality without antibiotic therapy can be as high as 30% to 60% for pneumonic and septic tularemia. current antimicrobial therapy results in a mortality rate of less than 2%. rapid diagnostic testing for tularemia is not widely available. the constellation of atypical pneumonia, pleuritis, and hilar lymphadenopathy in association with the previously described symptoms should raise suspicion for tularemia. in the wake of a bioterrorist attack, until a number of patients present, initial diagnosis may be delayed. most diagnoses are made serologically with a fourfold rise between acute and convalescent antibody titers. antibodies are slow to develop: titers will usually be negative at 1 week, positive at 2 weeks, and peak in 4 to 8 weeks. laboratories need to be specifically notified if tularemia is suspected, both to improve diagnostic accuracy and to protect laboratory workers. polymerase chain reaction (pcr) and antigen detection are rapid and available at reference laboratories in the united states through the laboratory response network. if the reference lab is alerted when specimens are sent, an answer can be given within hours. treatment recommendations for tularemia are presented in table 64 .2. in the event of a contained casualty situation, streptomycin is the preferred drug, with gentamicin as a first-line alternate. doxycycline, chloramphenicol, and ciprofloxacin are acceptable alternates. for mass casualties and for postexposure prophylaxis, oral doxycycline and ciprofloxacin are the recommended agents. treatment with aminoglycosides or fluoroquinolones should last 10 days; tetracyclines and chloramphenicol require a 14-day course. botulinum toxin, produced by the bacteria c. botulinum and a few other clostridium species, is the most potent known neurotoxin. botulinum toxin inactivates proteins necessary for the release of acetylcholine into the neuromuscular junction. the toxin could be disseminated as an aerosolized agent or as a food contaminant. fewer than 200 naturally occurring cases of botulism occur annually in the united states. the use of botulinum toxin by terrorists would result in inhalational botulism or foodborne botulism, depending on the mode of dispersal. the neurologic signs are identical regardless of whether the toxin enters the body via the lungs or the digestive tract. intestinal botulism may be preceded by gastrointestinal complaints. symptoms appear approximately 12 to 72 hours after exposure. botulism presents as an acute, symmetric, descending flaccid paralysis. there is no associated fever, and the bulbar musculature is always affected first. presenting complaints and findings are related to cranial nerve palsies and include diplopia, dysarthria, dysphonia, and dysphagia. hypotonia and generalized weakness ensue. loss of airway protection may necessitate intubation, and respiratory muscle paralysis may require mechanical ventilation. the course is variable and may require months of mechanical ventilation. during small outbreaks with sufficient medical resources, serial measurement of vital capacity can help identify patients with respiratory muscle weakness. elevation of paco 2 is a late finding, and positive pressure ventilation should be instituted prior to frank ventilatory failure. notably, cognitive function is not affected; patients are completely awake and alert. a fever should raise suspicion of secondary infection or an alternative diagnosis. the diagnosis of botulinum intoxication is clinical and is classically described as the triad of symmetric cranial neuropathies with descending paralysis, clear sensorium, and lack of fever. other diagnoses to consider are listed in box 64.3. in developed nations, the occurrence of a number of temporally related cases of acute paralysis points to botulinum intoxication. the edrophonium or "tensilon" test may be transiently positive in botulism, although it still may be helpful in distinguishing it from myasthenia gravis. csf is generally normal in botulism. an electromyogram demonstrating an incremental glandular ulceroglandular oculoglandular oropharyngeal pneumonic septic response with repetitive stimulation at 50 hz may suggest botulism when the conduction velocity and sensory nerves are normal. culture of stool or gastric contents (for foodborne exposure) may yield clostridium. confirmation usually requires the mouse bioassay (mice are exposed to samples and those given polyvalent and specific antitoxin survive) but takes several days. samples for the mouse bioassay must be collected prior to the patient's receiving antitoxin. a clinician who suspects botulism must immediately notify local public health authorities to aid with epidemiologic and diagnostic investigations. most laboratory testing cannot be performed at hospitals. laboratories must be notified of suspicion regarding botulism, since samples can be potentially harmful to laboratory personnel. specific therapy consists of treatment with equine antitoxin (see table 64 .2), which will not reverse extant paralysis but may prevent progression. it must therefore be administered as early as possible. in mass casualty situations, when the antitoxin supply may be limited, patients with weakness but not yet requiring mechanical ventilation may be the most appropriate for antitoxin therapy. supportive therapy is essential, with a specific focus on mechanical ventilation and efforts to prevent complicating events (e.g., ventilator-associated pneumonia, venous thromboembolism, and decubitus ulcers). nonventilated patients should be placed in the reverse trendelenberg position at 20 to 25 degrees to optimize respiratory muscle function and minimize the possibility of aspiration aminoglycosides and clindamycin should be avoided because of the potential for exacerbating the neuromuscular blockade. mortality for foodborne botulism averages 6%. the last naturally occurring case of smallpox was identified in 1977, and the last case (due to a laboratory accident) was in 1978. despite worldwide eradication, smallpox continues to concern biodefense experts due to uncertainties about available stocks of the virus. despite the mortality rate of smallpox (30%) being considerably lower than those of other bioweapons agents, its potential for harm is still very high because those who survive may be severely deformed or blinded and no proven specific therapy exists once exposed people become symptomatic. in addition, with cessation of worldwide vaccination, entire populations and especially younger persons are susceptible to infection. the agent of smallpox, the variola virus, belongs to the orthopoxvirus family. these viruses are quite stable in the environment, and hence an aerosol may be widely dispersed. any case of smallpox would be an emergency and must be considered to be the result of a deliberate act. the typical incubation period for smallpox is 7 to 17 days, with an average of 12 to 14 days for the majority of patients. initial symptoms include fever, rigors, backache, and headache. vomiting and delirium may develop in this prodromal phase. two or 3 days later, a nonspecific erythematous rash begins. the rash first appears in the mouth and throat, with red spots appearing on the buccal and pharyngeal surfaces. the usual dictum is that a person is not contagious until the rash begins. in most patients, the macular lesions become papular followed by vesicles. the lesions then become pustular, which umbilicate and are deeply seated in the dermis. the crops of lesions appear at the same time and are all at the same stage on the affected part of the body (fig. 64 .5). they usually begin and are more concentrated on the face and limbs rather than the trunk. this is in contrast to primary varicella, in which lesions on any given part of the body are in different stages of development (some macular, some vesicular, and some crusting) and in which the rash begins on the trunk and moves outward. after 8 to 10 days, scabs form at the sites of the pustules. in survivors, these become depressed depigmented scars. smallpox lesions also occur on the palms and soles, which rarely occurs with chickenpox. the previous description is seen in more than 90% of smallpox cases, but there are less common forms of smallpox as well. hemorrhagic smallpox is uniformly fatal (it tends to affect pregnant women more frequently), and it typically has a shorter incubation period and does not lead to the classic rash. instead, death follows development of a hemorrhagic rash. in the malignant or flat form of smallpox, the disease begins classically but does not progress to pustules. instead, the rash is confluent and may desquamate. also, variola minor is a less severe form of smallpox. suspicion of smallpox must initially be based on clinical findings; the possibility of this disease must be considered in any patient displaying fever and a characteristic centrifugal and uniform rash. definitive diagnosis requires specialized diagnostic techniques. electron microscopy can determine whether the virus is an orthopox, and confirmatory pcr techniques require primers specific to variola. laboratory confirmation will likely be required for the sentinel cases of an outbreak. after initial cases are confirmed, additional case identification can be based on clinically consistent criteria. miller-fisher variant of guillan-barré syndrome myasthenia gravis tick paralysis atropine poisoning paralytic shellfish/puffer fish poisoning if an exposure to smallpox is suspected but the patient is asymptomatic, administration of the vaccinia virus within a few days from exposure can prevent or greatly diminish the severity of the illness. once the disease develops, however, specific therapeutic options are limited (see table 64 .2). there is evidence that cidofovir may have activity against the variola virus, although the evidence is based on alternative orthopox disease models and in vitro assays. supportive care for critically ill patients may limit mortality, but since the last case occurred more than 25 years ago, it is uncertain what effect modern critical care will have on outcomes. mortality is expected to be approximately 30%, with far greater rates of disfigurement or disability. secondary transmission is most likely to occur through close contact with symptomatic patients (e.g., droplet and contact transmission), although fomite and airborne transmission have been documented. patients with a cough may be more likely to transmit droplet nuclei (i.e., airborne transmission), and those with atypical disease courses (e.g., hemorrhagic and malignant) may be more difficult to identify, so unprotected exposure of health care workers may be more likely. the viral hemorrhagic fevers believed to be possible agents of bt are listed in box 64.4. the filoviruses and arenaviruses are transmissible from person to person. although the limited information available suggests that transmission is primarily via infected body fluids, mucosal transmission has been documented in experimental animals and airborne or droplet transmission has been suggested in several outbreaks. clinical manifestations will vary with the particular virus. in general, the vhfs have an incubation period ranging from 2 to 21 days (commonly, 5-10 days). initial symptoms are nonspecific and may last up to 1 week. fever, malaise, headache, myal-gias, arthralgias, nausea, and gastrointestinal complaints are prominent. a rash may be present. on exam, patients are typically febrile, relatively bradycardic, hypotensive, and tachypnic. as the disease progresses, hemorrhagic manifestations, such as petechiae, mucosal bleeding, hematuria, hematemesis, and melena, may appear. eventually, disseminated intravascular coagulation, multiorgan system failure, and shock may develop. mortality rates vary greatly, but in the case of ebola virus they may be as high as 90%. confirmatory diagnosis of vhfs must be made at specialized laboratories. the diagnosis must be suspected in any patient presenting with acute fever, severe illness, and hemorrhagic manifestations. any patient who presents with a vhf who does not have a travel, contact, or exposure history consistent with the known natural occurrence of these illnesses must be considered as the possible victim of a bt attack. unfortunately, vhfs have a high lethality and supportive therapy is the only treatment (see table 64 .2). there is evidence that ribavirin may have some effect against arenaviridae and bunyaviridae. no therapies have been shown to be effective against the filoviruses or flaviviruses. additionally, there is no recommended agent or vaccine for postexposure prophylaxis to any of the vhfs. many of the operational functions for the response to bioterrorism events are similar to those for other disasters (intentional or natural), such as requiring coordination and communication among a number of government agencies, professions, citizens, and community stakeholders for the response. however, bioterrorism events pose specific challenges for the medical response that may be serious enough that if not addressed may shut down hospitals and leave many victims without adequate options for health care (table 64 .3). recognizing that a conventional disaster has occurred is usually immediate, and these events are limited both geographically and temporally. casualties have traumatic injuries, and a large portion of the survivors are taken to the nearest health care facility. within hours to a few days, the number of expected casualties is usually known. immediate death rates may be high, especially with structural collapse, but typically only a small bioterrorism and the intensive care unit fraction of survivors are critically ill (injury severity score > 15). enclosed space explosions may lead to higher proportions of survivors with critical injuries, but the absolute number of critically ill patients is usually less than 100. after the initial chaotic response period, medical staff and equipment are usually not in short supply. if local hospitals are overwhelmed, additional staff and resources can be transported to the disaster area, or patients can be evacuated to unaffected areas. the recovery plans for the affected health care facilities are usually initiated within the same day or a few days following a conventional disaster. unlike conventional disasters, a release of a bioweapons agent may go undetected. in such situations, exposed people would present for medical care after the incubation period has passed. since people travel extensively in developed countries, and most incubation periods are days to weeks, patients are likely to present to a number of hospitals rather than to the facility located closest to the release. having patients distributed to a number of hospitals may lead to delayed recognition that a bt event has occurred. in addition, most diseases resulting from serious bioweapons agents initially cause symptoms and signs that are commonly seen every day in hospital emergency departments and outpatient clinics. there may be no pathognomonic signs that a bioterrorist event occurred in the sentinel patients initially presenting with respiratory failure or hemodynamic collapse. no diagnostic tests are available to help clinicians rapidly diagnose most diseases, so a bt event may go unnoticed until scores of ill victims arrive at hospitals. the presentation of multiple previously healthy patients with unusual and severe symptoms should prompt suspicion. because of the specialized diagnostic techniques required for these organisms, and the biosafety precautions that are frequently beyond the capabilities of most hospital-based clinical laboratories, confirmatory diagnostic testing for the category a agents in the united states is handled at laboratories of the national laboratory response network, which includes local and state labs as well as federal facilities, such as the u.s. army medical research institute of infectious diseases and the cdc. there will necessarily be a delay in final diagnosis because samples for confirmatory testing must be sent to off-site laboratories. this increases the importance of accurate clinical diagnoses and proceeding with management on the basis of clinical suspicion. prompt diagnosis is critical not only for appropriate care of individual patients but also for instituting an epidemiological investigation. the community or nationwide response hinges on the results of this rapid investigation. once the source, agent, and location of a bt attack have been deduced, other clinicians can be notified, resources can be mobilized, and the at-risk population can receive postexposure prophylaxis. the difficulties of recognizing initial exposures to a bt attack have profound implications for hospital functionality, particularly if the pathogen is contagious. health care workers (hcws) and other patients without adequate infection control protections may be exposed to contagious patients. during an outbreak of severe acute respiratory syndrome (sars), unprotected exposure of hcws and hospitalized patients to patients with sars was thought to be the major risk to a hospital remaining open. most victims of serious bioweapons attacks (e.g., anthrax, plague, smallpox, botulism toxin, tularemia, and vhfs) will develop illness that is rapidly progressive (ultimately requiring mechanical ventilation, hemodynamic support, or other aggressive therapeutic interventions) if they do not receive early medical intervention or if no disease-specific medical countermeasures exist. these critically ill patients will also likely require extended critical care for survival. few hospitals can provide even usual critical care services for an additional 100 critically ill bt victims, especially if the pathogen is contagious. in the aftermath of a bt event, it may be very difficult to ascertain the extent of the exposure. incubation periods have a range, so the first cases may simply represent the tails of the gaussian distribution, and many more patients may require care in the following days. some ill patients may go unrecognized as cases, and patients may arrive at hospitals in a larger geographical area than is typical after a conventional disaster. the initially affected area may be quickly overwhelmed because of shortages of critical care resources. the unaffected regions may choose to wait to offer help until the size of the event becomes better delineated so that they do not send staff and resources away until they are certain they were unaffected. furthermore, if the pathogen is contagious, resources in affected areas may be more rapidly overwhelmed and unaffected regions may be even less likely to provide help. bt attacks resulting in a disproportion of critically ill victims to available icu beds are plausible. if such an event occurred today, many critically ill patients would have to forgo potentially life-sustaining critical care interventions. hospitals can plan to give traditional standards of critical care to the few who are fortunate to arrive early during the event, or they can modify critical care so that more patients have access to some of the most important critical care interventions (e.g., mechanical ventilation). methods to decide who should get critical care (e.g., triage algorithms), what critical care interventions should be provided, who should provide critical care, and where critical care should be provided need to be addressed before a bt event. through such planning, hospitals may "gracefully degrade" services rather than ceasing to function when overwhelmed. all efforts must be made to provide disease-specific therapies to victims of bioterrorism. unfortunately, not all of the serious bioweapons agents have effective treatments, and for those with treatments there is concern about development of antimicrobialresistant strains. systems must be in place for testing new treatments during an outbreak so that effective treatments can be rapidly communicated to other clinicians and ineffective or harmful treatments can be rapidly withdrawn. methodological issues, ethical concerns, skeleton protocol development, and information technology systems capable of making data rapidly available for analysis should all be developed and made functional before a disaster. patients seriously ill due to a bioweapons agent, regardless of whether a specific therapy exists, will likely require extensive supportive care, including interventions traditionally provided in icus. for small-scale events with few critically ill patients, traditional icu care will likely be provided. for larger events, decisions regarding which supportive care practices to continue and which to forgo will depend on the number of patients relative to the available resources. supportive care that is deemed most important can be better maintained if advanced planning and preparedness are undertaken. icu physicians should alert their hospitals to the potential need for rapid acquisition of additional mechanical ventilators, noninvasive respiratory aids, oxygen, palliative medications, and specialized staff in the event of a bt attack. perhaps the most critical aspect of caring for victims of a biologic attack or an emerging infective disease in an icu is the prevention of secondary transmission. in the sars outbreak of 2003, 77% of cases in canada resulted from in-hospital exposure. in taiwan, the percentage of hospital-acquired cases was 94%. these data include other patients in the hospital who contracted the disease as well as health care workers who suffered occupational exposures. category a biologic agents that are transmissible from person to person are listed in box 64.5. effective infection control measures are paramount in preventing the spread of disease through the hospital and, by extension, into the community. infection control is particularly important in the icu, in which a "perfect storm" for the rapid spread of an infection exists. victims of a bioterrorist attack who require icu-level care may be more contagious than those who are less sick due to higher levels of viremia or bacteremia. invasive procedures with their attendant risks of splashing or aerosolization of blood, respiratory secretions, or other bodily fluids are more commonly performed on critically ill patients. staff members in an icu are often called on to rapidly complete a number of tasks in stressful conditions, a situation conducive to errors in infection-control practices. since critically ill patients require a high level of frequent care, cumulative exposure to staff may be higher than in other areas of the hospital. finally, other patients in the icu are immunocompromised by virtue of their own critical illnesses, notwithstanding the disproportionate number of icu patients who are immunosuppressed secondary to organ transplantation, oncologic conditions, or infection with the human immunodeficiency virus (hiv). one of the mainstays of management of a chemical incident is rapid and effective decontamination of victims. decontamination serves both to limit the total dose of chemical agent received by the victims and to protect health care workers from remnant chemicals on patient skin or clothing. patients will not present for medical care after release of a biologic agent until the incubation period passes. decontamination is not necessary for these patients, since they are not likely to be grossly conta-minated at the time of presentation. for an overt attack, if the patient is grossly contaminated and there is concern about secondary aerosolization, it becomes reasonable to decontaminate the patient. since t2 mycotoxin can be transdermally absorbed, decontamination of patients grossly contaminated with this agent is also warranted. if possible, symptomatic victims of a communicable bioterrorism agent should be placed in a private room to prevent exposure to other patients. in the case of smallpox or a viral hemorrhagic fever, rooms should be under negative pressure and equipped with high-efficiency particulate air (hepa) filtration. the exhaust air from these rooms should be expelled directly to the outside, and the ventilation system should not be shared with other areas of the hospital. documented cases of smallpox transmission have occurred through ventilation systems. although the number of airborne infection isolation (aii) rooms in most hospitals is few, there are engineering modifications to increase modified aii capacity during an outbreak. planning for these modifications before an event is critical. assuming a large outbreak of disease, patients should be grouped together not only with respect to location but also with respect to nursing staff, physicians, and equipment. if no diagnostic test exists, care must be taken to minimize exposure of uninfected patients with similar signs or symptoms who may be inadvertently housed in the same location. although friends and family undoubtedly bring much comfort and support to critically ill patients, in the face of an infectious disease they become both potential victims and potential vectors. visitors of victims of bioterrorism with contagious agents, or victims of an emerging infectious disease, must be kept to an absolute minimum. they must be instructed and supervised in the use of proper protective equipment and notified that they must seek treatment immediately if they develop symptoms. in extreme circumstances, it may be necessary to completely preclude friends and family from visiting patients. health care providers are accustomed to putting the welfare of the patient ahead of their own; patient care, particularly in emergency situations, is often carried out without adequate protective equipment. this cannot be allowed in the case of extremely contagious agents, even in an emergency or "code" situation. a health care provider who has contact with a patient without suitable protective gear risks not only his or her own health but also the health of other patients, coworkers, visitors, and their own families. individual patient care issues must be secondary to adequate infection control practices, lest an epidemic of smallpox, sars, plague, or viral hemorrhagic fever spread unchecked. the cdc has developed categories of precaution that are to be applied to patients with potentially communicable diseases (box 64.6). these categories have been described at length elsewhere but are summarized in the following sections, along with their applicability to the category a biological agents. standard precautions should be applied to all patients and include measures designed to prevent transmission of bloodsmallpox viral hemorrhagic fevers pneumonic plague cutaneous anthrax borne illnesses such as hiv and hepatitis b/c. most interactions with patients do not require any protective equipment, but gloves, gown, and face shield should be used for any activity that could potentially result in an exposure to blood or bodily fluids. of the category a agents, anthrax, tularemia, and botulinum toxin require only standard precautions because these diseases are not transmissible from person to person. cutaneous anthrax should perhaps be treated with contact precautions because transmission has been suggested following contact with the lesions of this type of anthrax. contact precautions are applicable to diseases that can be spread by touching the patient directly or indirectly by coming into contact with contaminated objects. common examples include scabies, herpes, clostridium difficile, and methicillin-resistant staphylococcus aureus. contact precautions must be used, if applicable, during all patient interactions, regardless of whether body fluid contact is expected. protective equipment consists of gloves and gown, and a face shield is mandatory if splashing or spraying of body fluids is possible. patients with smallpox and vhfs must be placed in contact precautions. patient care equipment must also be dedicated to these patients and not used on patients not suffering from these diseases. droplet precautions apply to diseases that are transmissible by large-particle droplets, defined as those greater than 5 mm. due to the size of the droplets, transmission is highest over short distances (<1 m) and does not occur through ventilation systems. necessary equipment includes a face shield or surgical mask with eye protection, gown, and gloves. pneumonic plague requires droplet precautions. airborne precautions are required for diseases that are spread via droplet nuclei, which are less than 5 mm. tuberculosis is the most familiar of airborne infectious agents, but of the category a agents, both smallpox and vhfs fit into this category. droplet nuclei may travel through ventilation systems, underscoring the importance of placing patients with these diseases in negativepressure rooms with hepa filters and exhaust of air to the outside. required equipment includes gown, gloves, and adequate respiratory protection. either an n95 respirator (specifications are described elsewhere) with eye protection or a powered air purifying respirator (papr) is acceptable. a misconception exists that once a patient is intubated and mechanically ventilated, both large droplets and droplet nuclei are no longer expelled into the air. this is true only if the ven-tilator expiratory circuit is fitted with a filter that meets hepa guidelines. unfortunately, many of the filters and heat/moisture exchanging filters commonly used do not meet hepa criteria. this poses dangers to health care personnel, visitors, and other patients not only for bioterrorist agents but also for tuberculosis, sars, and other emerging infections. hospitals would be well advised to stockpile hepa-grade filters for ventilator expiratory circuits. correct hand washing is an essential component of hospital infection control in all circumstances. this is perhaps even more true in the circumstances of an outbreak of an emerging infectious disease or possible agent of bioterrorism. hands must be washed after each patient contact even when protective gloves are used because a surprisingly high percentage of protective gloves contain microscopic holes, and holes may develop during the activities of routine patient care. failure to thoroughly wash hands following patient contact places other patients and health care workers at risk. there has been an increase in the use of waterless alcohol rubs in icus rather than soap and water. although these gels are generally effective against bacteria and viruses, they have been shown to be ineffective against bacterial spores such as anthrax. soap and water, antimicrobial or not, are effective at removing anthrax spores from hands. accordingly, we recommend the use of antimicrobial soap and water for the washing of hands after patient contact. complete recommendations for postexposure prophylaxis are given in table 64 .1. in the case of anthrax and tularemia, postexposure prophylaxis with vaccination has not been proven effective. there is no need for prophylactic antibiotic therapy for health care workers unless they were potentially exposed in the initial attack. vaccination does not confer protection against pneumonic plague. health care workers caring for patients with pneumonic plague should, however, receive prophylaxis with 7 days of oral doxycycline. if symptoms such as fever develop, they should be aggressively treated with parenteral antibiotics. health care workers caring for patients with smallpox should be vaccinated as soon as possible because vaccination within 4 days of exposure can prevent or limit the severity of subsequent illness. immediate isolation must follow the development of fever after exposure to smallpox. no postexposure prophylaxis exists for vhfs. although decontamination of victims of a biological attack is rarely necessary, the rooms in which they are treated can become contaminated with infectious organisms, particularly if sprays of bodily fluids or respiratory aerosols are produced. virions of smallpox, in particular, can persist in linens for extended periods of time; documented cases exist in which laundry workers contracted smallpox from handling contaminated bedding and clothing. the causal agents of vhfs may also be transmitted via contaminated linens. commercial hospital disinfectants and household bleach at a 1:10 dilution are effective at eliminating surface contamination with anthrax, smallpox, ebola and marburg viruses, tularemia, and plague. linens from infected patients should be incinerated, autoclaved, or the bodies of deceased patients with smallpox, plague, or vhfs continue to pose an infectious risk. autopsies or postmortem examinations should be avoided if possible. the bodies of victims of smallpox should be cremated. people who have died of a vhf should preferably be cremated, although prompt burial without embalmment is a secondary option. proper use of personal protective equipment is essential in order to protect staff from infectious disease. the equipment required depends on the particular disease, as described previously. effort must be made to ensure that adequate supplies of equipment exist, and requirements will likely be far greater than expected. calculations estimating the amount of equipment necessary for one nurse caring for four patients with a communicable disease are striking and sobering. during an 8-hour shift, one nurse would likely require 64 sets of personal protective equipment: 64 pairs of gloves; 64 gowns; 64 surgical masks, n95 masks, or papr hoods; and 64 face shields. providing for the needs of physicians, respiratory therapists, and other ancillary personnel increases this equipment need markedly. beyond the availability of adequate stocks of equipment, health care workers must be adequately trained in their uses. n95 respirators require fit testing annually. the equipment must be used as designed, including donning and removing it correctly. removing equipment in the proper order is particularly important: the gloves must be removed first to avoid contamination of the face or clothing when removing the gown, mask, and eye shield. unfortunately, this correct sequence is not widely appreciated by health care workers. given the complexity of caring for victims of a bt attack or an emerging infectious disease, early planning for such an incident must be carried out at the institutional and icu levels. how will a hospital, or an icu, care for a potentially massive influx of patients with communicable disease or specific requirements? plans will differ in their specifics depending on each hospital and each icu's architecture and capabilities. general principles include the following: patients should be grouped according to infection, not necessarily by need. these patients should then be isolated from the remaining hospital population and staff. dedicated physicians, nurses, ancillary personnel, and equipment should be used so as to prevent exposure to other patients and keep the numbers of exposed health care workers to a minimum. the scale of a bioterrorist attack could potentially be enormous. as described previously, many of the category a agents could produce the same lethality as a nuclear explosion. the number of casualties could rapidly overwhelm any one hospital or even all of the hospitals in a community. staff safety must be paramount; if staff members believe themselves to be at risk, large numbers of nurses, physicians, and others may not show up to work, crippling or even forcing the closure of a hospital. although all risk cannot be avoided, all possible provisions must be made. all staff members must be trained in the proper use of personal protective equipment. beyond that, training should be provided in specific plans for the response to an infective or bt incident. physicians and nurses in particular should be educated with regard to possible agents of bt and presented with disease-specific issues. contingency plans must be made in advance for postexposure prophylaxis, either with antibiotics or vaccinations, as indicated. given the critical and limited time windows to initiate effective prophylaxis, plans for distribution of medication or vaccine must be thought through in advance and not made in an ad hoc manner. a large-scale bioterrorist incident could rapidly exhaust the resources of individual hospitals or even whole communities in a number of respects. the demand for personal protective equipment will be enormous. beyond that, there will be a need for pharmaceuticals of all types. antibiotics will be essential, but so will medications regularly employed in the icu setting: vasopressors, sedatives, narcotics, and others. mechanical ventilators may be at a premium for patients with acute respiratory distress syndrome, pneumonia, or ventilatory failure secondary to botulism. although in the united states, and likely in other countries, the federal government has assembled stockpiles of antibiotics, smallpox vaccine, and mechanical ventilators, it will take time for these to be deployed, and they may not contain all essentials. the prospect of bioterrorism is not an abstract one. it has been attempted before, and it will certainly be attempted again. that it has not happened is not reason for complacency. adequate response to a bioterrorist incident is possible, but it requires careful and thoughtful preparation. also, the preparation is hardly specific to the potential agents of bioterrorism. the principles of providing safe and effective care to victims of a bt incident are wholly applicable to the management of patients suffering from naturally acquired emerging infectious diseases. the age of bioterrorism is also that of sars, coronavirus, avian influenza, ebola virus, and, significantly, global travel. the potential for patients to present acutely ill with rare, unknown, and infectious diseases, whether naturally acquired or unleashed by criminals, has never been higher. planning and preparation are essential. botulinum toxin as a biological weapon. medical and public health management hemorrhagic fever viruses as biological weapons. medical and public health management tularemia as a biological weapon. medical and public health management the challenge of hospital infection control during a response to bioterrorist attacks smallpox as a biological weapon. medical and public health management plague as a biological weapon. medical and public health management updated recommendations for management key: cord-318683-1yxurnev authors: green, manfred s; leduc, james; cohen, daniel; franz, david r title: confronting the threat of bioterrorism: realities, challenges, and defensive strategies date: 2018-10-16 journal: lancet infect dis doi: 10.1016/s1473-3099(18)30298-6 sha: doc_id: 318683 cord_uid: 1yxurnev global terrorism is a rapidly growing threat to world security, and increases the risk of bioterrorism. in this review, we discuss the potential threat of bioterrorism, agents that could be exploited, and recent developments in technologies and policy for detecting and controlling epidemics that have been initiated intentionally. the local and international response to infectious disease epidemics, such as the severe acute respiratory syndrome and west african ebola virus epidemic, revealed serious shortcomings which bioterrorists might exploit when intentionally initiating an epidemic. development of new vaccines and antimicrobial therapies remains a priority, including the need to expedite clinical trials using new methodologies. better means to protect health-care workers operating in dangerous environments are also needed, particularly in areas with poor infrastructure. new and improved approaches should be developed for surveillance, early detection, response, effective isolation of patients, control of the movement of potentially infected people, and risk communication. access to dangerous pathogens should be appropriately regulated, without reducing progress in the development of countermeasures. we conclude that preparedness for intentional outbreaks has the important added value of strengthening preparedness for natural epidemics, and vice versa. the biological weapons convention prohibits the manufacture and use of biological weapons. it came into force in 1975, and has undergone periodic reviews, the last being in 2016. to date, 180 countries are signatories to the convention. unfortunately, terrorist groups or rogue governments are unlikely to feel bound by international agreements. the potential for bioterrorism is of particular concern, since it can cause disease, death, and panic-in great disproportion to the resources expended. 1 there have been a few well documented cases of bioterrorism. in 1984, a religious sect in the usa deliberately contaminated restaurant salad bars with salmonella typhimurium, intending to disrupt local elections. 2 the attack resulted in several hundred cases of salmonellosis and no deaths. the anthrax letters incident in 2001 in the usa resulted in 11 cases of inhalation anthrax, with five deaths, and another 11 cases of cutaneous disease. 3 extensive circumstantial evidence strongly suggests that the perpetrator was a civilian employee of the us military. however, no evidence of a clear motive was found. thousands of workers received prophylactic or post-exposure therapy, and affected buildings were decontaminated at huge expense. 4, 5 in 1993, a cult in japan carried out an attack using anthrax spores with no physical casualties, 6 but later, evidence of post-traumatic stress syndromes was found in victims of the attack. 7 the perpetrators were apparently planning to use other agents such as q fever bacteria, botulinum toxin, and ebola viruses, 8 but they were detained before they could implement further attacks. in this review, we discuss the threat of bioterrorism, potential perpetrators, and general preparedness principles. we examine the special characteristics of biological agents that could potentially be used for bioterrorism, advances in prevention and treatment of diseases caused by these agents, and the remaining deficiencies in the management and control of possible bioterrorist outbreaks. in all respects, the ways in which the resources developed for bioterrorism preparedness could be used for controlling naturally occurring epidemics remain a guiding principle. • preparedness for intentional outbreaks will strengthen the response to naturally occurring epidemics • high level leadership should be maintained with responsibility and authority • health-care providers should maintain awareness of biological agents with bioterrorism potential and consider the presence of unknown pathogens • emergency room and community physicians should be updated regularly about the clinical manifestations of diseases caused by potential bioterrorist agents and emerging infectious diseases. • personal protective equipment should be improved to become more user friendly • improved surge capacity (the ability to rapidly gear up the health system to cope with a sudden, large increase in patients with a serious, contagious disease) is required, particularly in peripheral areas • the capacity of general and reference laboratories should be increased, to keep developing faster, more reliable diagnostic tests • new and improved vaccines (pre-exposure and post-exposure) and treatment regimens should be developed • clinical and environmental surveillance needs to increase • syndromic surveillance systems can be maintained to register suspicious or confirmed cases reported by physicians, and the data can be used to improve risk communication programmes and to monitor the progress of an outbreak • an adequate stockpile of vaccines and medications should be maintained, both nationally and internationally • to improve preparedness for natural and bioterrorist outbreaks, international cooperation should include joint exercises involving multiple countries and constant improvement in the exchange of information on potential bioterrorism threats and management following the breakup of the former soviet union, there was concern that loss of control of their biological weapons programme could allow terrorist groups to gain access to both the weapons and scientific expertise. additionally, in the past few years, developments in the field of microbial genetics have heightened concern about the possible abuse of new technologies. since there are so many unknowns, it is extremely difficult to assess the risks and threats of bioterrorism. 9, 10 the most likely perpetrators could be disgruntled individuals, terrorist organisations, or rogue countries that are believed to support international terrorism. whereas individual attackers are unlikely to cause mass casualties, terrorist organisations could pose a substantial threat if they gain access to sophisticated biological weapons, materials, or scientific expertise. although regulations and safeguards for securing dangerous pathogens in research laboratories now exist in most countries, the scope of these regulations and the extent of the safeguards vary. 11 rogue countries have the necessary capabilities for a bioterrorist attack but might be restrained by the threat of the response of a unified global community. knowledge gained from legitimate research that could also be applied to bioterrorism 12 is considered dual-use. as a result, the regulation of legitimate research on infectious diseases has increased. there will always be a risk of the "insider threat", 12,13 which typically involves a single individual, so it is important to assure that new regulations truly increase security and have minimal negative effect on legitimate research. the cost of regulations applied to research on infectious diseases, in terms of missed opportunities for international collaboration, exchange of pathogens, and sharing of novel agents, is often intangible and overlooked. it is essential to promote healthy organisational cultures to enhance both safety and security in laboratories. 14 since a bioterrorist attack is a low-risk, high-impact event, effective and sustained preparedness is an essential component in both the deterrence and management of an attack. a bioterrorist attack has a lot in common with naturally occurring public health emergencies resulting from infectious diseases. however, there are some important differences. since it is a deliberate act to cause harm, there are the obvious security considerations. the resulting outbreak differs in some important ways from naturally occurring epidemics-for instance, it is more likely to be a point source outbreak initiated by simultaneous exposure to many people. the infectious agent used is likely to be uncommon and possibly not endemic to the region, might have been modified genetically to make it resistant to current medications and vaccines, and produced in a way that enhances its transmission or virulence. therefore, early clinical symptoms and signs after infection with a bioterrorist agent might be unusual, complicating both recognition and management of the disease. these factors could create greater public panic. despite the many similarities with naturally occurring outbreaks of infectious disease, preparedness for bioterrorist attacks is more complex. in many aspects, a bioterrorist attack has the characteristics of a mass casualty event, and thus preparedness involves strengthening of the specialised infrastructure that is required for treatment of seriously ill patients over a very short period of time. new prophylactic and treatment regimens for unusual diseases are required, to ensure their accessibility when needed, along with clear standards for the handling and study of dangerous pathogens. when the proportion of available resources poor international preparedness • delay before the who declared a public health emergency of international concern • delay in implementing coordinated international assistance • logistic challenges in delivering support to assist epidemic response • shortcomings in who's regional and country-level capacity exposed • lack of global plans to address an epidemic of a high-risk pathogen in the least developed urban centres • evaluation of promising vaccine and therapeutic interventions came too late 16 and ebola virus. 17 the local and international responses to the 2014 west african ebola virus epidemic revealed shortcomings that could allow highly contagious epidemics of infectious disease to spread widely before they are terminated (panel 1). 17 during the cold war, agents that could potentially be used as biological weapons were identified on the basis of the following characteristics: pathogenicity for humans, animals, or plants; ability to cause disability or death; stability and infectivity as small particle aerosols; and capability of being readily and rapidly produced and weaponised in munitions or delivery systems. more characteristics have been added, to include other features of biological agents such as the relative ease of medical prevention or treatment and the likelihood of harm to the perpetrator. the us centers for disease control and prevention (cdc) identified bacteria, viruses, and toxins that could potentially be weaponised (panel 2). in 2002, they categorised them into three groups-a, b, and cdepending on ease of dissemination, severity of illness caused, and ability to cause death. 18 biological agents can be infectious and contagious, infectious but not usually contagious, or toxins if they are neither. category a agents were considered the greatest risk to public and national security. the more recent classification of tier 1 select agents and toxins is similar to the category a classification (table 1) . other agents, such as naturally occurring pathogens, produce diseases that are considered of intermediate risk to the public (eg, brucellosis, glanders, q fever). they are moderately easy to disseminate, and include emerging and re-emerging infectious diseases. however, genetic modifications could make them more virulent, produce uncharacteristic clinical signs, increase their resistance to treatment and vaccines, and even change their transmissibility or host range. genetic modifications could be made using the tools of synthetic biology; such activities might be an example of dual-use research. 19, 20 for instance, in 2005, the 1918 spanish influenza pandemic virus was reconstructed, 21 and the poliovirus was synthesised nearly 20 years ago. 22 the addition of an immuno-modulatory gene to the mousepox virus genome in 2001, 23 rendered a mousepox vaccine ineffective, and this technology could potentially be applied to the smallpox virus. 24 the recent synthesis of the extinct horsepox virus 24 has been a reminder that the smallpox virus could be reconstructed, and that the regulations that have been put in place to prevent the misuse of powerful, cheap, and globally available tools must be reconsidered. 25 this possibility has also raised the issue of whether research results should sometimes be censored, or even refused publication, if the potential to cause harm is too high. 19 although bioterrorist agents could be disseminated through multiple routes, the aerosol route would likely maximise exposure. contagious agents could produce a large number of second and later generation cases, depending on the number of people initially exposed, the series average number of people who acquire the disease from one infected individual (r 0 ), and the disease generation time in humans. for instance, the r 0 of pneumonic plague has been estimated to be around 1·3, 26 whereas the r 0 for smallpox is likely to be around 5. 27 for diseases that are not contagious, such as inhalational anthrax, the number of cases of disease will depend almost entirely on the size of the population exposed and the timing of post-exposure antibiotic prophylaxis. 28 aerosolised agents remain the threat of most concern, but safety and security of food [29] [30] [31] and water supplies 32 are also important components of primary prevention (panel 3). new methods to detect toxins in food, such as antibody based assays, are being developed. 33 rapid diagnostics take on additional urgency in a bioterrorist event, because of both health and security concerns. 34 since the 2001 anthrax letters, there have been major advances in diagnostic capabilities. some of the greatest advances in the past decade have been in the speed and reduced cost of sequencing capabilities. 35, 36 highly sensitive and specific pcr-based systems, coupled with modern sample preparation technologies, have enabled sequencing technologies to become less costly, more portable, and multiplexed. with fieldable patient-side diagnostics and sequencing outputs directly connected via cloud-based networks, 37 health-care providers globally can make decisions more rapidly and respond more quickly for individual care or outbreak detection. a rapid, cartridge-based assay for francisella tularensis has been developed for use at point of care. 38 a system that uses a sensitive microsphere technology to detect both antibodies and antigens is now available to diagnose infections with ebola virus and lassa virus. 39 although diagnostic elisa tests are available for anthrax antibodies, 40,41 a compact system (genexpert) that includes both sample processing and pcr amplification can produce a result in about 90 minutes. 42 a rapid and sensitive method to detect smallpox virus has been developed for use at point of care, based on antibody immuno column for analytical processes (abicap) immunofiltration, that produces results in about 45 minutes. 43 however, diagnostic electron microscopy is also still considered a fast and efficient method 44 to identify smallpox and other viral agents. ebola virus was rapidly sequenced during the outbreak in sierra leone to link sporadic cases with the transmission chains. 35 advanced proteomics are also being developed as reference assays 45 and a new method for simultaneous immunodetection of anthrax, plague, and tularaemia from blood cultures has recently been reported, 46 using multiplexed suspension arrays. next generation safety and security of food [29] [30] [31] and water supplies 32 are important components of primary prevention: • intentional contamination of food should be considered, particularly during a large foodborne epidemic with a common source • salmonella and shigella species, enterohaemorrhagic escherichia coli (all serotypes), vibrio cholerae, cryptosporidium parvum, and noroviruses are all potential candidates for intentional contamination of food • contamination of water with biological agents should still be considered, even though it is unlikely to be the major target of bioterrorism, due to chlorination, dilution, and the need for large quantities of the agent to cause a substantial outbreak • c parvum and noroviruses are more resistant to chlorination than other agents, so can be a threat to the water supply • food-borne or water-borne dissemination of these biological agents might lead to higher rates of morbidity and case fatality than previously observed, if the population has been exposed to substantially higher infectious doses • algorithms could be developed to measure the likelihood that outbreaks of disease were a consequence of intentional contamination of food or water, using descriptive, analytical, and molecular epidemiologic tools (none are known to be available so far) 47 increased networking and collaboration of laboratories will also improve the response to intentional outbreaks. effective global surveillance of infectious diseases is essential to control both intentional and naturally occurring epidemics. 48 surveillance data can be used to monitor the progress of an outbreak, and for risk communication. to obtain information rapidly, the ongoing collection of health-related data (termed syndromic surveillance) has been introduced, to monitor patterns of symptoms and signs that are suggestive of an outbreak. [49] [50] [51] [52] although it was hoped that syndromic surveillance would be a more sensitive method for early detection of an epidemic, frequent reports of unusual increases in incidence of non-specific illnesses can desensitise and paralyse the system. 53 in fact, early detection will depend largely on alert, prepared clinicians. for example, when the 2001 anthrax attacks occurred, an astute clinician identified the index case. 34 emergency room and community physicians should be updated regularly on the clinical signs and symptoms associated with the most common bioterrorist agents. 54 the syndromic surveillance system would be more useful after suspicious or confirmed cases have been reported by physicians. a focused analysis of the surveillance data against non-specific, background disease rates could detect changes and provide information about the dynamics of the disease. special legislation might then be necessary, to gain access to medical records. 55 the internet facilitates other potential forms of surveillance and communication about infectious diseases. 56, 57 one such system is promed, which was established by the user community and has proven effective in connecting clinicians and scientists around the world; it has already served as an early warning system 58 standard contact and airborne precautions should be taken. 65 supportive therapy and antibiotics can be provided for secondary infections. there is some evidence of the potential efficacy of thiosemicarbazones. cidofovir has shown in vitro efficacy against variola, and has shown efficacy against other diseases caused by human orthopoxviruses, notably diseases caused by vaccinia viruses. it has also shown efficacy in animal models of orthopoxvirus infections. since 2012, the us food and drug administration (fda) has approved drugs and biologic agents developed under the animal rule. this rule allows for approval of a drug that cannot be tested for efficacy in humans, but is effective in animals and safe in humans. the first drug approved under this rule was the monoclonal antibody raxibacumab for treatment of inhalation anthrax. tecovirimat is a drug that inhibits all orthopoxviruses tested in vitro. it was found to be highly effective in treating monkeypox and rabbitpox in animals and is considered safe in humans. 66 tecovirimat is being considered by the fda for approval for use in humans to treat smallpox under the animal rule. standard and droplet precautions should be taken. 65 ciprofloxacin, levofloxacin, and doxycycline have been approved for the treatment of pneumonic plague. streptomycin and gentamicin have been found to be effective in treatment, although there is some evidence of the development of multiple resistance. 67, 68 tularaemia isolation of the patients is not necessary and standard precautions should be taken. ciprofloxacin, levofloxacin, and doxycycline are all approved for the treatment of tularaemia. streptomycin and gentamicin have been found to be effective. 68, 69 haemorrhagic fevers standard contact and airborne precautions should be taken until diagnosis is confirmed. subsequently, droplet precautions can be considered. supportive care and treatment of secondary infections can be provided. ribavirin is now approved for treatment of lassa fever and it also appears to be effective against new world arenaviruses and crimean-congo haemorrhagic fever. protective n95 respirators and clothing should be provided to health-care personnel. clothing of patients should undergo decontamination and thorough handwashing. supportive therapy is available, with antibiotics such as ciprofloxacin, doxycycline and ampicillin. if the bacteria are resistant to some of the antibiotics, the treatment regimen will depend on sensitivity testing. 60 the regulations require immediate reporting of serious health risks by all member countries. additionally, who has established the global outbreak alert and response network, and the european union has a programme called bichat to improve cooperation between member states in preparedness and response to biological and chemical attacks. they operate the early warning and response system for outbreaks of communicable diseases. the world organisation for animal health has developed plans to identify and deal with a bioterrorism attack on populations of food-producing animals. canada has established the global public health intelligence network for worldwide monitoring of threats to public health. 61 a major benefit of a less formal global collaboration is the development of networks of trust among knowledgeable scientists and clinicians, who are considered early warning posts for both natural and intentional outbreaks. 62 the one health initiative encourages collaboration between health professionals and is as important for bioterrorism preparedness as it is for management of emerging infectious disease and the global spread of antimicrobial resistance. 63 the management of patients that have been infected during incidents of bioterrorism can be challenging. 64 precautions and treatment regimens for several bioterrorist agents are summarised in panel 4. although supportive care serves as the basis of management for all agents, treatments for some of the relevant diseases have substantially progressed. the management of inhalation anthrax has advanced since the 2001 attack, 64, 70, 71 with improvements in critical care, and in treatment of acute lung injury and acute respiratory distress syndrome, severe sepsis, and septic shock. pleural effusions are routinely drained and there are more options for antimicrobial therapy. antibiotics are still recommended for 60 days after exposure or diagnosis, together with the anthrax vaccine. if the vaccine is given concurrently with antibiotic treatment, the period of treatment could be shortened. antibiotics for treatment of other bacterial infection are usually given for shorter periods, since the causative agents do not sequester spores. tularaemia is treated with ciprofloxacin or doxycycline. 72 for smallpox, antivirals such as cidofovir or a related acyclic nucleoside phosphonate analogue appear to be more effective than post-exposure vaccines in preventing mortality, according to experiments in non-human primates infected with monkeypox virus. 73 this suggests that antivirals might play an important role when preparing for a smallpox outbreak. for viral haemorrhagic fevers, ribavirin might have some efficacy in post-exposure prophylaxis. a small-molecule antiviral drug, gs-5734, has been developed that appears to be effective in treating ebola virus infection. 74 for intentional and other sudden outbreaks of contagious disease, isolation of patients and controlling risks to health-care workers remains extremely challenging, as was noted in the mers coronavirus, sars, ebola virus disease, and avian influenza epidemics. 75 hospital units adequately equipped for isolation are needed, similar to those equipped to care for filovirus-infected patients, with negative pressure air filtration. 76 if facilities are too small for the number of patients, a lower level of isolation with strict barrier nursing should be implemented, and in the event of a very large outbreak, there might be a need to set up isolation facilities in public buildings. in regions with poor infrastructure, it might be necessary to consider treating patients in their homes. people who have died should be regarded as infectious and handled with the same precautions used for patients. burial procedures might have to be modified, but every effort should be made to respect religious practices and traditions of the local culture. quarantining of people who might have been exposed to the infectious agent can be problematic, 77 as was the case in the sars and west african ebola virus epidemics. 78 since the quarantined population includes both people who were exposed and people who were not, the risk of disease transmission is higher. during the ebola virus outbreak in west africa, suspect cases were held until they could be cleared as negative, which took about 3 days pending the pcr results. on a national level, reducing the movement of populations is a sensitive issue, potentially interfering with commerce. 79 closure of schools is an important means of achieving social distancing to reduce spread. whether the public should use masks during an outbreak of a contagious disease is less clear, 80 and the efficacy of the masks is extremely variable. the most important reasons for variable efficacy are differences in facial shape, incorrect application, and duration of use. a substantial proportion of the cases and fatalities in the sars and ebola virus epidemics were among healthcare workers. clear guidelines specific to each agent are available to health-care personnel, public health workers, and emergency workers for the use of masks and personal protective equipment. the national ebola virus training and education center has been established in the usa to train health-care workers and assist hospitals in preparing for patients infected with high hazard virus in the usa and other countries. laboratory workers must be trained to work with dangerous pathogens and wear protective gear. designated threat pathogens must be stored, handled, and transported under a different set the role of vaccines in pre-exposure and post-exposure prophylaxis measures should be in place to protect the population from biological agents likely to be used in an attack before an incident occurs. however, since a bioterrorist incident is likely to be caused by biological agents not covered by routine immunisation, pre-exposure prophylaxis is generally confined to vaccines for military forces, health-care workers, and emergency response personnel. to the majority of the population, only postexposure prophylaxis is relevant. post-exposure prophylaxis for an intentional outbreak would include both those people known to be exposed during the incident and those people who were infected by others. the prophylaxis itself might consist of both vaccines and antimicrobials. when using live, attenuated vaccines, the relatively large proportion of the population who has some form of immunodeficiency has to be taken into account. 81, 82 monoclonal antibody preparations are now being considered for prophylaxis in select, high risk groups. 83 table 2 summarises pharmacological prophylaxis for tier 1 pathogens. currently, the vaccines that would most likely be used for pre-exposure or post-exposure prophylaxis are the smallpox and anthrax vaccines. since routine vaccination against smallpox was stopped in the 1980s, less than 50% of the world's population has been vaccinated, and antibody titres usually decline markedly after 5 to 10 years. 84 residual cellbased immunity can persist for many years. 85, 86 postexposure prophylaxis involves ring vaccination, which requires intensive tracing and vaccination of primary contacts, followed by vaccination of secondary contacts, 86 and finally, vaccination of all people in a defined affected region. for post-exposure prophylaxis of those directly exposed during the incident, vaccination can be effective if given within 3 to 4 days of exposure. 87 since it might take that amount of time to detect the first cases after exposure, the vaccine will generally only be effective for secondary and subsequent contacts. immune-boosting adjuvants 88 and toll-like receptor agonists have the potential to improve the immune response to post-exposure vaccination. 89 serious side-effects are relatively rare 90,91 but can affect com pliance. 92 in those cases, lower doses of vaccine might be administered, to provide adequate protection with fewer side-effects. 93 newer smallpox vaccines are in development, 94 including those that could immunise people who have atopic dermatitis. 95 because it is produced in small quantities by collecting antiserum from immunised humans, there might be a shortage of vaccinia immune globulin, used to treat people who would have serious side-effects with the vaccine. one possible solution to this shortage would be to use antibodies against other poxviruses, such as cowpox and monkeypox, 96 because of their cross-protective properties. since naturally occurring inhalation anthrax is extremely rare, there have been safety and immunogenicity profiles of the anthrax vaccine in humans, 97 but the efficacy has only been tested in animal models, and not in clinical trials. [98] [99] [100] the current anthrax vaccine is made from culture filtrates of a toxigenic, avirulent, nonencapsulated mutant of the bacillus anthracis vollum strain, and is administered in five intramuscular doses, followed by annual boosters. the protective, antigen specific memory b cells persist for many years after vaccination and are associated with humoral immunity. 99 serum igg response to the vaccine has been 100% after the fourth dose. 101 for post-exposure prophylaxis in unvaccinated people, the vaccine should be administered as a three-dose subcutaneous series (at 0, 2, and 4 weeks), in conjunction with a 60-day course of appropriate antimicrobial drugs. it has been given to thousands of us military personnel, and notable adverse events have been rare. 102 the anthrax vaccine is not recommended for pregnant women, although one study 103 with women in the us military, inadvertently vaccinated during pregnancy, did not show evidence of an increase in birth defects. in this study, 4418 women received the vaccine in the first trimester and 423 in the second and third trimeseter. more effective anthrax vaccines that require fewer doses are constantly being tested, 104, 105 such as the neat protein anthrax vaccine, 106 a dual purpose influenza vaccine that protects against anthrax, 107 and a combined anthrax-plague vaccine. 108 apart from the 17d yellow fever live attenuated vaccine and the junin virus vaccine, no vaccines for haemorrhagic fevers have been licensed. since the west african ebola virus epidemic, new ebola virus vaccines that have been long under development are being used successfully in the 2018 epidemic in the democratic republic of the congo. 109 essentially no other licensed vaccines are available for other tier 1 select agents. the previously licensed, formalin-inactivated, whole-bacilli plague vaccine has not proven effective against primary pneumonic plague in non-human primate models, 110 series us military. the vaccine has not been used widely for preexposure prophylaxis and has no place in post-exposure prophylaxis. new vaccines against tularaemia are under development 113, 114 including one that might provide crossprotection between plague and tularaemia. 115 the investigational pentavalent (abcde) botulinum toxoid vaccine was provided by the cdc for laboratory workers at high risk of exposure to botulinum toxin and it has also been given to military members that are at risk. the botulinum toxoid vaccine produces effective immunity after several months and has no value for postexposure prophylaxis because of the short latent period of the toxins. this vaccine was discontinued in 2012, 116 because of declines in immunogenicity and adverse events. new recombinant botulism vaccines are being developed, in addition to vaccines against glanders and rift valley fever. [117] [118] [119] efforts are ongoing to greatly shorten the time required to develop and produce new vaccines and other immune approaches. improved technologies are important for rapid scale-up and production of new treatment regimens, particularly following an attack with a contagious agent. the largely unpredictable nature of an epidemic initiated intentionally is likely to increase uncertainty and reduce public trust in the authorities. public education and effective risk communication are essential to increase public confidence and improve cooperation and compliance with recommended medical counter-measures. 120 the anthrax vaccine in military populations has caused considerable scepticism regarding the need for, safety, and efficacy of the vaccine. [121] [122] [123] clinicians and public health personnel should have access to up-to-date information, and the general public should be provided with nontechnical information and simple instructions on how to act during an emergency. sandman 120 has proposed that "one should not over-reassure, ack nowledge uncertainty, and share dilemmas". this behaviour would only cause overreaction or panic when new information about the risk is made public. risk communication will be necessary at all stages: before a bioterrorist incident occurs, when an incident is suspected, when it is confirmed, while it is taking place, and in the aftermath. credible and trusted spokespersons, including respected clinicians, scientists, and public servants for a country, should be adequately informed before an incident. during an outbreak, there could be unexpected events, such as atypical presentation of cases, varying responses to treatment (including unusual side-effects), and false positive and false negative diagnoses. the public might lose trust in the authorities if apparently unexposed people become ill. the advent and global distribution of social networking increases the risk of the dissemination of false or misleading information. lastly, a major infectious disease incident will also require flexibility and possible changes of established government policy. environmental detection of biological agents is another area of research that should be developed. to date, most systems of environmental detection have focused on anthrax, as a result of the anthrax attacks. 124, 125 however, a sensitive and specific set of recombinase polymerase amplification assays for fast screening, detection, and identification of b anthracis in a field setting has recently been developed. 126 the rare occurrence and likely small effect of an aerosol bioterrorist attack limits the practical use of environmental detection to special event venues, public transportation systems, and possibly some government buildings thought to be likely targets. many countries have national stockpiles of drugs and vaccines, for use in the event of a biological or chemical attack, or for serious outbreaks that might achieve epidemic proportions. the usa, for instance, maintains a strategic national stockpile of vaccines and other medical countermeasures. global stores of smallpox vaccines are held by who, in addition to stores held by individual countries. some countries have undertaken active vaccination programmes against smallpox and anthrax in the military and first responder populations. preparedness for bioterrorist incidents requires constant re-evaluation of policies. 127 although there is no evidence that the 2009 h1n1 influenza pandemic or the 2014-16 ebola virus epidemic in west africa were initiated intentionally, the local and international responses revealed strengths and weaknesses in the current state of preparedness for bioterrorist incidents. 128 the ebola virus epidemic spread to a number of countries, with more than 20 000 cases reported worldwide and a case fatality rate of more than 60%. imported cases of ebola virus disease were identified in the usa and spain. locally, in the affected countries in west africa, 10% of the people who died because of ebola virus disease were health-care workers. various shortcomings in the response to the epidemic have been identified since (panel 1). accurately predicting the intentional misuse of a biological agent to cause harm is difficult without intelligence data, but several attempts have been made to rationally predict the categories of risk: man-made, natural, accidental, contagious, and non-contagious. 129 series risks. 132 the risk of bioterrorism has called into question some of the dogmas related to eradication of diseases such as poliomyelitis and measles. 133 for example, if polio is successfully eradicated, universal vaccination might have to continue because of the risk of poliovirus being used as a bioterrorist agent. [134] [135] [136] emerging and reemerging infectious diseases will continue to be a threat, 137 but preparedness for bioterrorism is, in many ways, similar to preparedness for naturally emerging disease. all countries should collaborate to address the root causes of terrorism, and develop appropriate preventive strategies. effective preparedness is, in itself, a deterrent to bioterrorism, since it reduces the incentive to use biological weapons by making a country or region a hard target. it is also the cornerstone of consistent and effective responses to naturally occurring epidemics. the abuse of biological agents can be further reduced or discouraged with reliable intelligence and an effective response if it does occur. national and regional resources and capabilities will vary, but all will require infrastructures that are capable of recognising and dealing with a variety of biological agents. 138 the needs of specific populations, such as the paediatric population, pregnant women, 139 elderly people, and people with immunological disorders, must also be addressed. 140 funding for biodefence is crucial to adequate preparation and response to bioterrorist threats. 141 international preparedness for bioterrorism has the dual benefit of strengthening the infrastructure for responding to naturally occurring epidemics of highly pathogenic organisms. lessons from the 2014 west african ebola virus epidemic show that health-care providers must always be watchful for unusual presentations of disease, and new and improved approaches must be developed for early detection and response. health-care providers need more effective means of isolating infected patients, and better methods to control the movement of potentially infected people outside of the affected areas. personal protective equipment should be inexpensive and effective, and available to use with minimal training and under harsh environments. protection of health-care workers against infection remains particularly prob lematic, and should be a focus of research and development. the ebola virus epidemic has highlighted the importance of improving the logistics of moving human and material resources in areas with relatively poor infrastructure. risk communication and public education before and during an outbreak need to be improved. more clinical trials should be fast-tracked during development of new vaccines and antiviral drugs. preparedness for a low-risk, high-impact event that is bioterrorism should be monitored constantly, tested in tabletop exercises, 142, 143 and integrated into the routine functioning of the health system. here it would serve the dual purpose of ensuring that countries are prepared to meet the challenges of controlling epidemics of emerging and re-emerging infectious diseases. msg designed the review, did the literature search and was responsible for writing the manuscript. jld assisted in the literature search, revised the manuscript, and supplied technical expertise. dc assisted in the literature search and revised the manuscript. drf helped design the review, contributed source material, did the literature search, and helped write and revise the manuscript. we declare no competing interests. we searched pubmed and google scholar using the terms "bioterrorism", "sustainable bioterrorism preparedness", "all-hazards infectious disease preparedness", "biological threat agents", "bioterrorism preparedness", "emerging infectious diseases", "smallpox", "anthrax", "plague", "tularemia", "botulism", "hemorrhagic fevers", and "risk communication". we searched national and international reports from the who and us centers for disease control using the terms "bioterrorism", "bioterrorism preparedness", and "emerging infectious diseases". we also completed web searches for "disease surveillance", "infectious disease diagnostics", "medical countermeasures", "emergency healthcare delivery", and "risk management". we focused on academic literature in english and restricted most of our searches to documents published since 1990, with an emphasis on those published after 2001, and included mainly those published since 2008. federal funding for health security in fy2016 a large community outbreak of salmonellosis caused by intentional contamination of restaurant salad bars investigation of bioterrorism-related anthrax, united states, 2001: epidemiologic findings remediation of bacillus anthracis contamination in the u.s. department of justice mail facility total decontamination cost of the anthrax letter attacks bacillus anthracis incident post-traumatic stress disorder symptoms in victims of tokyo subway attack: a 5-year follow-up study biological warfare and bioterrorism: a historical review the threat of bioterrorism: identifying the unknown. ffi focus biodefense in the 21st century department of health and human services (hhs). possession, use, and transfer of select agents and toxins-addition of bacillus cereus biovar anthracis to the hhs list of select agents and toxins. interim final rule and request for comments destruction of microbial collections in response to select agent and toxin list regulations implementing the select agent legislation: perfect record or wrong metric? gaps remain in china's ability to detect emerging infectious diseases despite advances since the onset of sars and avian flu health system resource gaps and associated mortality from pandemic influenza across six asian territories ebola: lessons learned and future challenges for europe public health assessment of potential biological terrorism agents responsible conduct by life scientists in an age of terrorism mitigating the risks of synthetic biology. council on foreign relations characterization of the reconstructed 1918 spanish influenza pandemic virus chemical synthesis of poliovirus cdna: generation of infectious virus in the absence of natural template expression of mouse interleukin-4 by a recombinant ectromelia virus suppresses cytolytic lymphocyte responses and overcomes genetic resistance to mousepox labmade smallpox is possible, study shows the de novo synthesis of horsepox virus: implications for biosecurity and recommendations for preventing the reemergence of smallpox epidemiologic determinants for modeling pneumonic plague outbreaks risk assessment and risk communication strategies in bioterrorism preparedness. nato security through science series a: chemistry and biology estimating time and size of bioterror attack threat of a biological terrorist attack on the us food supply: the cdc perspective a large food-borne outbreak of group a streptoccocal pharyngitis in an industrial plant: potential for deliberate contamination german outbreak of escherichia coli o104:h4 associated with sprouts water and bioterrorism: preparing for the potential threat to u.s. water supplies and public health a proteomics assay to detect eight cbrn-relevant toxins in food index case of fatal inhalational anthrax due to bioterrorism in the united states rapid outbreak sequencing of ebola virus in sierra leone identifies transmission chains linked to sporadic cases real-time pcr to identify variola virus or other human pathogenic orthopox viruses pocket dna sequencers make real-time diagnostics a reality sensitive detection of francisella tularensis directly from whole blood by use of the genexpert system comparison of magpix assays and enzyme-linked immunosorbent assay for detection of hemorrhagic fever viruses the early humoral immune response to bacillus anthracis toxins in patients infected with cutaneous anthrax specific, sensitive, and quantitative enzyme-linked immunosorbent assay for human immunoglobulin g antibodies to anthrax toxin protective antigen rapid detection of bacillus anthracis bloodstream infections by use of a novel assay in the genexpert system rapid and sensitive point-of-care detection of orthopoxviruses by abicap immunofiltration rapid viral diagnosis of orthopoxviruses by electron microscopy: optional or a must? striking against bioterrorism with advanced proteomics and reference methods simultaneous immunodetection of anthrax, plague, and tularemia from blood cultures by use of multiplexed suspension arrays the changing face of pathogen discovery and surveillance public health surveillance and infectious disease detection evaluation of a syndromic surveillance system using the wsare algorithm for early detection of an unusual, localized summer outbreak of influenza b: implications for bioterrorism surveillance syndromic surveillance during pandemic (h1n1) 2009 outbreak internet-based surveillance systems for monitoring emerging infectious diseases digital surveillance for enhanced detection and response to outbreaks if syndromic surveillance is the answer, what is the question? clinical recognition and management of patients exposed to biological warfare agents public health. ethics and the conduct of public health surveillance digital disease detectionharnessing the web for public health surveillance factors influencing performance of internet-based biosurveillance systems used in epidemic intelligence for early detection of infectious diseases outbreaks the internet and the global monitoring of emerging diseases: lessons from the first 10 years of promed-mail social and news media enable estimation of epidemiological patterns early in the 2010 haitian cholera outbreak geneva: world health organization the global public health intelligence network and early warning outbreak detection: a canadian contribution to global public health with the changing biological threat…smart international engagement policy would lower cost and increase national security animals as sentinels of bioterrorism agents clinical management of potential bioterrorism-related conditions wisconsin department of health services. infection control and prevention-standard precautions oral tecovirimat for the treatment of smallpox plague: recognition, treatment, and prevention the sanford guide to antimicrobial therapy new therapeutic approaches for treatment of tularaemia: a review centers for disease control and prevention expert panel meetings on prevention and treatment of anthrax in adults antitoxin treatment of inhalation anthrax: a systematic review consequences of delayed ciprofloxacin and doxycycline treatment regimens against francisella tularensis airway infection a single cidofovir treatment rescues animals at progressive stages of lethal orthopoxvirus disease therapeutic efficacy of the small molecule gs-5734 against ebola virus in rhesus monkeys risks to healthcare workers with emerging diseases: lessons from mers-cov, ebola, sars, and avian flu implementing a negative-pressure isolation ward for a surge in airborne infectious patients large-scale quarantine following biological terrorism in the united states: scientific examination, logistic and legal limits, and possible consequences ebola and fda: reviewing the response to the 2014 outbreak, to find lessons for the future assessment of the potential for international dissemination of ebola virus via commercial air travel during the 2014 west african outbreak transmission of ebola viruses: what we know and what we do not know estimating the size of the u.s. population at risk of severe adverse events from replicating smallpox vaccine influence of population immunosuppression and past vaccination on smallpox reemergence next-generation monoclonal antibodies: challenges and opportunities analysis of historical data suggests long-lasting protective effects of smallpox vaccination duration of neutralizing antibody persisting in thai individuals after childhood vaccination against smallpox a model for a smallpox-vaccination policy can postexposure vaccination against smallpox succeed? immune-boosting adjuvants tlr3 and tlr9 agonists improve postexposure vaccination efficacy of live smallpox vaccines risks of serious complications and death from smallpox vaccination: a systematic review of the united states experience adverse events following smallpox vaccination with acam2000 in a military population factors associated with healthcare worker acceptance of vaccination: a systematic review and meta-analysis reducing the dose of smallpox vaccine reduces vaccine-associated morbidity without reducing vaccination success rates or immune responses comparing new-generation candidate vaccines against human orthopoxvirus infections long-term safety of replication-defective smallpox vaccine (mva-bn) in atopic eczema and allergic rhinitis cross-neutralizing and protective human antibody specificities to poxvirus infections randomized, double-blind, placebo-controlled, safety and immunogenicity study of 4 formulations of anthrax vaccine adsorbed plus cpg 7909 (av7909) in healthy adult volunteers select human anthrax protective antigen epitope-specific antibodies provide protection from lethal toxin challenge protective antigen-specific memory b cells persist years after anthrax vaccination and correlate with humoral immunity lethal factor antibodies contribute to lethal toxin neutralization in recipients of anthrax vaccine precipitated serum igg antibody response to the protective antigen (pa) of bacillus anthracis induced by anthrax vaccine adsorbed (ava) among u.s. military personnel an overview of adverse events reported by participants in cdc's anthrax vaccine and antimicrobial availability program safety of inadvertent anthrax vaccination during pregnancy: an analysis of birth defects in the u.s. military population progress and novel strategies in vaccine development and treatment of anthrax host immunity to bacillus anthracis lethal factor and other immunogens: implications for vaccine design progress toward the development of a neat protein vaccine for anthrax disease a dual purpose universal influenza vaccine candidate confers protective immunity against anthrax a bivalent anthrax-plague vaccine that can protect against two tier-1 bioterror pathogens, bacillus anthracis and yersinia pestis experimental treatment of ebola virus disease with tkm-130803: a single-arm phase 2 clinical trial plague vaccines: current developments and future perspectives plague vaccines: status and future intranasal delivery of a protein subunit vaccine using a tobacco mosaic virus platform protects against pneumonic plague protective immunity against lethal f. tularensis holarctica lvs provided by vaccination with selected novel cd8+ t cell epitopes francisella tularensis live vaccine strain deficient in capb and overexpressing the fusion protein of igla, iglb, and iglc from the bfr promoter induces improved protection against f. tularensis respiratory challenge yopp-expressing variant of y. pestis activates a potent innate immune response affording cross-protection against yersiniosis and tularemia notice of cdc's discontinuation of investigational pentavalent (abcde) botulinum toxoid vaccine for workers at risk for occupational exposure to botulinum toxins recombinant botulinum neurotoxin hc subunit (bont hc) and catalytically inactive clostridium botulinum holoproteins (cibont hps) as vaccine candidates for the prevention of botulism burkholderia pseudomallei and burkholderia mallei vaccines: are we close to clinical trials? safety and immunogenicity of a mutagenized, live attenuated rift valley fever vaccine, mp-12, in a phase 1 dose escalation and route comparison study in humans bioterrorism risk communication policy why do uk military personnel refuse the anthrax vaccination? a longitudinal study of uk military personnel offered anthrax vaccination: informed choice, symptom reporting, uptake and pre-vaccination health notes from the field: compliance with postexposure prophylaxis for exposure to bacillus anthracis among u.s. military personnel-south korea rapid detection of bacillus anthracis spores using immunomagnetic separation and amperometry rapid detection of viable bacillus anthracis spores in environmental samples by using engineered reporter phages sensitive and specific recombinase polymerase amplification set of assays for fast screening, detection and identification of bacillus anthracis in a field setting the pitfalls of bioterrorism preparedness: the anthrax and smallpox experiences pandemic preparedness and response-lessons from the h1n1 influenza of 2009 global catastrophic biological risks: toward a working definition expert views on biological threat characterization for the u.s. government: a delphi study the risk of bioterrorism re-analysed risks of paralytic disease due to wild or vaccine-derived poliovirus after eradication eradicating polio: a balancing act vaccination against polio should not be stopped vaccines should be kept even if polio is wiped out ebola and zika: cautionary tales medical countermeasure development since 2001: a long way yet to go preparing for biological threats: addressing the needs of pregnant women safety of live vaccinations on immunosuppressive therapy in patients with immune-mediated inflammatory diseases, solid organ transplantation or after bone-marrow transplantation-a systematic review of randomized trials, observational studies and case reports biosecurity: assessing the bioweapons threat a plague on your city: observations from topoff shining light on "dark winter key: cord-022003-cvawdes6 authors: darling, robert g.; noste, erin e. title: future biological and chemical weapons date: 2015-10-23 journal: ciottone's disaster medicine doi: 10.1016/b978-0-323-28665-7.00080-7 sha: doc_id: 22003 cord_uid: cvawdes6 nan robert g. darling and erin e. noste biological and chemical weapons have been used throughout history. 1 for millennia, indigenous south american peoples deliberately used plant-derived arrow poisons such as curare and toxins from poison dart frogs, although these preparations were used mainly for hunting. similar toxins were used in africa. the ancient greeks, for whom toxikon meant "arrow poison," tipped arrows with winter aconite, and this practice continued into medieval europe, persisting into the seventeenth century in spain and portugal. 2 soldiers in india used smoke screens, incendiary weapons, and toxic fumes as early as 2000 bce, and the sung dynasty in china employed a wide variety of arsenical smokes and other poisons in battle. the military use of toxins dates from at least the sixth century bce, when assyrian soldiers poisoned enemy wells with ergotcontaminated rye. in 423 bce, during the peloponnesian war, thracian allies of sparta captured the athenian fort at delium by using a long tube and bellows to blow poisonous smoke from coals, sulfur, and pitch into the fort. greek fire (likely composed of rosin, sulfur, pitch, naphtha, lime, and saltpeter) was invented in the seventh century ce and proved to be a very effective naval weapon. various poisons saw battlefield use during medieval times, and the use of poisons for murder (including assassinations) became widespread. other examples before the twentieth century include the contamination of water by dumping the corpses of dead humans or animals into wells, the use of snakes and other creatures as poisonous vectors, and occasionally, fomites to transmit infections such as smallpox to unsuspecting victims. this latter technique was used with remarkable success during the french and indian war (1754-1767), when sir jeffrey amherst was alleged to have given "gifts" (blankets) harboring the pus and scabs from smallpox victims to unsuspecting native americans. the indians possessed no immunity against smallpox and thus experienced very high rates of infection and mortality as smallpox swept through the local tribes. 3 during the late nineteenth and early twentieth centuries, the science and technology necessary for the development of sophisticated biological and chemical weapons proceeded apace. world war i saw the first large-scale use of "poison gas," including lacrimators, chlorine, phosgene, arsenicals, cyanide, and sulfur mustard. by the end of the war, nearly one in every three rounds was a chemical munition. dr. shiro ishii and other japanese scientists in the infamous unit 731 worked on the weaponization of anthrax, plague, smallpox, and tetrodotoxin as well as a variety of chemical agents during world war ii. there are even suspicions that the bomb used in the assassination of reinhard heydrich in czechoslovakia in 1942 contained botulinum toxin. 4 after world war ii, ricin was used as an injectable assassination weapon, and in the 1970s and 1980s t-2 toxin, a trichothecene mycotoxin, was alleged to have been the toxic component of the "yellow rain" employed against h'mong refugees from laos. more recently, iraq and iran both used chemical weapons against each other in the iran-iraq war of the 1980s, and iraq had a weapons program that included the development of sulfur mustard, nerve agents, "agent 15" (an anticholinergic incapacitating agent), botulinum toxin, epsilon toxin from clostridium perfringens, and aflatoxin. 5 militia groups in the united states and terrorist groups throughout the world have used ricin for political purposes. american scientists started developing chemical weapons as a response to the use of chemical warfare in europe during world war i and conducted both offensive and defensive research on biological and chemical weapons. however, in 1969 the united states unilaterally renounced the first use of chemical agents, halted chemical-agent production, and terminated its offensive biological weapons program. in 1972 the biological weapons and toxins convention was created; it was signed by representatives from 104 nations, including the united states (which ratified the convention in 1975), the soviet union, and iraq, although many signatories did not consider toxins to be biological weapons and did not consider the treaty binding on toxin use. since that time, at least 140 nations have either signed or ratified this treaty. 6 however, the soviet union and iraq began violating the treaty in short order. in the soviet union, weapons scientists stepped up research and development of numerous biological and chemical weapons as part of one of the largest and most comprehensive biological-weapons programs in history. soviet scientists created large stockpiles of weaponized anthrax, plague, smallpox, tularemia, nerve agent, mustard, and other biological and chemical agents. 5 in 1979 the world was put on notice of the devastating potential that biological weapons pose to humanity. in that year, a small quantity of weapons-grade anthrax was accidentally released from a manufacturing plant located in the former city of sverdlovsk (now yekaterinburg) in russia, resulting in 77 cases and 66 deaths. dr. matthew meselson, a harvard scientist, was permitted to study the event many years later and reported the results of his work in a 1979 science article. meselson determined that the majority of the deaths had occurred among victims living in a narrow, 4-km-wide band downwind from the plant. animal deaths were confirmed as far as 30 km downwind. meselson further concluded that less than 1 g of weapons-grade anthrax had been released from the plant. 7 if his calculations are accurate, weaponized anthrax possesses staggering potential as a biological weapon given its stability, its relative ease of production, and its ability to be dispersed in a clandestine manner over great distances. in march 1995, after having unsuccessfully attempting to deploy biological agents, members of the aum shinri kyo cult executed a coordinated attack with the nerve agent sarin (gb) on the tokyo subway system. more than 5500 people sought medical treatment, and a dozen died. the aum shinri kyo had used sarin in matsumoto 9 months earlier in an attack that had exposed more than 300 people and killed 7 in an attempt to assassinate judges unfavorable to their cause. 8, 9 the anthrax attacks in the fall of 2001 involved the use of letters containing weapons-grade anthrax mailed through the u.s. postal system. five people died, and 17 became ill with either cutaneous or inhalational anthrax. buildings contaminated with spores included the hart senate office building and the brentwood postal facilities in washington, dc. it cost millions of dollars to rehabilitate these buildings. the anthrax used in the attacks was determined to be extremely potent and could have caused far greater numbers of casualties had it been dispersed more widely. 10, 11 the use of chemical weapons also occurred in recent history. in september 2013 the united nations (un) released their investigations on the use of chemical weapons in syria. the un concluded that sarin gas was used on august 21, 2013, in the ghouta area of damascus against "civilians, including children, on a relatively large scale." 12 these findings were based on interviews with survivors and other witnesses, documentation of munitions and their components, collection of environmental samples for subsequent analysis, assessment of symptoms of survivors, and collection of hair, urine, and blood samples. according to dr. ken alibek, former deputy director of biopreparat, the soviet union's nominally civilian medical research institute, soviet scientists and physicians spent large sums of money and manpower during the 1980s and 1990s developing the most lethal and potent biological weapons known to man. in addition to weaponizing the etiologic agents of anthrax, smallpox, marburg fever, and others, they created antibiotic-resistant strains of yersinia pestis (plague), francisella tularensis, and other pathogens. furthermore, by applying genetic engineering techniques, the soviets are also alleged to have created pathogens with novel characteristics and strains of several organisms capable of defeating certain vaccines. 13 as we enter the biotechnological revolution of the twenty-first century, our understanding of molecular biology, genetics, and biochemistry is exploding. the human genome has been sequenced, and it is now possible to manipulate genes from disparate organisms to create new and novel pathogens. scientists are also able to synthesize and weaponize a number of different endogenous biological-response modifiers including cytokines, hormones, neurotransmitters, and plasma proteases. but even nature continues to surprise us. new, naturally occurring infections with the potential to cause large-scale human diseases and death continue to emerge at an ever-increasing rate throughout the world, and it is conceivable that these pathogens could also be weaponized by enterprising scientists. this chapter briefly reviews the future of chemical and biological weapons as we enter this new era of explosive growth in our understanding of the life sciences. we are presented with an extraordinary opportunity to solve a host of human afflictions or to create new classes of biological and chemical weapons that have the capacity to destroy our civilization as we know it today. the appearance of a new or reemerging infectious disease has global implications. during the past 20 years, more than 30 new lethal pathogens have been identified. 14 a classic example of this emerging threat is pandemic influenza. in 1918, as world war i was coming to an end, the spanish flu struck with devastating consequences. in less than 1 year, this virus was able to circumnavigate the globe and kill an estimated 40 million people. 15 more recently, the emergence of severe acute respiratory syndrome (sars) in southeast asia resulted from a coronavirus that jumped species from animals to humans and rapidly spread to 29 countries in less than 90 days. finally, the 2014 outbreak of ebola in western africa, still raging as of this writing, is an example of how devastating these agents can be when they emerge in a region previously naïve to them. novel and dormant infectious agents such as sars, influenza, and ebola appear to be emerging or reemerging with increasing frequency and with greater potential for serious consequences. many factors contribute to the emergence of new diseases: environmental changes, global travel and trade, social upheaval, and genetic changes in infectious agent, host, or vector populations. once a new disease is introduced into a suitable human population, it often spreads rapidly and has a devastating impact on the medical and public health infrastructure. if the disease is severe, it may lead to social disruption and have a profound economic impact. outbreaks of emerging or reemerging diseases may be difficult to distinguish from outbreaks resulting from intentional introduction of infectious diseases for nefarious purposes. as scientists develop more sophisticated laboratory procedures and increase their understanding of molecular biology and the genetic code, the possibility of bioengineering more virulent, antibiotic, and vaccineresistant pathogens for military or terrorist uses becomes increasingly likely. it is already theoretically possible to synthesize and weaponize certain biological response modifiers (brms) as well as to engineer genomic weapons capable of inserting novel dna into host cells. the potential to cause widespread disease and death with any of these weapons is incalculable and concerning. scientists and policy makers have begun to address the issue with a robust research agenda to develop medical countermeasures. ebola hemorrhagic fever, as of december 4, 2014, had caused 6055 deaths among 17,111 confirmed cases in western africa. scientists have debated whether ebola could be weaponized into a weapon of mass destruction by terrorists. the consensus seems to be that this would be a very difficult undertaking because of the knowledge and laboratory skills that are required. moreover, the biology of the virus does not lend itself well to weaponization. 16 of course, nefarious individuals could use ebola in a number of ways, in much the same way that a suicide bomber straps on a vest laden with explosives. in theory a person could deliberately infect themselves with the virus and then attempt to infect others once they become symptomatic. important existing biological agents with the potential for weaponization for military or terrorist use include the following: another way to view the relative importance of the above list of agents and diseases list is to consider the centers for disease control and prevention (cdc) strategy. the cdc categorizes bioterrorism agents or diseases as category a, b, or c. 17 category a agents pose the highest risk to the public and are characterized as follows: • easily disseminated or transmitted from person to person • can cause high mortality rates and possess the potential for profound public impact • could cause public panic and social disruption • require special preparations for adequate public health preparedness diseases and agents • anthrax (bacillus anthracis) • botulism (clostridium botulinum toxin) • plague (yersinia pestis) • smallpox (variola major) • tularemia (francisella tularensis) • viral hemorrhagic fever: filoviruses (e.g., ebola and marburg) and arenaviruses (e.g., lassa and machupo) category b agents are the next highest priority and are characterized by • μoderately easy to disseminate • cause moderate morbidity and low mortality • require specific enhancements of cdc's diagnostic capacity and enhanced disease surveillance diseases and agents • epsilon toxin of clostridium perfringens • food safety threats (salmonella species, escherichia coli o157:h7, • ricin toxin (ricinus communis-castor beans) • staphylococcal enterotoxin b • typhus fever (rickettsia prowazekii) • viral encephalitis (alphaviruses: venezuelan equine encephalitis, eastern equine encephalitis, western equine encephalitis) • water safety threats (vibrio cholerae, cryptosporidium parvum) category c agents form the third highest priority and include emerging pathogens that could be engineered for mass dissemination in the future because of the following: • availability • ease of production and dissemination • potential for high morbidity and mortality rates and major health impact agents • emerging and reemerging infectious diseases such as nipah virus, hantavirus, human influenza, avian influenza, sars and sarsassociated coronavirus (sars-cov), and middle east respiratory syndrome (mers) because emerging diseases are so diverse and endemic to different geographic locations, their complete description is beyond the scope of this chapter. however, some of these infections may become future threats as agents of biological warfare or terrorism. the most worrisome emerging infectious disease may well be the one we do not know about. recent experience with hiv, ebola hemorrhagic fever, sars, monkey pox, west nile fever, and hundreds of other "new" diseases reveal that we will continue to be surprised. avian influenza, or highly pathogenic avian influenza, has periodically caused human infections primarily through close contact with avian species, most often through occupational contact at chicken or duck farms in southeast asia. a large outbreak of avian influenza involving the h5n1 strain and human cases occurred in 2004 and originated in two countries from this region. 18 no sustained human-to-human transmission was reported, but there is some evidence that isolated episodes did occur, and the potential exists for genetic reassortment between avian and human or animal strains of influenza. a recent report in the journal science linked the influenza virus responsible for the 1918 epidemic to a possible avian origin. 19 if true, avian influenza may pose a much greater danger to human populations than previously reported. the disease presents in humans in a fashion similar to other types of influenza viruses. it usually begins with fever, chills, headaches, and myalgias and often involves the upper and lower respiratory tract with development of cough, dyspnea, and, in severe cases, acute respiratory distress syndrome. laboratory findings may include pancytopenia, lymphopenia, elevated liver enzymes, hypoxia, a positive reverse transcriptase-polymerase chain reaction (rt-pcr) test for h5n1, and a positive neutralization assay for h5n1 influenza strain. in vitro studies suggest that the neuraminidase (na)-inhibitor class of drugs may have clinical efficacy in the treatment and prevention of avian influenza infection. 20 the threat for pandemic spread of human influenza viruses is substantial. the pathogenicity of human influenza viruses is directly related to their ability to alter their eight viral rna segments rapidly; the new antigenic variation results in the formation of new hemagglutinin (ha) and na surface glycoproteins, which may go unrecognized by an immune system primed against heterologous strains. two distinct phenomena contribute to a renewed susceptibility to influenza infection among persons who have had influenza illness in the past. clinically significant variants of influenza a viruses may result from mutations occurring in the ha and na genes and expressed as 491 chapter 80 future biological and chemical weapons minor structural changes in viral surface proteins. as few as four amino acid substitutions in any two antigenic sites can cause such a clinically significant variation. these minor changes result in an altered virus able to circumvent host immunity. moreover, genetic reassortment between avian and human or avian and porcine influenza viruses may lead to the major changes in ha or na surface proteins known as antigenic shift. in contrast to the gradual evolution of strains subject to antigenic drift, antigenic shift occurs when an influenza virus with a completely novel ha or na formation moves into humans from other host species. global pandemics result from such antigenic shifts. influenza causes in excess of 30,000 deaths and more than 100,000 hospitalizations annually in the united states. pandemic influenza viruses have emerged regularly in 10-to 50-year cycles for the last several centuries. during the last century, influenza pandemics occurred 3 times: in 1918 ("spanish influenza," an h1n1 virus), in 1957 (asian influenza, an h2n2 subtype strain), and in 1968 (hong kong influenza, an h3n2 variant). the 1957-1958 pandemic caused 66,000 excess deaths, and the 1968 pandemic caused 34,000 excess deaths in the united states. the 1918 influenza pandemic illustrates a worst-case public health scenario; it caused 675,000 deaths in the united states and 20 to 40 million deaths worldwide. 19 morbidity in most communities was between 25% and 40%, and the case-mortality rate averaged 2.5%. a reemergent 1918-like influenza virus would have tremendous societal effects, even in the event that antiviral medications were effective against this more lethal influenza virus. sars-associated coronavirus (sars-cov) emerged as the cause of sars during 2003. that year, sars was responsible for approximately 900 deaths and more than 8000 infections in people from at least 29 countries worldwide. before a case definition had been clearly established, chinese authorities reported to the world health organization (who) more than 300 cases of an atypical pneumonia with 5 related deaths, all originated from guangdong province in china during february 2003. the infection quickly spread as infected patients traveled to hong kong and from there to vietnam, canada, and other locations. only eight laboratory-confirmed cases occurred in the united states, but there is concern that the u.s. population is vulnerable to a widespread outbreak of sars such as the one that occurred in china, hong kong, singapore, toronto, and taiwan in 2003. 21 a sars case definition evolved from this initial report to the who by chinese health authorities in february 2003. a case was initially defined by clinical criteria; a suspected or probable case was defined as an illness that included potential exposure to an existing case and fever with pneumonia or respiratory distress syndrome. in april 2003, a confirmed case was defined as a case from which sars-cov was isolated from culture. 22 sars-cov infections have an incubation period of 2 to 10 days. systemic symptoms such as fever and chills followed by a dry cough and shortness of breath begin within 2 to 7 days. patients may develop pneumonia and lymphopenia by days 7 to 10 of the illness. most patients with sars-cov have a clear history of exposure either to a patient with sars or to a setting in which sars-cov is known to exist. laboratory tests may be helpful but do not reliably detect infection early during the illness. sars-cov should be suspected in patients requiring hospitalization for radiographically confirmed pneumonia or acute respiratory distress syndrome of unknown etiology and one of the following risk factors during the 10 days before the onset of illness: (1) travel to china, hong kong, or taiwan, or close contact with an ill person having a history of such travel; (2) employment in an occupation associated with a risk for sars-cov exposure; or (3) inclusion in a cluster of cases of atypical pneumonia without an alternative diagnosis. a "respiratory hygiene/cough etiquette" strategy should be adopted in all sars-affected health care facilities. all patients admitted to the hospital with suspected pneumonia should receive the following measures: (1) they should be placed in droplet isolation until it is determined that isolation is no longer indicated (standard precautions are appropriate for most community-acquired pneumonias; droplet precautions for nonavian influenza); (2) they should be screened for risk factors of possible exposure to sars-cov; and (3) they should be evaluated with a chest radiograph, pulse oximetry, complete blood count, and additional workup as indicated. if the patient has a risk factor for sars, droplet precautions should be implemented pending an etiologic diagnosis. when there is a high index of suspicion for sars-cov disease, the patient should be treated in terms of sars isolation precautions immediately (including airborne precautions), and all contacts of the ill patient should be identified, evaluated, and monitored. 22 although ribavirin, high-dose corticosteroids, and interferons have been used in treatment, it is unclear what effect they have had on clinical outcome. no definitive therapy has been established. empiric antibiotic treatment for community-acquired pneumonia following the current american thoracic society/infectious diseases society of america guidelines is recommended pending etiologic diagnosis. diagnostic tests for sars-cov include antibody testing using an enzyme immunoassay and rt-pcr tests for respiratory, blood, and stool specimens. 23 in the absence of known sars-cov transmission, testing is recommended only in consultation with public health authorities. testing for influenza, respiratory syncytial virus, pneumococcus, chlamydia, mycoplasma, and legionella should be conducted, as the identification of one of these agents excludes sars by case definition. clinical samples can be obtained during the first week of illness with a nasopharyngeal swab plus an oropharyngeal swab and a serum or a plasma specimen. after the first week of illness, a nasopharyngeal swab plus an oropharyngeal swab and a stool specimen should be obtained. serum specimens for sars-cov antibody testing should be collected when the diagnosis is first suspected and at later times as indicated. an antibody response can occasionally be detected during the first week of illness, is likely to be detected by the end of the second week of illness, and at times may not be detected until more than 28 days after the onset of symptoms. respiratory specimens from any of several different sources may be collected for viral and bacterial diagnostics, but the preferred specimens of choice are nasopharyngeal washes or aspirates. 23 mers-cov is a disease caused by a coronavirus and results in severe acute respiratory disease including fever, chills, cough, and dyspnea. some patients also develop nausea, vomiting, and diarrhea. it has a 30% mortality rate. most patients who have died had underlying comorbidities and developed pneumonia or renal failure. the illness was first reported in saudi arabia in september 2012; most cases appear to be limited to the arabian peninsula. the incubation period ranges from 2 to14 days. the illness can spread from person to person but it requires close contact, and no sustained transmission had been reported by late 2014. the virus could evolve and lead to sustained transmission, but this cannot be predicted with certainty. there is no vaccine for mers-cov and there is no specific treatment, but there is ongoing research by the national institutes of health (nih) and other entities to fill this gap. 24 the nipah and hendra viruses are closely related but distinct paramyxoviruses that compose a new genus within the family paramyxoviridae. the nipah virus was discovered in malaysia in 1999 during an outbreak of a zoonotic infection, now called nipah virus encephalitis, involving mostly pigs and some human cases. 25 hendra, the causative 492 section vii topics unique to terrorist events and high-threat disaster response agent of hendra virus disease, was identified in a similar outbreak involving a single infected horse and three human cases in southern australia in 1994. 26 it is believed that certain species of fruit bats are the natural hosts for these viruses and remain asymptomatic. horses and pigs act as amplifying hosts for the hendra and nipah viruses, respectively. the mode of transmission from animal to humans appears to require direct contact with tissues or body fluids or with aerosols generated during butchering or culling. personal protective equipment including gowns, gloves, and respiratory and eye protection is advised for agricultural workers culling infected animal herds. thus far, human-to-human transmission of these viruses has not been reported. in symptomatic cases, the onset of disease begins with flu-like symptoms and rapidly progresses to encephalitis with disorientation, delirium, and coma. fifty percent of those with clinically apparent infections have died from their disease. there is currently no approved treatment for these infections, and, therefore, therapy relies heavily on supportive care. the antiviral drug ribavirin has been used in past infections, but its effectiveness remains unproven in clinically controlled studies. 27 although no person-to-person transmission is known to have occurred, barrier nursing and droplet precautions are recommended because respiratory secretions and other bodily fluids are known to harbor the virus. the clinical laboratory should be notified before specimens are sent as these may pose a laboratory hazard. specimens for viral isolation and identification should be forwarded to a reference laboratory. requests for testing should come through public health departments, which should contact the cdc emergency operations center at 770-488-7100 before sending specimens. brms direct the myriad complex interactions of the immune system. brms include erythropoietins, interferons, interleukins, colonystimulating factors, granulocyte and macrophage colony-stimulating factors, stem cell growth factors, monoclonal antibodies, tumor necrosis factor inhibitors, and vaccines. 28 a growing understanding of the structure and function of brms is driving the discovery and creation of many novel compounds including synthetic analgesics, antioxidants, and antiviral and antibacterial substances. for example, brms are being used to treat debilitating rheumatoid arthritis by targeting cytokines that contribute to the disease process. 29 by neutralizing or eliminating these targeted cytokines, brms may reduce symptoms and decrease inflammation. brms may also be used as anticarcinogens, with the following goals: (1) to stop, control, or suppress processes that permit cancer growth; (2) to make cancer cells more recognizable, and therefore more susceptible, to destruction by the immune system; (3) to boost the killing power of immune system cells, such as t cells, natural killer cells, and macrophages; (4) to alter growth patterns in cancer cells to promote behavior like that of healthy cells; (5) to block or reverse the processes that change a normal cell or a precancerous cell into a cancerous cell; (6) to enhance the ability of the body to repair or replace normal cells damaged or destroyed by other forms of cancer treatment, such as chemotherapy or radiation; and (7) to prevent cancer cells from spreading to other parts of the body. 30, 31 more of these promising new drugs are currently in development. it can be readily theorized that research to develop various brms can be subverted to a malicious end. that is, instead of using brms to suppress cancer growth or to decrease disease susceptibility, researchers could develop compounds to cause illness and death. other drugs could be designed to alter certain metabolic processes or to alter brain chemistry to affect cognition or mood. the opportunity for mischief is limited only by the imagination of the person with ill intent. the field of synthetic biology had its beginnings near the turn of the millennium with the idea that basic engineering principles could be applied to biological systems at the cellular and genetic levels to create new and improved organisms. synthetic biology is the "engineering of biology." 32 the goal and end products of the engineering are conventionally to be used for the benefit of humankind. however there has been increasing concern that synthetic biology could be used for nefarious purposes. [33] [34] [35] [36] a precise definition of synthetic biology has not yet been established; however, a consensus is building that synthetic biology is defined as the use of molecular biology tools and techniques to forward the engineering of cellular behavior. there is disagreement among scientists whether the new field of synthetic biology will allow terrorists to create biological agents with more lethal characteristics or create completely novel pathogens with enhanced pathogenicity and weaponization potential in an easier fashion. 37 some argue that synthetic biology causes "de-skilling" of biological techniques and allows laboratory processes to become easier for less experienced scientists or even laypeople to master. 35 others maintain that the tacit or unwritten laboratory skills that only a few highly trained scientists possess and which are very difficult to transfer and very difficult for a terrorist to acquire. 38 social scientists have carefully studied these tacit skills and argue that these techniques are very difficult to pass on from scientist to scientist without considerable effort that is often impossible even under the most ideal circumstances. 38 this difficulty was historically present in both the u.s. and soviet biological weapons programs each of which were extremely well funded and staffed with competent scientists. nevertheless, the rapid advance of synthetic biology has the potential to alter the present and future threat of biological weapons. 9, 35, 36 already, complete or partial genomic sequence data for many of the most lethal human pathogens (such as anthrax, plague, and the smallpox virus) have been published and are widely available via the internet. 39 in addition to the enormous explosion in our knowledge of human pathogens, there is a parallel increased understanding of the complexities of the human immune response to foreign agents and toxins. such knowledge has led to a deeper understanding of the development of basic immunity to a variety of different human infectious diseases. with this increase in scientific knowledge has come the power to manipulate the immune system at its most fundamental level. as we prepare for future threats, we must not ignore the potential quantum leap that synthetic biology offers to terrorists for developing new biological-warfare threats. examples of biological threats that could be produced through the use of synthetic biology include the following: (1) microorganisms resistant to antibiotics, standard vaccines, and therapeutics; (2) innocuous microorganisms genetically altered to produce a toxin, a poisonous substance, or an endogenous bioregulator; (3) microorganisms possessing enhanced aerosol and environmental stability characteristics; (4) immunologically altered microorganisms able to defeat standard threat identification and diagnostic methods; (5) genetic vectors capable of transferring human and foreign genes into human cells for therapeutic purposes 39 ; and (6) combinations of these with improved delivery systems. [40] [41] [42] the threats associated with the use of chemical weapons as battlefield or terrorist weapons are not easy to assess. 43, 44 risk assessment of use must take into account national laws, international treaties and 493 chapter 80 future biological and chemical weapons conventions, and the likelihood of adherence to these legal obligations. loopholes in existing agreements can be exploited to develop weapons that are technically not prohibited by international law. goals and objectives may vary depending on whether military use is planned at the strategic, tactical, or operational level and whether the developer is a national government, a breakaway republic, a kidnapped or recruited scientist, or a terrorist cell. risk of use may also depend on whether the targets are military versus civilian, human versus nonhuman (animals or plants, including livestock and crops), or individual (as in assassinations) versus large groups, and on whether the aim is death versus incapacitation. risk also depends on agent availability and on the technology available for production, storage, and dissemination; current advances in technology are associated with a higher risk of weaponization. two examples from the twentieth century and one from the twenty-first can illustrate the fallibility of intelligence: 1. during most of world war ii, the allied perception of risk from possible chemical agent use by axis powers focused on those agents, primarily pulmonary agents and vesicants, known from world war i. in fact, germany had developed a new kind of chemical-warfare agent, the compounds later to be called g-series nerve agents. their existence came as a complete surprise to western governments when, in the waning days of the european campaign, allied soldiers advancing into germany discovered buried nerve-agent munitions and entire nerve-agent factories. why these agents were never used on the battlefield is a topic of much speculation, but in retrospect they clearly posed the most lethal, yet unrecognized, threat from germany. 45 report was released and reported that "while a small number of old, abandoned chemical munitions have been discovered, isg judges that iraq unilaterally destroyed its undeclared chemical weapons stockpile in 1991." 47 although the initial intelligence proved to be wrong, it does not invalidate the argument that the risk from these agents, if possessed, would be very concerning. chemical agents originally used during world war i are sometimes considered obsolete, especially in comparison to the more potent nerve agents and incapacitating agents. however, agent potency is only one part of the story. to deliver the 5 μg that represents an estimated lethal dose for half of an exposed group (ld 50 ) of the nerve agent vx would seem to be easier than delivering the 3 to 7 g that constitute the ld 50 of sulfur mustard and more difficult than delivering the much smaller lethal doses of toxins such as botulinum toxin. 48 in fact, sulfur mustard is easier to synthesize than is a nerve agent and is easy to disseminate in a clandestine manner to create delayed effects. thus mustard still lays claim to being the "king of gases," and it has allegedly been used in a variety of venues since the end of world war ii. most known chemicals with toxicities equal to or greater than that of ammonia could theoretically be used as chemical warfare or terrorism agents. existing chemicals capable of weaponization for military or terrorist use include the following: existing chemicals remain candidate agents for future use. some compounds not developed to cause injury or incapacitation nevertheless can be very dangerous; hexachloroethane (hc) smoke, for example, can cause the same type of pulmonary damage induced by phosgene, a chemical weapon used in world war i. the cdc lists nearly 70 separate chemicals, including a variety of toxic industrial chemicals and poisons, as potential agents for terrorism. these include osmium tetroxide, long-acting anticoagulants, heavy metals, toxic alcohols, and white phosphorus. 49 the april 21, 2000, morbidity and mortality report included an even longer list of chemical agents that might be used by terrorists. 50 pyrolysis, the thermochemical change of an organic material in the absence of oxygen by heat, and products from explosions and conflagrations may release large quantities of cyanide and other toxicants that, although different from the original chemicals present, may still cause death. industrial chemicals are readily available in large quantities as preformed compounds and should be considered high on the list of potential terrorist agents. 51, 52 toxins that are chemicals produced within biological organisms also represent high-threat agents. 53 new chemicals are currently being synthesized on rigid three-dimensional molecular skeletons, the most promising of which are the norbornanes. norbornane is a bicyclic crystalline hydrocarbon (c7h12). 54 building on norbornane geometry allows for a modular enhancement of the number of functional sites on a given molecule. many norbornane derivatives, such as the mixture of chlorobornanes known as the toxaphenes, are persistent and have significant acute and chronic toxicity. these norbornane derivatives have been considered as potential candidates for new agents. novichok 55-58 (russian for "newcomer") refers to the alleged russian development of a highly toxic binary nerve agent or generation of nerve agents (sometimes called "fourth-generation" agents). only sketchy and unverifiable information is available in the unclassified literature, but the existence of these agents would demonstrate the possibility of creating new section vii topics unique to terrorist events and high-threat disaster response chemical compounds toxic enough to be used as chemical warfare or terrorist agents. one of the sources of unclassified information is from a dissident russian scientist who wrote newspaper articles and published a book about the novichok program and the types of chemical agents that were produced. 59 so-called gv analogs combining some of the properties of g-series and v-series nerve agents have also been suggested as potential new agents. 38 the use in 2002 of an incapacitating gas in the siege of a moscow theater taken over by chechen rebels was evidence of use of a chemical aerosol. 60, 61 the russian health minister at the time, after significant international pressure identified the aerosol as a fentanyl derivative and then stated that use of a fentanyl derivative was not prohibited by the chemical weapons convention. further investigations of survivors have suggested that carfentanil and remifentanil were possibly used in the siege. 62 organofluorines have been investigated because of their reported ability to defeat protective-mask or chemical-filtration systems. 38 other incapacitating agents under development exert primarily physical rather than chemical effects and include immobilizing agents ("stickums"), antitraction gels ("slickums"), and malodorants. 63, 64 an effective incapacitating agent must be highly potent and reversible. it also must have rapid onset, short duration of action, and a high safety margin. 63 nontraditional agents (ntas) are chemicals that do not fall in the traditional chemical weapons category but have been reportedly researched or developed for use as chemical weapons. "ntas are novel chemical threat agents or toxicants requiring adapted countermeasures," according to homeland security presidential directive/ hspd-18. 65 developing defenses against ntas is a listed priority for the u.s. department of defense. 41 ways in which existing or future battlefield chemical agents and delivery systems could be modified to improve performance must be considered. these modifications include the following: 1. agent thickening 2. binarization 3. micronization: "dusty agents" 4. developments in delivery systems a. dual-use cyberinsects and biorobots b. nanotechnology small quantities of thickening agents, such as acrylates, can be added to chemical agents to increase their viscosity. thickened agents are more persistent in the environment and in wounds than are nonthickened agents, and they are less easily decontaminated. 66 although no nation is currently known to stockpile thickened agents, the technology for their production is relatively simple and requires only standard chemical-warfare agents and the right proportion of a thickener. 67 many industrial chemicals and other poisons could theoretically be rendered more effective as battlefield or terrorist agents by thickening. in the 1950s, the u.s. army began to investigate the then-new technology of binarization, although production did not accelerate until the 1960s and deployment was not widespread until the 1980s. 58 a binary chemical weapon did not employ a new kind of agent but rather represented a novel way of producing and storing an already existing type of agent. the idea was to make storage of chemical rounds safer by stopping the production process at the penultimate synthetic step, resulting in two precursor compounds that when mixed would create the desired agent. these two precursors could then be stored separately. just before use, one component could be inserted into a round, where it would be separated from the other precursor by a thin membrane. the impact and momentum of the launch of the projectile would burst the membrane to allow for mixing of the components and in-flight production of the chemical agent. in practice, this process was often not complete, but the 20% or so of ancillary reaction product was often extremely toxic by itself. binarization or some similar production-arrest method could theoretically be used by a clandestine terrorist cell to help evade detection and to decrease the risks associated with the production, transportation, and use of chemical agents. micronization is a type of particularization involving the production of extremely fine particles onto which a chemical agent can be adsorbed. during world war ii, germany explored particularization of sulfur mustard onto small carrier particles of silica (silicon dioxide), although other powdered silicates (e.g., talc, diatomite, and pumice) and clays (e.g., kaolinite and fuller's earth) can also be used. 68 the advantages of such "dusty agents" are increased volatility, facilitation of the movement of relatively nonvolatile agents such as sulfur mustard and the persistent nerve agent vx into the alveoli, and increased penetration of clothing and chemical protective equipment. 31 iraq used a "dusty mustard" composed of 65% sulfur mustard adsorbed onto silica particles ranging in diameter from 0.1 to 10 μm during its war with iran. micronization of a variety of chemical, biological, and toxin agents requires a certain degree of technological sophistication that is becoming increasingly easy to acquire. agent delivery can potentially be modified in a variety of ways in addition to thickening and micronization. the jordanian government released a report in 2004 of the discovery of an elaborate plot by al qaeda terrorists for a two-stage attack using a massive vehicle-borne improvised explosive device followed by the release of toxic chemicals to include acetones, nitric acid, and sulfuric acid. 69 similarly, enhancedfragmentation munitions could be used in combination with chemical agents to drive the agents more effectively into the body. innovative new delivery systems taking advantage of advances in robotics include the proposed use of cyberinsects and biorobots to deliver biological agents, chemical agents, or toxins. 70 engineering on an even smaller scale is the purview of nanotechnology, also called "micromechanical engineering" and "micro-electromechanical systems." 71 nanotechnology takes advantage of the unique properties of materials on the scale of about a nanometer (10 to 9 m) 72 and deals with the molecule-by-molecule or even atom-by-atom assembly of materials. nanoparticles behave in unusual and unpredictable ways, are small enough to enter cells easily, and in fact are being developed to provide not only better storage and dispersal of pharmaceutical products but also more efficient transport of both biological organisms (e.g., viruses) and chemical compounds into the body. 71 in some cases they may be surprisingly toxic, partly because of the ease with which they can cross membranes, including the blood-brain barrier, and enter cells. 73 this toxicity could be exploited by governments or terrorist organizations interested not only in small-particle delivery of chemical agents but also in the ancillary and perhaps synergistic effects of the carrier materials themselves. nanomaterials can be encapsulation compounds such as fullerenes, or buckyballs, which are hollow 60-carbon geodesic shells; nanoshells (e.g., a gold shell surrounding an inert silica core); a "self-assembled, polyamino acid nanoparticles system" under development in france; or dendrimers, which are onion-like layers of shells surrounding a biologically active core. 72 any of these materials could be used to deliver existing or new chemical agents. other nanomaterials include selfassembling liquids composed of cylindrical nanofibers (each 6 to 8 nm in diameter) that solidify upon injection to form structured scaffolds capable of presenting ordered peptide signals to cells. a ferrofluid such as a colloidal suspension of nanoscale ferrous oxide can be coupled 495 chapter 80 future biological and chemical weapons with antibodies in a laboratory to detect and concentrate rare human cells in a diagnostic setting, but this technology could easily be adapted to target those cells in vivo. quantum dots are nanoscale semiconductor crystals that show promise in the in vitro and in vivo diagnosis of a variety of conditions; although their main use is projected to be in the laboratory, animal experimentation involving injected quantum dots has demonstrated successful targeting of lymph nodes and of prostate-cancer xenografts in mice. adverse health effects from any of these kinds of nanoparticles could represent a primary goal for military or terrorist operatives in addition to the toxicity of any other chemicals delivered by the nanoparticles. for example, water-soluble fullerenes (or buckyballs) have caused brain damage in largemouth bass. 74 also, dendrimers can cause osmotic and membrane damage and can activate the clotting and complement systems. quantum dots composed of selenium, lead, and cadmium could release those metals into cells, depending on the composition of the surface coating of the dots, and cause damage. 72 "designer" chemicals from biotechnological processes biotechnology refers to "any technological application that uses biological systems, living organisms, or derivatives thereof, to make or modify products or processes for specific use." 75 biotechnology includes such time-honored practices as the baking of bread and the brewing of beer, but in the twenty-first century refers in particular to genetic engineering, that is, the artificial transfer of genes from one organism to another and the consequent alteration of the genetic structure of a cell. 76 it is founded on the basic sciences of genomics (the study of the genetic composition of an organism) and proteomics (the study of the expression of the genome by means of protein synthesis). "designer" chemicals could be produced from biotechnological processes. these processes include the following: (1) combinatorial chemistry and ligand modification; (2) genomics and target identification; (3) microarrays, proteomics, and rational agent design; and (4) toxicogenomics, database mining, and the prediction of toxicity. 77 these developments, if used for chemical warfare agents, would be considered "dual-use technology." this is technology that can be used for both peaceful and military aims. combinatorial chemistry is the production of complex sets, or socalled libraries, of related compounds, as in the case of the norbornane derivatives previously described. automated screening techniques to select for library elements with desired toxic effects on specified target organs can process several hundred thousand compounds a day against several dozen different proteins. this obviously accelerates tremendously the development of new chemical agents. genomics has benefited enormously from three modern scientific efforts: the human genome project, the human genome diversity project, and gene therapy. 78 identification and cataloging of hundreds of single-nucleotide polymorphisms (individual sequence variations) allow for the selection of genomic sequences to be mass-produced for insertion into cells to create a specific effect. targeting unusual sequences of high prevalence in certain populations raises the specter of genomic, or ethnic, weapons, as previously described. less appreciated is the potential for genomics to be used to develop drugs and chemical or toxin agents that can also be targeted to specific variants within a population of humans, animals, or crops. the widespread availability of genome libraries on the internet makes it nearly impossible to control or restrict access to the already published genomic libraries on over a hundred microbial pathogens. 79 proteomics complements genomics by characterizing the protein expression of segments of the genome and by making it easier to develop compounds that target or produce a specific protein. direct gene insertion, genetic delivery via virus or bacteria, or drug tailoring to affect a given protein can be used. for example, a scorpion toxin has been successfully engineered into a virus that acts as a pesticide against caterpillars. protein sequences in toxins are partly responsible for resistance to light, oxygen, moisture, and desiccation; the insertion of genes to create altered proteins or the introduction of chemical agents engineered to cause structural changes in expressed proteins could significantly alter the toxicity of a given compound. 77 furthermore, the widespread use of dna microarrays (glass slides or chips imprinted with thousands of specific single-stranded dna sequences) allows for fast-automated screening of candidate compounds. scientists involved in the selection and evaluation of specific chemical agents can now use toxicogenomics (the study of genetic variation of response to toxins) and data mining (the computerized analysis of databases of drug and chemical information via sophisticated neural nets) as tools to eliminate less likely candidates and to algorithmically predict compounds with high toxicity or with other desired characteristics relating to environmental persistence, toxicokinetics (absorption, distribution, biotransformation, and elimination), and toxicodynamics (mechanism of action). such tools will undoubtedly lead to the development not only of new pharmaceutical agents but also of designer toxins for military or terrorist use. 77 if history is any guide, new biological and chemical weapons and novel "mid-spectrum" agents (e.g., toxins, bioregulators, synthetic viruses, and genocidal weapons) will be developed in the future, and new modifications will be found to improve the production, weaponization, storage, delivery, and action of existing agents. 33, [80] [81] [82] naturally occurring emerging infectious diseases provide examples of newly identified pathogens with weaponization potential, and midspectrum agents such as toxins and bioregulators will undoubtedly assume more prominence with the accelerating pace of synthetic biology. agents of any category can theoretically be engineered to target specific genes or proteins with differential population prevalence to produce genomic or ethnic weapons; and advances in proteomics, toxicogenomics, and computerized database mining could be used for the rapid and efficient development of not only new drugs but also new chemical agents for terrorism. 9, 17, 68 synthetic biology has now advanced to the point that no special equipment is required beyond that available to any modern molecular-biology laboratory, and the scale of operations is also well within the means of governments and terrorist groups. 78 the threats from future modification of existing agents and from the development of new agents, new agent-development technologies, and innovative delivery systems should not and must not be underestimated. historical aspects of medical defense against chemical warfare the great smallpox epidemic of 1775-82 unit 731: japan's secret biological warfare in world war ii. london: hodder and stoughton biological warfare. a historical perspective the biology of doom: america's secret germ warfare project the sverdlovsk anthrax outbreak of 1979 the history and threat of biological warfare and terrorism toxic terror: assessing terrorist use of chemical and biological weapons follow-up of deaths among u.s. postal service workers potentially exposed to bacillus anthracis-district of columbia airborne infection with bacillus anthracis-from mills to mail united nations mission to investigate allegations of the use of chemical weapons in the syrian arab republic report on the alleged use of chemical weapons in the ghouta area of damascus on 21 the chilling true story of the largest covert biological weapons program in the world, told from the inside by the man who ran it the challenge of emerging and re-emerging infectious diseases the spanish influenza pandemic of 1918-19: new perspectives. london ebola: not an effective biological weapon for terrorists centers for disease control and prevention. emergency preparedness and response: bioterrorism agents/diseases avian influenza a virus infections of humans structure of the uncleaved human h1 hemagglutinin from the extinct 1918 influenza virus interim guidance for protection of persons involved in u.s. avian influenza outbreak disease control and eradication activities factors associated with transmission of severe acute respiratory syndrome among health-care workers in singapore surveillance case definition for severe acute respiratory syndrome (sars) and update on sars cases-united states and worldwide severe acute respiratory syndrome: public health guidance for community-level preparedness and response to sars. version 2. supplement f: laboratory guidance. appendix f8-guidelines for laboratory diagnosis of sars-cov infection hendra virus disease & nipah virus encephalitis emerging viral diseases: an australian perspective ribavirin therapy for nipah virus infection bioregulators as instruments of terror therapeutic strategies for rheumatoid arthritis an evaluation of bioregulators as terrorism and warfare agents office of technology assessment. technologies underlying weapons of mass destruction a brief history of synthetic biology the future of biological warfare biology's brave new world: the promise and perils of the synbio revolution could terrorists exploit synthetic biology the promise and perils of synthetic biology synthetic biology and biosecurity: how scared should we be. london: king's college london the social context shaping bioweapons (non)proliferation genome projects and gene therapy: gateways to next generation biological weapons is all fair in biological warfare? the controversy over genetically engineered biological weapons department of defense chemical and biological defense annual report to congress synthetic biology: biology by design defending america: iraq and other threats to the us involving weapons of mass destruction combating terrorism: need for comprehensive threat and risk assessments of chemical and biological attacks medical aspects of chemical and biological warfare, part i cbrne: incapacitating agents national intelligence estimate: iraq's continuing programs for weapons of mass destruction overall view of chemical and biochemical weapons biological and chemical terrorism: strategic plan for preparedness and response. recommendations of the cdc strategic planning workgroup industrial chemicals: terrorist weapons of opportunity agency for toxic substances disease registry. industrial chemicals and terrorism: human health threat analysis, mitigation and prevention toxins as weapons of mass destruction. a comparison and contrast with biological-warfare and chemical-warfare agents neutron spectroscopy of norbornane russia's toxic threat ex-soviet scientist say gorbachev's regime created new nerve gas in '91 the chemical weapons convention: implementation challenges and solutions center for nonproliferation studies history of chemical and biological warfare: an american perspective state secrets: an insider's chronicle of the russian chemical weapons program chemical weapons: documented use and compounds on the horizon incapacitating chemical weapons: a year after the moscow theatre siege analysis of clothing and urine from moscow theatre siege casualties reveals carfentanil and remifentanil use incapacitating biochemical weapons introduction to chemical and biological weapons homeland security presidential directive/hspd-18. medical countermeasures against weapons of mass destruction medical aspects of chemical and biological warfare, part i the chemical warfare threat and the military healthcare provider the danger to the chemical weapons convention from incapacitating chemicals. first cwc review conference paper no. 4. strengthening the chemical weapons convention was chemical bomb target biological warfare, bioterrorism, biodefence and the biological and toxin weapons convention news: nanotechnology: the potential for new wmd. available at the ups and downs of nanobiotech nanotechnology precaution is urged: miniscule particles in cosmetics may pose risk, british scientists say manufactured nanomaterials (fullerenes, c60) induce oxidative stress in the brain of juvenile largemouth bass the free encyclopedia united nations environment program. governments to advance work on cartagena protocol on biosafety the plague makers: the secret world of biological warfare nrc wants genome data unfettered the threat of mid-spectrum chemical warfare agents the future of weapons of mass destruction: their nature and role in 2030 bioregulators as prototypic nontraditional threat agents the authors gratefully acknowledge the previous contributions of james m. madsen. key: cord-340131-refvewcm authors: kache, tom; mrowka, ralf title: how simulations may help us to understand the dynamics of covid‐19 spread. – visualizing non‐intuitive behaviors of a pandemic (pansim.uni‐jena.de) date: 2020-06-04 journal: acta physiol (oxf) doi: 10.1111/apha.13520 sha: doc_id: 340131 cord_uid: refvewcm the new coronavirus sars‐cov‐2 is currently impacting life around the globe (1). the rapid spread of this viral disease might be highly challenging for health care systems. this was seen in northern italy and in new york city for example(2). governments reacted with different measures such as shutdown of all schools, universities and up to a general curfew. all of those measures have a huge impact on the economy. the united nations secretary general has stated recently: “the covid‐19 pandemic is one of the most dangerous challenges this world has faced in our lifetime. the most dangerous challenges this world has faced in our lifetime. it is above all a human crisis with severe health and socio-economic consequences" 3 . according to the european mortality observing network 4 : "pooled mortality estimates from the euromomo network continue to show a markedly increased level of excess all-cause mortality overall for the participating european countries, coinciding with the current covid-19 pandemic" (see figure 1 ). there are many aspects that need to be considered when considering measures such as lockdowns. one of the main problems is that aspects of the dynamical behavior are hard to grasp with our thinking in the context of our experiences. our understanding of cause and relationship is mostly related to "linear thinking". this means that there is a linear relationship between a potential influence and a result. an example would be "if you buy twice as many apples you have to pay twice the amount of money." this concept however does not work for exponential growth rates. those are governed by rules that are highly sensitive to the conditions of the underlying process. in the context of a viral disease we can consider the concept of growth rate describing the factor that describes the number of affected individuals divided by the number the previous day. for example, if you change the growth rate in an ideal exponential scenario from 1.1 by 10 % to 1.21 you will get after only 10 days more than twice and after 20 days more than 6 times the amount (see figure 2 ). in the case of covid19 it has been estimated that the reproductive number can be as large as 3.9 5 in this particular study. the exponential growth however could only be found in a situation where there are an infinite number of individuals that can get viral disease. this is obviously not the case. exponential-like growth can be observed only in the beginning of the spread, where the number of people that do not have the disease is a lot bigger in comparison to the infectious individuals. in order to understand the spread and aspects of the dynamics of the disease we can use a modeling approach 6 . here we explore an agent-based model where dots as surrogates for individuals move around in a given space. the simulation can be found at https://pansim.uni-jena.de. each of the dots has one of the following states: susceptible, infectious without symptoms, infectious with symptoms, recovered, and immune. once infected, the dots go through the disease cycle described by the states. the dots move around in the space with a given "mobility". once the dots get aware of the disease they change their mobility. the ultimate goal would be to eradicate the sars-2 virus worldwide. this is unlikely to achieve on a shortterm scale due to its pandemic characteristic. when the pandemic runs through the population a key figure is the maximum number of active cases. since a certain percentage of the active cases will require this article is protected by copyright. all rights reserved intensive care the maximum number of active cases determines whether or not the healthcare system runs at its limit or not. in order to model different theoretical scenarios, we have implemented predefined parameter settings. the different scenarios described in the following relate a "default" parameter setting that is used to compare the modifications. if the population has been in contact with the virus previously. in scenario iii, a third is immune to the disease, whereas in scenario iv two thirds is. as expected the maximum number of active cases is drastically reduced. scenario v: sick people remain mobile: if sick people remain mobile, the maximum number of active cases will increase. this suggests that it is beneficial for slowing the spread that the number of contacts of an infectious person low. scenario vi: infection detected earlier: this would reflect a situation where testing of the disease is available on a large scale to identify infections individuals. in the model, infected agents reduce their mobility earlier and therefore the transmission rate is reduced. here we simulate the case that a given number of individuals have a much higher mobility than the rest of the simulated population. this high mobility is kept at all times in the simulation. the simulation shows much higher values for the maximum number of active cases. scenario viii and ix: lock down with a threshold: these scenarios describe a population-wide lockdown based on a threshold of active cases at a given time. the lockdown measure reduces the maximum this article is protected by copyright. all rights reserved number of active cases in the model. if the lock-down is imposed at a lower threshold the effect is stronger. a second value allows setting, to which fraction of the lockdown threshold the number of active cases needs to drop in order to lift the lockdown. as pointed out earlier, the behaviour of the spread is highly sensitive to the initial parameters. that is why it is difficult to predict the spread and hard to estimate when to impose drastic measures such as lockdowns. there are however features in the model that are important to note and give consistent results over a broad range of parameter sets. for example that if one reduces the infection probability; the speed of the spread is slowed down. that means that any measure that leads to the reduction of that probability would reduce the maximum number of active cases and hence would help to reduce the risk of pushing the health care system to its limit. a complicating issue with covid-19 is that a substantial number of people show only mild or even no symptoms 7 , hence remaining undetected while spreading the virus. one other problem of controlling the disease is that there are delays in the underlying biology. it is known systems that have time delays in the feedback loop might exhibit oscillatory behaviour 8 . this is what we see in the model (figure 3) for the active cases. understanding the dynamical aspects of the disease is crucial for proper control. it is therefore important to educate the public about the unintuitive behavior of the spreading of a disease like covid-19. the agent-based model might be a good tool to communicate the epidemiological characteristics. the model makes it possible to explore the effects of the different parameters on the behaviour of the spread and key outcomes such as peak number of active cases or total number of affected individuals. the input parameters can be changed interactively on the website. according to the model one effective way to slow down the disease would be to reduce the contacts of infectious individuals. in this regard it is discussed whether it might be helpful to trace contacts and to test for the disease in a targeted manner 9 . the world health organization (who) has defined priorities for research related to covid-19. among those top priorities is "the natural history of the virus, its transmission and diagnosis" 10 . from an epidemiological point of view the way of transmission has a prominent role. for example, it would have a tremendous impact on the prevention strategy if the virus would be transmitted over food or drinking water such as e. coli bacteria or sapovirues 11 . the model also shows that super mobile individuals that have contact to many others contribute to a faster spread if they are infectious. this article is protected by copyright. all rights reserved it appears to be certain that the topic of covid-19 with currently 41.36.056 confirmed cases and 283.478 total deaths worldwide (as of 11 th of may 2020) 12 will engage our interest for the next months. however one should not forget that other health topics are also on the table. according to the latest numbers: "tobacco kills more than 8 million people globally every year. more than 7 million of these deaths are from direct tobacco use and around 1.2 million are due to non-smokers being exposed to secondhand smoke" 13 , this is all due to an behaviour that could been stopped immediately, at least in theory. agent based model: the agent-based model (abm) presented here is based on the idea of the classical sir model and expanded by additional states. the sir model describes the spread of an infectious disease in a population using three population states quantities: susceptible (s), infectious (i), and recovered (r) 5 . the dynamics of the sir model is governed by different transition rates between the states, that can be explicitly modeled by a system of differential equations. pansim, on the other hand, simulated a population of entities (agents) that behave individually according to simple rules in a simulation space. in contrast to a mathematically formulated model, such rules can be easily understood and yet can produce similar outcomes despite the stochastic nature of abms 14 . a simple abm can approximate the sir model by using three instructions that each agent obeys: i.) move through the simulated space, ii.) if a susceptible agent is near an infected agent, change state from susceptible to infected, iii.) if an agent is infected and a certain time has passed, change the state to recovered (see fig. 1 ). behavior rules can be programmed to reflect individual behavior of agents even more realistically. this makes it possible to produce complex interactions between agents, which may lead to unexpected, emergent behavior from those simple rules 15, 16 . technical implementation: pansim was written in javascript using the p5.js 17 and plotly.js 18 libraries. p5.js is a graphical library that allows for simple creation of visual elements in javascript. build-in functions of p5.js were used to create graphical representations of the agents within the pansim simulation.p5.js is built around the 'draw()' main loop function, which is periodically called and is used to progress the simulation and the graphical representation at a specified frame rate. plotly.js has been employed for generating live-updating plots of the data produced during the model run. the pansim model uses object-oriented programming to produce individual agents with the 'person' object. 'person' describes the states agents can assume (susceptible, infected, recovered, immune), how this article is protected by copyright. all rights reserved they behave, and how agents are graphically represented. the 'town' object takes user input and builds a corresponding population of agents and passes down user-defined behavior parameters to the agents. thus, users of pansim can change simulation parameters as desired. after the model parameters are set, the simulation starts and updates the agents position using a random walk by iteratively calling p5.js's 'draw()' function. after each update interval, the agents check if they are in contact with an infected agent, i.e. if the distance between any agent is less than their radius, and change their status accordingly. the probability of infection upon contact can be specified, so that not every contact leads to an infection. if an agent is infected, an internal timer is incremented after each update interval. an infected agent progresses from an infectious but asymptomatic time interval (time infectious without symptoms, tiwos) to an infectious time interval with symptoms (tiws) 19 . agents can change their mobility during the symptomatic time interval. the user can set for how many simulation updates the tiwos and tiws phase should last. in addition to the three classic states from the sir model, a population of super mobile agents can be included. those agents have three times mobility of susceptible agents throughout the model run despite their infection state. the user has the option to introduce a population-wise lock down by specifying a threshold of symptomatic agents. in that case, all agents reduce their mobility except super mobile agents. the number of susceptible (grey), symptomatic (red) and total number of infected agents (black) is counted after each update interval and displayed in a live-updating plotly.js graph. the fraction of the population that got infected and the maximum fraction of active cases is displayed in the control panel of pansim to allow users to compare different simulation outcomes. conflict of interest: none. : exponential growth is highly sensitive to its parameters. in this example the growth rate of 1.1 (blue) has been increased only by 10% (red). figure 3 : behaviour of the model with a lockdown criterion based on a threshold (screenshot from https://pansim.uni-jena.de). in this setting a second lockdown was required. notice after imposing the lockdown the number is still increasing, in this case by a factor of two to three. this overshoot is much smaller in the second lock-down phase. this is due to the fact that a considerable fraction got infected and turned to the recovered/immune stage. if however this fraction is small, steep increase would be possible again. a: sars-cov-2: what do we know so far? acta physiologica team: coronavirus: the first three months as it happened statement by the secretary-general -on covid-19 w: preliminary prediction of the basic reproduction number of the wuhan novel coronavirus 2019-ncov centre for mathematical modelling of infectious diseases, c-wg: early dynamics of transmission and control of covid-19: a mathematical modelling study estimating the asymptomatic proportion of coronavirus disease 2019 (covid-19) cases on board the diamond princess cruise ship sh: death by delay south korea is reporting intimate details of covid-19 cases: has it helped? nature, 2020. 10. who news relaease: world experts and funders set priorities for covid-19 research it: two drinking water outbreaks caused by wastewater intrusion including sapovirus in finland who statement: tobacco use and covid-19 j: an open-data-driven agent-based model to simulate infectious disease outbreaks lm: brief introductory guide to agent-based modeling and an illustration from urban health research km: agent-based modeling in public health: current applications and future directions processing simplicity times javascript flexibility, 2020 (www document this article is protected by copyright. all rights reserved key: cord-255140-3dwqqgv1 authors: christian, michael d. title: biowarfare and bioterrorism date: 2013-07-04 journal: crit care clin doi: 10.1016/j.ccc.2013.03.015 sha: doc_id: 255140 cord_uid: 3dwqqgv1 bioterrorism is not only a reality of the times in which we live but bioweapons have been used for centuries. critical care physicians play a major role in the recognition of and response to a bioterrorism attack. critical care clinicians must be familiar with the diagnosis and management of the most likely bioterrorism agents, and also be adequately prepared to manage a mass casualty situation. this article reviews the epidemiology, diagnosis, and treatment of the most likely agents of biowarfare and bioterrorism. weapons and their potential use by states with the consideration of use by nonstate actors (eg, terrorists). 3 more recent definitions of bioterrorism include both a broader range of potential biological agents as well as considering more diverse groups of potential targets and impacts of biological weapons. spencer 4 defines bioterrorism as "the use of micro-organisms as weapons of catastrophic effect which can be described as: the category or method of use of a weapon system that results in a significant negative impact on a nation's physical, psychological or economic well-being, thereby causing a major modification of routine activity." this definition highlights several key points. first, it highlights that a wide range of microorganisms must be considered and that their impact is not merely physical but may also include psychological and economic factors. spencer further elaborates on his definition of bioterrorism by stating "bioterrorism is best described as the use of micro-organisms (pathogens) or the products of living organisms (toxins) to inflict harm on a wider population, including animals and crops." 4 the elaboration on his original definition highlights that not only are humans directly vulnerable to bioterrorism but we are vulnerable through indirect attacks on our livestock or crops, which has also been termed agroterrorism. 5, 6 other investigators broaden the definition further to not only include microorganisms and biotoxins but also larger organisms, specifically insects. 6 as with many other medical issues that intensivists face in their busy clinical and academic practices, pressured by ever-increasing time and budgetary restraints, there is a necessity to prioritize efforts and resources toward the most common and higherimpact concerns. it is difficult to provide a clear-cut answer as to where bioterrorism should be prioritized on this list. although some experts state that the risk of a largescale bioterrorist attack is low, 7 in a more recent analysis, us senators graham and talent quote their conclusion form the commission on the prevention of weapons of mass destruction proliferation and terrorism in 2010, which stated "unless the world community acts decisively and with great urgency, it is more likely than not that a [biologic] weapon of mass destruction will be used in a terrorist attack somewhere in the world by the end of 2013." 8 anthrax in particular remains such a concern, because of both the lethality of the agent and also the potential availability given the number of governments that produced weaponized anthrax in the past. 9 the later issue is of concern because of both the availability of expertise as well as the risk of residual caches of anthrax in failed states that are vulnerable for misappropriation. in addition, it may require less expertise to develop aerosolizable anthrax then previously believed. 8, 9 box 1 lists the capabilities required of any organization, whether state or nonstate, to conduct and deliver a bioterrorist attack. even a small or moderate bioterrorism event has the potential to overwhelm local medical resources and cause significant civil and economic disruption as a result of the psychological impact of such an attack. given the potential risk for a bioterrorism event and the major impact that could occur, these factors alone provide a strong argument as to why an understanding of bioterrorism is warranted for critical care physicians. however, if these arguments are not persuasive enough, the knowledge that critical care physicians will play a key role in a bioterrorist event, and that their effectiveness in responding to the event is dependent on their medical knowledge regarding bioterrorism agents, should compel one to take the time to read this review. agents used for bioterrorism typically cause critical illness and therefore are of clinical relevance to the intensivist. several of the category a and b organisms (see later discussion) also produce human infections in nature, and therefore, knowledge of their presentation and treatment can also be applied in nonbioterrorism settings. the us centers for disease control and prevention (cdc) has identified and categorized a list of potential bioterrorism agents ( table 1 ). the agents identified by the cdc have been accepted, by most authorities globally, as the highest priority for preparedness and research. those agents from category a form the primary focus of this article. however, relman 10 reminds us that it is important not to solely focus on the agents from the cdc list because they were largely driven by past military programs and do not include agents that were not of particular interest or relevance in the past but may be in the future given technological advances. in addition, military programs focused on weaponizing agents, whereas terrorists could seek to manipulate the natural spread of existing organisms or the development of novel strains. 10 agents this section provides a basic overview and description of the commonly considered bioterrorism threats. specific details about the epidemiology, diagnosis, treatment, and outcomes associated with each of the organisms follow in the subsequent sections. anthrax the bacteria causing anthrax is bacillus anthracis, an encapsulated, grampositive, spore-forming bacilli. 7, 9, 11, 12 when seen on a gram stain, it is often described as box cars given its appearance as a series of railway boxcars in a train viewed from above ( fig. 1) . bacillus anthracis is a soil-borne organism and can be found in the environment globally. the organism grows quickly on standard culture media (6-24 h) and its spores are highly resistant, potentially being viable for decades. in naturally occurring infections, the organism may infect humans by transcutaneous inoculation (cutaneous anthrax), ingestion (gastrointestinal [gi] anthrax, including oropharyngeal), or inhalation (thoracic anthrax). in terms of use as an organism of bioterrorism, it is most likely to be delivered in its spore form as an inhaled agent. once inhaled, bacillus anthracis spores enter alveolar macrophages by phagocytosis and are transported to regional lymph nodes, where they germinate, typically within 2 to 5 days but they may be delayed up to 60 days. 9, 13, 14 symptoms start after germination and bacterial replication begins to occur. what causes the variability in incubation period associated with the time from infection with spores to germination to the vegetative bacillus is unknown. bacillus anthracis produces 2 exotoxins: edema and lethal, comprised of 3 components: (1) edema factor (ef), which impairs neutrophil function and disrupts cell water hemostasis, resulting in massive edema; (2) lethal factor (lf), which causes release of tumor necrosis factor a and interleukin 1b, which are believed to mediate the severity of illness produced; and (3) protective antigen, found in combination with both ef and lf, which allows binding and transportation of the other 2 toxins to and across cell membranes. in addition to the toxins, there are several other virulence factors, including an antiphagocytic capsule. anthrax is not transmitted person to person, but patients should be isolated with droplet precautions as part of standard febrile respiratory infection precautions until the cause of their illness is confirmed. patients with cutaneous anthrax require contact precautions. decontamination should include removing clothing and jewelry and washing skin with soap and water. caution must be exercised not to generate secondary aerosols when handling contaminated clothing items. environmental contamination from patients exposed to an aerosol of bacillus anthracis spore can be performed with a 0.5% hypochlorite bleach solution. this procedure is not sufficient for site decontamination after the release of bacillus anthracis spores, because much more extensive decontamination is required. [15] [16] [17] plague plague is caused by yersinia pestis, a nonmotile, gram-negative, bipolar coccobacillus that can be found worldwide. 7, 18, 19 human infections occur in nature regularly, and plague is endemic in regions such as the southwest united states. yersinia pestis virulence factors include: v and w antigens, lipopolysaccharide endotoxin, capsular envelope (antiphagocytic), coagulase, and fibrinolysin. the natural reservoir for yersinia pestis is rodents and the vector to humans is the oriental rat flea (xenopsylla cheopis). plague is highly communicable in the pneumonic form and may present as any of the following clinical syndromes: bubonic plague, primary pneumonic plague, primary septicemic plague, plague meningitis, plague pharyngitis, pestis minor, and subclinical infection. aerosolized yersinia pestis likely produces a clinical presentation identical to pneumonic plague. however, yersinia pestis does not produce spores and is susceptible to destruction by drying, heat, and ultraviolet light, therefore making it significantly more challenging to weaponize it then anthrax. it is more readily transmitted via an infected vector, such as the oriental rat flea, or by person-to-person transmission. the incubation period depends on the clinical presentation; bubonic plague takes 2 to 8 days and pneumonic plague as short as hours to 3 days. isolation for plague requires droplet precautions until 48 hours of effective antibiotic therapy. decontamination is not required specifically for plague. health care institutions should follow their usual procedures for cleaning after patient discharge from a room. tularemia francisella tularensis is the organism responsible for causing tularemia; it has often been associated with rabbits and various rodents and is occasionally referred to as rabbit fever and deer fly fever. 7, [19] [20] [21] francisella tularensis is a fastidious, small gram-negative, facultative intracellular coccobacillus able to live in soil, water, and decomposing carcasses for long periods. francisella tularensis can be transmitted by direct contact with mucous membranes, cutaneous inoculation through broken skin or bites from infected ticks (or other arthropods), ingestion and inhalation (although human-to-human transmission has not been reported 20 ) . as discussed further in the epidemiology section, tularemia has been previously used as a biological weapon. the mode of deployment as a biological weapon in the past has often been through infected vectors 22 ; however, a modern bioterrorist would most like deploy the agent via aerosolization and it could present as: primary pneumonic tularemia (inhalation), oculoglandular tularemia (eye contact), ulceroglandular (broken skin contact), or oropharyngeal (mucous membrane contact without deep inhalation). tularemia is not communicable between humans, so specific isolation is not required; only standard universal precautions should be used. decontamination for aerosolized exposure should involve remove clothing and jewelry and washing the skin with soap and water. q fever the only category b bacterial organism to receive any extensive discussion in this article is the zoonotic rickettsial organism coxiella burnetii, the causative agent of q fever in humans. [23] [24] [25] coxiella burnetii is a small, obligate intracellular, gramnegative highly pleomorphic coccobacillus. it has a typical gram-negative cell wall structure, but does not stain well with gram stain although it can be seen with a gimenez stain. coxiella burnetii has 2 morphologic forms: large and small cell variants. the small cell form is the extracellular form, which is metabolically inactive and resistant to chemical agents as well as environmental and physical conditions. the large cell variant is the metabolically active and pathogenic intracellular form. the disease is called q fever for query fever, because the causative organism had not been identified at the time of the first documented large outbreaks. q fever occurs worldwide and is associated with contact with sheep, goats, or cattle, particularly during birthing when placental exposure occurs. human infection is through the inhalation of aerosolized organisms from infected animals. coxiella burnetii is highly infectious to humans, and its sporelike small cell variant form makes it a potentially viable biological weapon. most human cases of disease are zoonoses, with human-to-human infection only rarely reported; therefore, only standard infection control precautions are required in the clinical setting. however, coxiella burnetii is highly infectious from culture and therefore should always be handled under biosafety level 3 conditions. environmental decontamination should a culture spill or bioterrorism release occur is difficult given the highly resistant features of the small cell variant. 23 other bacterial agents in addition to the bacterial agents already discussed, there are several other potential organisms that could be used as bioweapons (see table 1 ). [24] [25] [26] rickettsia prowazekii is the causative agent of louse-borne typhus and, similar to q fever, efforts have been made in the past to develop it into a bioweapon. brucella sp are small, aerobic, intracellular gram-negative coccobacilli, which cause primarily zoonotic infections in sheep, cattle, goats, and other animals. brucellosis can manifest as either christian systemic or localized infections and is acquired by contact with or ingestion of fluids from infected animals, particularly by the consumption of unpasteurized milk or cheese. brucella suis was reportedly weaponized by the united states, and possibly other countries, in the past. 24 the organism was to have been aerosolized from a bomb. burkholderia pseudomallei is a gram-negative, facultative anaerobic, motile bacillus that causes melioidosis. there are naturally occurring cases of melioidosis in southeast asia and australia yearly. the disseminated form causes an acute illness, with high mortality. the disease is also known to cause abscesses, which, in some cases, do not present until many years after the initial exposure. burkholderia mallei is a gram-negative, aerobic, nonmotile bacillus that causes glanders. it exists only in living organisms and cannot survive in the external environment. glanders usually presents in humans as a nodular disease with regional lymphangitis; however, systemic dissemination of the organism does occur on occasion, producing septic shock, potentially leading to death. control measures have led to the eradication of glanders from most countries in the world, with the exception of ongoing zoonotic endemics in parts of the middle east, asia, africa, and south america. both burkholderia sp tend to be transmitted to humans through inoculation via a break in the skin, although inhalation is occasionally also a means of human infection. both organisms have been studied, and reportedly developed, as biological weapons in the past by the germans and russians. 26 given the possibility of human-to-human transmission, droplet precautions should at least initially be used for patients with systemic or pulmonary involvement. the more significant transmission risk is from cultured organism, and therefore, any cultures should be managed in a biosafety level 3 laboratory. smallpox once believed to be a disease of the past, smallpox is now one of the most significant bioterrorism threats to the world. 7, 19, [27] [28] [29] there are no animal reservoirs for smallpox and the last naturally occurring case was in somalia in 1977. 28 smallpox is caused by the dna variola virus, a member of the genus orthopoxvirus (family poxviridae), which also includes molluscum contagiosum, vaccinia (virus used in the smallpox vaccine), and monkeypox. a large vulnerable population has existed since smallpox vaccination was stopped in the late 1960s and early 1970s. variola is highly communicable, and can be transmitted by airborne, droplet, and fomite (bed linen and clothing) transmission in addition to direct contact. infection starts in mucosa of airway, then the virus replicates in regional lymph nodes before subsequently resulting in an asymptomatic primary viremia 3 to 4 days after infection with spread to the bone marrow, spleen, and lymph nodes. secondary viremia occurs (day [8] [9] [10] [11] [12] , in which the virus localizes in the dermis and oropharyngeal mucosa and is marked by the onset of symptoms and infectivity. five clinical syndromes of variola infection are seen: classic (unvaccinated only), modified (vaccinated or unvaccinated), flat (vaccinated or unvaccinated) hemorrhagic, and variola sine eruptione (vaccinated only). variola has several features that make it a potentially good bioweapon, including the ability to be made into a lyophilized form, which can be aerosolized and is heat resistant (box 2). however, the virus is easily killed, even in the lyophilized from, by ultraviolet light and disinfectants. the primary concern related to smallpox as a bioterrorism agent is that it was produced in large quantities as a bioweapon by the soviet union and the security of its viral stock remains uncertain after the collapse of the superpower. 29 mathematical models 30 suggest that even a moderate-sized attack infecting 100 to 1000 people would lead to a massive global pandemic that would require significant interventions to control, which in turn would carry with them significant economic and civic ramifications. 31 airborne and contract precautions should be used to manage patients with smallpox. clearly, in a large-scale attack or outbreak, isolation of all patients in negative pressure rooms would not be feasible. all fomites (bed linen and patient clothing) should be transported in sealed biohazard bags and autoclaved before washing or incineration. standard hospital procedures for room cleaning can be followed once a patient has been discharged from the room. viral hemorrhagic fevers viral hemorrhagic fevers (vhf) ( table 2 ) are a group of infections caused by 4 families of rna viruses (arenaviridae, bunyaviridae, filoviridae, and flaviviridae), which are believed to primarily exist within animal or arthropod reservoirs, but occasionally infect humans. 7, 19, 32 each virus is generally contained to a specific geographic region. all of the viruses have in common that they attack the vascular endothelium, leading to vascular leak and potentially producing shock or coagulopathies. with the exception of hantavirus, natural transmission is primarily by respiratory droplets and body fluids but not by airborne aerosol. aerosol transmission of hantaviruses from the urine of rodents has been well documented. most secondary (humanto-human) transmission of vhf is to family members or health care workers caring for patients with vhf. the primary mode of transmission in these cases is direct contact with infected body fluids and percutaneous exposures. however, questions still remain regarding the possibility of aerosol transmission, particularly for lassa fever or from aerosol-generating procedures. 33 several of the vhf viruses have reportedly been weaponized by the russian and the us militaries. isolation for all vhf is strict droplet and contact precautions. patients who are end stage or have significant pulmonary involvement may warrant airborne isolation in a negative pressure room, if available. for specific details regarding isolation and environmental decontamination procedures for vhf, see the cdc's interim guidance for managing patients with suspected viral hemorrhagic fever in us hospitals, available online at http://www.cdc. gov/hai/pdfs/bbp/vhfinterimguidance05_19_05.pdf (accessed january, 2013). in addition to clinical infection control precautions, it is important to minimize risk to box 2 why smallpox is a good candidate as a bioterrorism agent 1. it is transmissible by the aerosol route, both in a weaponized from and from infected to susceptible persons. 2. it is heat stable in a lyophilized form. 3. the populations of most countries contain a high proportion of susceptible persons. 4. smallpox is associated with high morbidity and about 30% mortality. 5. there is a significant psychological fear of smallpox among both the public and health care workers. 6. initial diagnosis of cases of the disease may be delayed, given that it has not been seen clinically for more than 20 years. 7. other than the vaccine, which may be effective only in the first few days after infection, there is no proven drug treatment for clinical smallpox. the laboratory and other health care workers by strictly limiting the number of samples and blood work collected. no fungal organisms are currently on the cdc list of category a or b agents, although it could be argued that some should be considered as potential category c agents. 34 coccidioidomycosis is caused by the soil-borne spore-forming organisms coccidioides immitis and coccidioides posadasii found in the americas. these organisms typically cause a self-limited respiratory illness in immunocompetent individuals. coccidioidomycosis is typically transmitted by inhalation, particularly after events such as earthquakes or dust storms, which generate airborne particulate mater from the soil. human-to-human transmission does not occur. overall, the long incubation period and mild clinical presentations in most individuals make coccidioides an unattractive agent as a bioweapon. however, given the ease of access to the organism and its natural propensity for aerosol transmission, its potential use should not be ignored. keeping an open mind about possible future bioterrorism agents is essential if we are truly going to be prepared. 10 biological toxins technically toxins are chemicals but because of their source (living organisms), biotoxins are typically classified as biological weapons. 7, 19, 24, 35 the cdc list of potential bioterrorism agents (see table 1 ) includes several toxins. in addition, there are several other toxins that also have the potential to be used as agents of bioterrorism (box 3). toxins can be deployed via aerosol dispersal devices or by contaminating food sources and have been used as bioweapons in the past ( table 4) . to be effectively deployed as an airborne agent, the toxin must be 1-3 mm for optimal aerosolization. ricin ricin is a protein toxin obtained from the organism ricinus communis (castor beans). castor beans are available worldwide and ricin is a natural by-product produced in the processing of castor beans to castor oil. the waste products resulting from the processing of castor beans contain approximately 5% ricin. 24 in its pure form, ricin is a white powder that is soluble in water and stable over a wide range of ph. chemically, ricin is a glycoprotein lectin with a and b chains joined by a disulfide bond. the b chain binds to galactose-containing proteins and lipids on cell surfaces, causing direct membrane damage and release of cytokines, whereas the a chain inhibits eukaryotic ribosomes by removal of adenine from the 28 rrna loop in the 60 subunit, therefore stopping protein synthesis, leading to cell death. ricin is less toxic by weight compared with botulinum toxin or staphylococcal enterotoxin type b (seb), but is more readily available and easy to produce in large quantities than either of these agents. ricin can be deployed as a bioweapon via several routes, including inhalation, ingestion, or injection. clearly, as a bioterrorism agent with the intent to cause mass casualties, inhalation or ingestion is more likely than injection. after exposure, decontamination should include removal of clothing and jewelry, washing with soap and water, followed by rinsing with copious amounts of water. clostridium botulinum toxin (botulism) botulinum toxin is the most poisonous substance known to man with a dose lethal to half those exposed (ld 50 ) of 1 ng/kg. clostridium botulinum is a gram-positive, spore-forming, anaerobic bacillus found in soil and water globally. rarely, the toxin is also produced by clostridium butyricum and clostridium baratii. botulinum toxin is produced in pharmaceutical grade and is commercially available for cosmetic and other medical purposes. seven types of toxins exist and are identified as types a to g, with a, b, and e causing most cases of human disease. the toxin contains 3 chains (heavy, light, nontoxic hemaglutinin). the heavy chain permits the toxin to bind with the cell, and the light chain contains zinc-dependent endopeptidase, which stops calcium release. botulinum toxin binds to peripheral presynaptic cholinergic terminals (including muscarinic and nicotinic) and blocks the calcium-dependent exocytosis process. the central nervous system is not affected by the toxin. botulism is most often considered a food-borne illness, given that most of the naturally occurring cases are the result of ingesting contaminated food products. other naturally occurring forms of toxicity are wound botulism (soil-contaminated wounds) and infantile botulism (ingestion of spores). the spores themselves are heat resistant, tolerating temperatures up to 100 c, but the toxin itself denatures and becomes inactivated with even brief exposures to temperatures higher then 85 c. use as a potential bioweapon includes contaminating food sources or via aerosolization. however, the toxin quickly degrades when exposed to environmental conditions, thus limiting its usefulness as a bioterrorism agent. botulinum toxin cannot cross the skin but can be absorbed by mucous membranes. should exposure to the toxin occur via aerosolization, decontamination should include removing clothing and jewelry, washing with soap and water, and rinsing with copious amounts of water. wash contaminated surfaces with a 0.1% hypochlorite bleach solution to destroy the toxin. micotoxins several hundred varieties of toxins are derived from fungi, with the trichothecenes and aflatoxins being most concerning. aflatoxins are produced by aspergillus flavus or aspergillus parasiticus and are common contaminants in harvested food. aflatoxins bind to and damage dna and cellular proteins. trichothecenes are produced by a large number of fungi including: fusarium, stachybotrys, trichoderma, myrothecium, and cephalosporium. the most likely toxins from the class trichothecenes to be used as bioweapons are: t2 (trichothecene) and deoxynivalenol (vomitoxin). t2 causes skin irritation/pain and can be absorbed through the skin or inhaled, and then disseminated systemically, binding to peptidyltransferase inhibiting protein synthesis. it also interferes with dna polymerase and monoamine oxidase. the organs affected first are those are with rapidly reproducing cell lines (gi tract, bone marrow, and skin) as well as impairing proteins involved in coagulation and the krebs cycle. in addition, the breakdown of serotonin, epinephrine, and norepinephrine is impaired. t2 is a yellow droplet in appearance. should exposure occur, decontamination includes removal of clothing and jewelry, scrubbing the skin with soap and water, and isolating clothing or other contaminated objects. staphylococcal enterotoxins staphylococcal enterotoxins are heat-stable toxins produced by the common staphylococcus aureus bacteria. there are more than 20 types of enterotoxins, with type a commonly known for causing food poisoning and type f causing toxic shock syndrome. the genes for staphylococcal enterotoxins are found on plasmids and bacteriophages, allowing transfer between different strains. seb can be aerosolized and has been studied as a potential incapacitating biological agent. seb is a superantigen, which produces massive stimulation of t cells and cytokine storm, which is rarely fatal but is significantly incapacitating to the victim. if there is exposure, decontamination requires removing clothing and jewelry, followed by washing skin with soap and water. clostridium perfringens epsilon toxin clostridium perfringens is an anaerobic, grampositive, spore-forming bacillus that is found in soil in all parts of the world. clostridium sp produce several toxins, one of which is the epsilon toxin, which commonly causes food poisoning. similar to seb, epsilon toxin could be another potential incapacitating bioweapon. insects were some of the earliest bioweapons ever used against an enemy (see table 4 ). 6, 22 there are 3 potential ways to use insects in bioterrorism: direct attacks, agents of agroterrorism, or as vectors of disease. direct attacks most often use stinging insects and have been used in the past to defend fortifications and rout enemies from entrenched positions (the bee hive in the log, as seen in cartoons). another insect that has been studied by the indian military as a potential bioweapon is the paederus beetle, found in the middle east, which produces the toxin pederin. pederin is potent, causing intense pain, festering lesions, and blindness if it contacts the eyes. ingestion of the beetle can be lethal, as is injecting pederin into the bloodstream. agroterrorism is "the deliberate introduction of an animal or plant disease as well as damage to crops and livestock with the goal of generating fear, causing economic losses and/or undermining social stability." 6 an example of insects used as an agent of agroterrorism is the medfly (ceratitis capitata), a fruit fly found in parts of the world but not the united states. their larvae eat many plants, causing significant crop destruction. in 1989, a group threatened to release medflies in california if the government did not stop a pesticide-spraying program. if the insects had been released, the estimated damage to the economy would have been $13.4 billion. 6 insects can be used in bioterrorism as disease vectors. table 3 lists potential vector-borne diseases of concern. epidemiology it is challenging to describe the epidemiology of agents of bioterrorism in the traditional manner for several reasons. first, as discussed in the introduction, theoretic risks are largely being discussed, with many unknown factors involved. further, given that most of the work done in the area of bioweapons has been conducted by military or state organizations, only a small percentage of the overall activities have been publically reported. to provide the best possible overall perspective on the epidemiology of bioterrorism agents, table 4 offers a comprehensive overview of past use of bioweapons as compiled from several sources. 1, 4, 6, 7, 13, 19, 22, 24, [35] [36] [37] in addition to the general risks discussed in the introduction, academic health centers potentially play a unique role in the epidemiology of bioterrorism. any public area where mass gatherings occur or any infrastructure critical for the smooth functioning of society are potential targets for bioterrorism attacks. not only do hospitals in general, and academic health centers specifically, fit this profile, but academic health centers may also be at an increased risk because they are also potential sources of agents of opportunity (ao). the term ao is used to connote the use of a routine and unregulated chemical, biological, or radiologic agent by terrorists. 36 table 3 potential bioterrorism vector-borne diseases of concern for humans vector disease two key features of an ao are (1) its availability, and (2) a practical means of dissemination. box 4 lists potential biological aos available in many academic health centers. an ao does not have to cause significant morbidity or mortality to be effective; the psychological effects on the public and impact on society/critical infrastructure functioning are sufficient to have a serious effect. aos are of significant concern to law enforcement officials because potential barriers to the development of a highly effective biological weapon are the cost and expertise required to develop them, whereas aos mitigate both of these factors. in the past, the development of bioweapons has required significant funding and labor, which typically only states can mobilize. it has been said that the us military bioweapon program employed more than 3000 staff and the ussr bioweapons sections employed more than 60,000 staff, both at costs of millions of dollars. 7 although presenting a significant obstacle in bioweapon development, cost does not always prevent nonstate actors from pursuing bioweapons; the aum shunrikyo cult may have spent more than $10 million dollars in its failed attempt to develop weapon-grade anthrax. 7 the diagnosis, or identification, of a bioterrorism event is likely the most challenging step in the response. 38 recognition of a bioterrorism event has 2 components: (1) identifying that an intentional rather then natural phenomenon has produced several cases of illness, and (2) diagnosing the specific organism or agent causing the illness. they may occur in either order depending on the circumstances. bedside clinicians, specifically critical care physicians, 39 play a crucial role in both aspects of recognizing that a bioterrorism event has occurred. 2, 39, 40 recently, a great deal of attention and money has been directed toward bioterrorism detection technologies, such as environment sampling and syndromic surveillance. however, these efforts are highly unlikely to replace the bedside clinician in recognizing the event. 2, 41 in a simulated anthrax attack designed to test the performance of a syndromic surveillance system currently in use, nordin and colleagues 41 found that performance of syndromic surveillance varies with the infection rate. based on their analysis in a metropolitan area if there was a release of anthrax at a large public venue, a syndromic surveillance system that monitored approximately 9% of the local population would detect the event most of the time if the infection rate was 20% and all of the time if the infection rate was more than 40%, but it would take 3 to 6 days in both situations. therefore, these investigators concluded that "a suspicious clinician may detect the first case of anthrax before a syndromic surveillance system sounds an alarm and public health determines it is an anthrax release." 41 identifying the important role that the beside clinician plays in recognizing a bioterrorism event is not intended to diminish or exclude the role of others in the process, especially not public health. the rapid recognition, and effective response, to a bioterrorism event requires that clinicians and public health officials work together. [42] [43] [44] in some circumstances, such as smallpox, the diagnosis of a single case triggers the alarm that a bioterrorism event has occurred. 45 however, for any of the diseases with a naturally occurring incidence, as well as with a novel organism or illness that has not yet been identified, it is the existence of a cluster of cases in a geographic area that triggers suspicion of a possible bioterrorism attack or emerging epidemic ( table 5) . recognition of a cluster of patients typically requires a level of situational awareness that is beyond what an individual clinician can acquire and is usually attainable only by a regional, state, or federal public health agency depending on how widely distributed the cases are. for example, if an anthrax attack took place in a busy international airport, the cases would be widely distributed across a nation or internationally (eg severe acute respiratory syndrome [sars]). 46 in this case, to the individual clinicians receiving the cases of inhalational anthrax, each would be an unusual, but not necessarily alarming, occurrence. however, the reporting of several cases of inhalational anthrax to a federal or international public health agency simultaneously would certainly trigger an alarm. however, we are often left with trying to distinguish between a naturally occurring event (epidemic) and an intentional event (bioterrorism). it can be difficult to distinguish between an artificial incident (terrorism) and a natural occurrence (epidemic). 47,48 box 5 lists some of the characteristics that might help christian table 5 case number triggering criteria for considering a potential bioterrorism event disease(s) a a this factor is in part based on the background incidence of disease in a geographic area and therefore varies from location to location as well as based on seasonal variations. for example, a viral encephalitis that is naturally transmitted by an endemic vector occurring in the summer when mosquitoes are prevalent would not necessarily raise concerns; however, a single case occurring in the winter or in a geographic area where that infection is not typically seen may raise concerns immediately. features that suggest a bioterrorism event rather than an epidemic 1. an epidemic curve that suggests a point source (common source) outbreak or extended source rather then a naturally propagated (transmitted) source (fig. 2) 2. identification of a cluster of cases (large numbers of patients from a similar geographic area with similar symptoms) distinguish between the 2 events based on epidemiologic features. recently, radosavljevic and belojevic 38 have developed a scoring system to help distinguish between a bioterrorism incident and an epidemic ( table 6) . these investigators use an interesting approach, which scores several qualitative and quantitative features of the incident categorized under 3 groupings: person (cases), place (spatial distribution), and time. a score of 8 or higher on this system is said to suggest that an artificial (intentional or accidental) event is more likely then a naturally occurring epidemic. the questions considered in this score have to be answered at the public health level, not the individual clinician level, because they require a level of situational awareness not attainable by a clinician in a hospital. the importance of clinicians working effectively with public health agencies is further reinforced by computer modeling work by hupert and colleagues, 49 which suggests that even minor delays in detecting (diagnosing) and initiating the response to a large-scale anthrax attack would make even perfectly executed prophylaxis campaigns ineffective in preventing the health care system from being overwhelmed by cases. as noted in the earlier discussion, identifying the disease or organism involved can be key to identifying the event as a bioterrorism incident. this section provides an overview of the clinical and investigational findings that assist in diagnosing the specific diseases. key to any biological diagnosis are laboratory, particularly microbiological, tests. 50 table 7 summarizes the appropriate specimens and tests for diagnosing potential agents of bioterrorism. however, as with many aspects in critical care, it is important not to be overly reliant on technology. the college of american pathologists regularly tests the ability of laboratories to accurately diagnose, or appropriately refer to a reference laboratory, clinical potential agents of bioterrorism. 51 in their recent testing of laboratories, the college of american pathologists found rates of acceptable identification responses were as follows: bacillus anthracis, 90% (2007) 51 although the rates of appropriate diagnosis are improving, there still remains a delay in notification of the reference laboratory and a not insignificant failure rate for diagnosing category a organisms other then anthrax. anthrax as discussed earlier, anthrax presents as 3 typically distinct clinical syndrome: cutaneous, gi, and thoracic. 7, 11, 12, 19 diagnosis of anthrax often involves a combination of clinical, radiographic, and microbiological data. the clinical presentation of the 3 anthrax syndromes is outlined in table 8 . cutaneous anthrax most often affects the hands, arms, and face. one of the main differential diagnoses for cutaneous anthrax are spider bites, and table 9 presents clues to help distinguish between the 2. the diagnosis of cutaneous anthrax can often be made clinically followed by laboratory confirmation. other laboratory and radiologic investigations are of limited use in cutaneous anthrax. conversely, in the cases of gi and thoracic anthrax, the clinical presentation is relatively nonspecific, and radiologic investigations, especially for inhalational anthrax, 52 along with laboratory and microbiological testing are essential for making the diagnosis. with thoracic anthrax, sputum cultures are rarely positive but first blood cultures are almost always positive. 7 with the relatively nonspecific findings associated with thoracic anthrax, it can be challenging for clinicians to differentiate it from the more commonly seen community-acquired pneumonia (cap). kyriacou and colleagues 53 conducted a retrospective review of cases of bioterrorism-acquired thoracic anthrax and compared them with cases of cap or influenza. these investigators found that the most accurate predictor of anthrax was mediastinal widening or pleural effusions on chest radiograph, which were 100% sensitive, 71.8% specific compared with cap, and 95.6% specific compared with influenza (see fig. 2 ). other features more common in anthrax were: nausea, vomiting, pallor, cyanosis, diaphoresis, altered level of consciousness, and increased hematocrit level. two groups 54, 55 have developed screening criteria to help diagnose anthrax after a known bioterrorism incident. however, a costeffectiveness analysis showed that although both criteria would have identified cases abbreviations: ct, computed tomography scan; cxr, chest radiograph; mri, magnetic resonance imaging. a for microbiological findings, refer to table 7 . data from refs. 7, 11, 12, 19, 52 of anthrax, the mayer criteria 55 would have screened only 4 patients (cost $1900 usd), whereas the hupert criteria 54 would have screened 273 patients ($126,025 usd). 56 although it is rare with cutaneous anthrax, any case of anthrax has the potential to develop hemorrhagic meningitis. therefore, any evidence to suggest meningitis should prompt appropriate investigations such as neuroimaging and a lumbar puncture. plague as with anthrax, the diagnosis of plague varies significantly based on the clinical syndrome as which it presents. 7, 18, 19 plague may present as any of the following clinical syndromes: bubonic plague, primary pneumonic plague, primary septicemic plague, plague meningitis, plague pharyngitis, pestis minor, and subclinical infection. the first 3 syndromes are the most common presentations and their diagnostic features are outlined in table 10 . bubonic plagues is characterized by initial lymphadenitis, most commonly inguinal, followed by the development of a systemic illness, which may include secondary plague pneumonia. primary septicemic plague is similar to bubonic plague, without the initial appearance of buboes. it remains unclear if it is the same entity with subclinical lymphadenitis or a distinct pathophysiologic process. pneumonic plague may be primary, which presents as an acute respiratory infection, or may occur as a result of hematologic spread (secondary) from bubonic or primary septicemic plague. hemoptysis is frequently seen with pneumonic plague and can help differentiate it from anthrax or tularemia. 7 general laboratory and radiologic investigations are nonspecific and the diagnosis is primarily made via microbiology testing (see table 7 ). automated culture detection systems may not detect or can misidentify yersinia pestis, and therefore, if there is a high level of suspicion for plague, the microbiology laboratory should be made aware of this when the cultures are ordered. tularemia tularemia is another disease with several distinct presentations that are variable and dependent on the mode of transmission. 7, 19 the clinical spectrum of disease includes: typhoidal, ulceroglandular, glandular, oculoglandular, oropharyngeal, and pneumonic tularemia, with the last being the most likely presentation from a bioterrorism incident. the average incubation period is 3 to 5 days, but ranges from 1 to 21 days. the pneumonic form typically presents with: fever, chills, headache, malaise, coryza, cough, chest pain, pharyngitis, abdominal pain, arthralgia, septic shock, respiratory failure, and acute respiratory distress syndrome (ards). the ulceroglandular form presents with cutaneous ulcers 0.5 to 3 cm with heaped-up edges and regional lymphadenopathy. ulceroglandular infections may progress to systemic disease similar to the pneumonic form. the typhoidal form does not involve lymphadenopathy. the typhoidal form usually involves the lung primarily, whereas the ulceroglandular form first affects the mediastinal lymph nodes, then progresses to parenchymal involvement. microbiological testing is required to make the diagnosis, although cultures are generally low yield given the difficulty growing the organism. therefore, alternative techniques for identification such as immunofluorescence and polymerase chain reaction (pcr) are preferred. 7 radiologic findings are generally nonspecific, although approximately 75% of patients with tularemia present with bronchopneumonia, typically bilateral, and may develop cavities. only approximately 33% have lymphadenopathy and pleural effusions, which is significantly lower than would be expected with thoracic anthrax. 52 q fever q fever, when seen in naturally occurring infections in humans, causes a spectrum of disease from asymptomatic fevers and fevers of unknown origin through to life-threatening infections, typically associated with endocarditis or pneumonia. 23, 24 the primary concern in a bioterrorism setting is inhaled coxiella burnetii, which presents with fever and cough alone in cases of lower inoculation through to severe pneumonia and respiratory distress with high inoculums. the diagnosis of q fever requires microbiological testing. the chest radiograph and computed tomography (ct) finding are nonspecific and consistent with any other type of pneumonia. 52 there is no significant adenopathy seen on the chest radiograph or ct, which may help to rule out anthrax. other bacteria the 2 other bacterial agents that are discussed are brucella sp and burkholderia sp. 24 brucellosis has multiple clinical presentations and can be difficult to diagnose. typically brucellosis infects organs (lung, liver, spleen), bone marrow, bones, or central nervous system and is an insidious infection, presenting primarily with constitutional symptoms and organomegaly. the incubation period ranges from 2 to 8 weeks. diagnosis of brucellosis is best made by bone marrow cultures or blood cultures, although they are less sensitive. burkholderia mallei causes the clinical disease glanders, which presents as skin nodules and lymphadenopathy after inoculation through the skin, but in a bioterrorism event with inhalation exposure, the most likely presentation would be a nonspecific systemic febrile illness, which may or may not be accompanied by a pustular rash. burkholderia pseudomallei causes the clinical syndrome of melioidosis, a disease with a broad spectrum of presentations, including skin ulcers, pneumonia, acute fulminate sepsis, and chronic abscesses. if used in a bioterrorism setting with inhalational exposure, the acute presentation would most likely be either pneumonia or sepsis. smallpox smallpox can present as 1 of 5 clinical syndromes (table 11) : classic, modified, flat, hemorrhagic, and variola sine eruptione. 7, 19, 28 the primary challenge in the diagnosis of smallpox is the lack of clinicians who have experience with the disease. the cdc has published a case definition to aid in the diagnosis of cases (box 6). classic smallpox begins with a prodrome of fever and constitutional symptoms. after 2 to 4 days, the rash begins to emerge, initially as small red spots in the mouth and throat, which develop into sores that erupt and discharge virus, making the patient highly infectious at this point. at about the same time that the sores in the mouth begin to open, a rash develops on the skin. the rash is said to spread in a centrifugal pattern, starting on the face, then spreading to the extremities. the lesions progress from macules, to papules (day 2 of rash), umbilicated vesicles (day [4] [5] , and pustules (day 7) before crusting over (fig. 3) . in contrast to chickenpox (varicella), in smallpox, all the lesions progress at the same stage and the lesions are more densely concentrated on the face and extremities (figs. 4 and 5) . a single suspected case of smallpox is a public health emergency and the local public health agency should be notified immediately if there is a suspected case. imaging does not play a major role, because the respiratory symptoms tend to develop after the skin lesions. a few individuals who care for smallpox victims, or who have been vaccinated, develop a pulmonary form of smallpox without skin lesions. in these cases, chest radiography or ct may be helpful. the chest radiograph shows ill-defined nodular opacities in the upper lung fields, which may persist for months before calcifying. 52 vhf clinical presentation of vhf depends on several factors, including specific virus, virulence of the strain, route of exposure, dose, and host factors (table 12) . 32,57 the general feature that all vhf have in common is that they cause microvascular damage, leading to symptoms associated with increased vascular permeability, potentially resulting in hypovolemic shock or frank hemorrhage. medical imaging is not helpful in diagnosing these conditions, and the general laboratory findings are nonspecific, but often reveal evidence of hepatitis and hypovolemia. the specific diagnosis depends on microbiological testing (see table 7 ). the diagnosis of a bioterrorism event secondary to the use of biological toxins depends on the presentation of unusual clusters of cases. identifying the specific toxin can be challenging but the clinical presentation of some toxidromes can be useful in suggesting the agent involved. the clinical syndromes associated with ricin poisoning vary based on the mode of entry into the body. inhalation of ricin leads to symptoms within hours, which include dyspnea, fever, cough, pulmonary edema, and chest tightness, whereas ingestion of ricin results in vomiting and diarrhea, leading to dehydration, shock, hallucinations, seizures, and hematuria. the typical onset of symptoms after ricin poisoning averages from 4 to 6 hours up to 10 hours. laboratory findings associated with ricin poisoning are nonspecific but include metabolic acidosis, increased results for liver function tests, anemia, increased creatinine level, leukocytosis, and hematuria. aerosolized botulinum toxin, likely to be experienced in a bioterrorism setting, produces disease that mimics food-borne, infant, and wound botulism clinically. the clinical features that are anticipated with the inhalation of botulinum toxin include descending paralysis starting with diplopia, ptosis, fixed dilated pupils, dysphagia, respiratory failure, urinary retention, and constipation. with the ingestion of botulinum toxin, the clinical features begin with gi symptoms, including nausea, vomiting, diarrhea, bloating, and pain, before the development of the paralysis. onset of symptoms after exposure to botulinum toxin occurs as early as 6 hours after inhalation, but if ingested typically there is a 12-hour to 36-hour delay before symptoms begin. however, the onset may be delayed as much as 10 days. the diagnosis of botulinum intoxication is based on laboratory detection of the toxin in blood, food, or stool. the clinical presentations of mycotoxins vary based on the specific toxin. aflatoxin ingestion produces clinical features acutely, including hemorrhagic liver necrosis and pulmonary edema with chronic features, including the development of hepatic cancer. the effects and clinical presentation of inhaled aflatoxin are not known. t2 can be absorbed through the skin or inhaled. the clinical presentation in the acute phase includes pain on contact with skin or eyes, conjunctivitis, blurred vision, rhinorrhea, epistaxis, dyspnea, wheezing, tachycardia, shock, vomiting, diarrhea, erythema, and blistering. delayed symptoms present approximately 1 week after exposure and include thrombocytopenia, neutropenia, anemia, and coagulopathy. seb is a superantigen, which causes massive stimulation of t cells and cytokine storm, which produces the associated clinical features. ocular exposure results in conjunctivitis and eye swelling, whereas inhalation causes fever, chest pain, gi symptoms, pulmonary edema/ards, and shock. the respiratory effects may not be seen for up to 48 hours after exposure. as with most other infectious diseases, considerations for the management of bioweapons should include consideration of (1) vaccination to prevent infection/illness, (2) treatment of infection/illness, and (3) prophylaxis after exposure to prevent clinical infection/illness. also, as is common with most other infectious diseases, there are often several antimicrobial regimens that can effectively treat a specific bioterrorism agent, and for others there are no effective treatments. table 13 presents the generally accepted international recommendations for vaccination, treatment, and prophylaxis of bioterrorism agents. however, clinicians should always consult the most recent guidelines published by their public health authority when making clinical management decisions regarding bioterrorism agents. in addition to the current therapies, new drugs and vaccines are being developed specifically to address the threat of bioterrorism. 58 with the exception of possibly cutaneous anthrax, there are no specific indications for surgical management of any bioterrorism organisms. 11 even in the case of anthrax, surgical debridement has primarily been used for injection anthrax. 59,60 supportive care for victims of a bioterrorism event involves 2 components: the care of the individual patient and the response to a mass casualty event. despite the heightened awareness of the potential for bioterrorist events since 2001, many hospitals remain unprepared for biological threats. a study of uk emergency departments (eds) revealed that 24% did not have isolation facilities, and only 61% had departments with independent ventilations systems that would allow for the department to be isolated from the rest of the hospital. 61 in addition, the survey found that isolation procedures in many eds were poor; for example, 27% would not have isolated a patient with potential sars and 23% would not isolate a patient with chickenpox. 61 the task force for emergency mass critical care provides guidelines for hospitals to prepare for and manage mass critically ill casualties from events such as a bioterrorism attack. 62 critical care physicians should be familiar with these guidelines and prepared to respond in such circumstances, because if they are not, all of their skills and knowledge for treating individual patients are of little use as the system becomes overwhelmed. supportive care of individual patients can generally be grouped into categories based on the class of agent involved: bacterial, viral, or toxin. bacterial and viral agents typically produce a sepsis syndrome with vascular leak, particularly in the case of vhf, as well as variable degrees of a systemic inflammatory response. supportive management is similar to the standard best practices for managing septic patients (box 7). particular attention needs to be directed toward ventilatory support, with lung protective ventilatory strategies to minimize the development or exacerbation of adult respiratory distress syndrome. significant fluid shifts are to be expected given the vascular leak, requiring judicious volume resuscitation and cardiovascular support with vasopressors and inotropes based on goal-directed therapy. supportive therapy for patients involved in bioterrorism attacks with biotoxins varies based on the specific toxin. in the case of exposure to botulinum toxin, the primary support required is mechanical ventilation, hydration, and nutritional support until the paralysis resolves. toxins such as seb produce a response similar to sepsis and are supported by the sepsis protocols described earlier. ricin and mycotoxins both act at the cellular level, and although supportive therapies should be attempted, they may be of little benefit in altering the outcomes for patients. for mycotoxins specifically, there may be a role for the addition of steroids as a component of supportive care. 35 data regarding patient outcomes after exposure to bioweapons or bioterrorism agents are, for the most part, lacking, given the paucity of cases and that most research has been carried out by government or military organizations and the results cannot be shared publically. table 14 summarizes the published outcomes after exposure to bioterrorism agents when data are available. however, the outcomes from previous incidents, and in particular from naturally occurring illness, may not predict future outcomes, because bioterrorism agents may undergo genetic manipulation to increase their virulence or introduce antibiotic resistance, which decreases the efficacy of treatments. although to the average clinician, bioterrorism may seem to be only a remote possibility, it is not only a reality of the times in which we live but also bioweapons have been used for centuries. critical care physicians play a critical role in the response to a bioterrorism attack, and even more importantly, in identifying that an attack has occurred in the first place. an effective response to a bioterrorism incident requires a coordinated effort between clinicians and public health. critical care clinicians must be familiar with the diagnosis and management of the most likely bioterrorism agents, and also be adequately prepared to manage a mass casualty situation. index case of fatal inhalational anthrax due to bioterrorism in the united states the anthrax attacks 10 years later biological weapons control. prospects and implications for the future potential bio-terror agents entomological terrorism: a tactic in asymmetrical warfare bioterrorism: preparing for the impossible or the improbable the threat of bioterrorism anthrax as a biological weapon, 2002: updated recommendations for management bioterrorism-preparing to fight the next war anthrax infection an overview of anthrax infection including the recently identified form of disease in injection drug users the sverdlovsk anthrax outbreak of 1979 bioterrorism-related inhalational anthrax: the first 10 cases reported in the united states a biological decontamination process for small, privately owned buildings comparative sporicidal effects of disinfectants after release of a biological agent decontamination after a release of b. anthracis spores threats in bioterrorism. i: cdc category a agents tularemia as a biological weapon: medical and public health management tularemia: the disease and the weapon insects as weapons of war, terror, and torture q fever: the neglected biothreat agent threats in bioterrorism. ii: cdc category b and c agents pathogenic rickettsiae as bioterrorism agents pathogenesis of burkholderia pseudomallei and burkholderia mallei rethinking smallpox an overview on the use of a viral pathogen as a bioterrorism agent: why smallpox? modeling potential responses to smallpox as a bioterrorist weapon the economic impact of a bioterrorist attack: are prevention and postattack intervention programs justifiable? viruses of the bunya-and togaviridae families: potential as bioterrorism agents and means of control interim guidance for managing patients with suspected viral hemorrhagic fever in us hospitals coccidioides species as potential agents of bioterrorism bioterrorism: toxins as weapons developing a consensus framework and risk profile for agents of opportunity in academic medical centers: implications for public health preparedness the history and threat of biological warfare and terrorism unusual epidemic events: a new method of early orientation and differentiation between natural and deliberate epidemics bioterrorism preparedness and response: clinicians and public health agencies as essential partners biologic and chemical weapons of mass destruction simulated anthrax attacks and syndromic surveillance late recognition of sars in nosocomial outbreak critical care and biological disasters: lessons learned from sars and pandemic influenza planning definition and functions of health unified command and emergency operations centers for large-scale bioevent disasters within the existing ics medical management of the suspected victim of bioterrorism: an algorithmic approach to the undifferentiated patient severe acute respiratory syndrome nato handbook on the medical aspects of nbc defensive operations bioterrorism and critical care predicting hospital surge after a largescale anthrax attack: a model-based analysis of cdc's cities readiness initiative prophylaxis recommendations diagnostic analyses of biological agent-caused syndromes: laboratory and technical assistance a review of sentinel laboratory performance: identification and notification of bioterrorism agents radiologic manifestations of potential bioterrorist agents of infection clinical predictors of bioterrorismrelated inhalational anthrax accuracy of screening for inhalational anthrax after a bioterrorist attack inhalational anthrax due to bioterrorism: would current centers for disease control and prevention guidelines have identified the 11 patients with inhalational anthrax from screening for inhalational anthrax due to bioterrorism: evaluating proposed screening protocols the office of the surgeon general at tmm publications state-of-the-art therapeutic medical countermeasures for viral threat agents injection anthrax causing compartment syndrome and necrotising fasciitis the surgical management of injectional anthrax emergency departments (eds) in the united kingdom (uk) are not prepared for emerging biological threats and bioterrorism summary of suggestions from the task force for mass critical care summit prevention of laboratory-acquired brucellosis comparison of efficacy of ciprofloxacin and doxycycline against experimental melioidosis and glanders key: cord-348106-agwdmtug authors: shankar, venkatesh; kushwaha, tarun title: omnichannel marketing: are cross-channel effects symmetric? date: 2020-09-07 journal: nan doi: 10.1016/j.ijresmar.2020.09.001 sha: doc_id: 348106 cord_uid: agwdmtug the rapid growth in omnichannel (e.g., web, call center, sales agent, store) shopping and the need to effectively allocate resources across channels are prompting managers and researchers to better understand cross-channel effects, that is, the effects of marketing efforts in one distribution channel on shopping outcomes in other channels. we develop a broad set of hypotheses about cross-channel effects based on channel richness and influence roles (informative, persuasive). to test the hypotheses, we model the effects (own and cross) of channel marketing efforts on shopping outcomes in different channels through a simultaneous equation system. we estimate these models using data from the auto insurance industry that comprises the exclusive agent, the independent agent, the web, and the call center channels. our results offer novel insights. they show that cross-channel effects and elasticities are significant and asymmetric. while the effect of marketing efforts in a channel on shopping outcomes in a dissimilar (with a different primary influence role) channel is positive (e.g., exclusive agent, the web, and the call center channels are complementary), the magnitudes of the cross-channel effects are asymmetric. similarly, while the effect of marketing efforts in a channel on shopping outcomes in a similar (with the same primary influence role) channel is negative (e.g., independent agent and exclusive agent channels are substitutional), they are also asymmetric. exclusive agent efforts have a greater negative effect on the outcomes of independent agent efforts than vice versa. based on the results, we develop a channel influence vs. influenceability analysis tool for managers to better plan their channel efforts. we also illustrate a resource allocation model that shows substantial incremental profits from the reallocation of marketing efforts based on our model with cross-channel effects relative to a model without cross-channel effects. omnichannel marketing-the practice of simultaneously offering shoppers 1 information, products, services, and support through two or more synchronized distribution channels in a seamless manner-is continuing to grow at a phenomenal rate (verhoef, kannan, and inman, 2015) . an overwhelming majority of consumers shop in more than one channel (sopadjieva, dholakia, and benjamin, 2017) , and these consumers are more profitable than others (montaguti, neslin, and valentini, 2016) . different distribution channels include the web, call center, and direct sales force (neslin and shankar, 2009) . omnichannel marketing offers organizations greater opportunities to interact with customers through different channels and efficiently use all the channels (kauferle and reinartz, 2015; kumar and venkatesan, 2005; montoya-weiss, voss, and grewal, 2003) . omnichannel marketing typically goes beyond multichannel marketing to include coordination of the distribution channels that offer a seamless customer experience. as channels evolve, cross-channel effects are becoming a significant imperative for omnichannel marketers (mark et al., 2019; watson, storm, palmatier, and ganesan, 2015) . cross-channel effects refer to the effects of marketing efforts in one distribution channel or format on shopping outcomes in other channels (gauri, 2013; talukdar and gauri, 2011; the dma, 2014) . the rapid growth in e-commerce and multichannel shopping is significantly altering shopping behavior (nielsen, 2020) and prompting managers and researchers to better understand cross-channel effects, to effectively allocate resources across distribution channels, managers need to better understand the effects of efforts in different channels on shopping outcomes (e.g., inquiries, quotes, orders, purchases) in other channels. cross-channel effects are important in several industries with multiple distribution channels. for example, in the auto insurance industry, state farm, like other leading brands, uses the 1 for expositional use, we use the terms, shopper, consumer, and customer interchangeably throughout the paper. complementary, substitutional, or insignificant in nature. two channels are complementary (substitutional) if efforts in one channel positively (negatively) influence purchases in another channel. in addition, some cross-channel effects can be high or low. a better understanding of the nature, direction, and magnitude of cross-channel effects is critical for researchers and managers from theoretical and practical standpoints, respectively. to deeply understand cross-channel effects, many firms do not have data on each customer's use of different channels. although more data on online customer journey are available through advanced google and adobe analytics, much data on offline customers' activities are still private and proprietary. furthermore, privacy regulations, such as general data protection regulation (gdpr) and california consumer privacy act, make it harder to track every customer's data. thus, many firms cannot always fully track each customer's exact journey (e.g., whether she first visited the web, called the firm, or contacted an agent). however, firms typically have access to channel level data on variables such as channel efforts and aggregate shopping outcomes in each channel. therefore, an important question is: how can managers use channel level aggregate data to estimate and infer cross-channel effects? we address the research questions by first developing a broad set of hypotheses about crosschannel effects. we then formulate a model of cross-channel effects. we estimate the model and the cross-channel effects using channel level aggregate data from the auto insurance industry. finally, we discuss the results, derive useful theoretical and managerial implications, and illustrate a resource allocation exercise based on our model of cross-channel effects. our results offer novel insights, showing cross-channel effects and elasticities are significant and asymmetric. while the effect of marketing efforts in a channel on shopping outcomes in a dissimilar (with a different primary influence role and richness) channel is positive (e.g., j o u r n a l p r e -p r o o f journal pre-proof exclusive agent, the web, and the call center channels are complementary), the directions and extent of cross-channel effects are asymmetric. similarly, while the effect of marketing efforts in a channel on outcomes in a similar (with the same primary influence role) channel is negative (e.g., independent and exclusive agents are substitutional), it can be asymmetric. exclusive agent efforts have a greater negative effect on independent agents' outcomes than vice versa. our research makes important contributions to the omnichannel management literature. first, it offers an approach for researchers and practitioners to develop and estimate a cross-channel effects model using only channel level aggregate data that are more commonly available. second, it develops broad hypotheses about cross-channel effects based on channel richness and influence roles (informative vs. persuasive), leading to complementary and substitutional crosseffects. third, it produces several interesting substantive insights regarding asymmetric crosschannel effects. finally, it offers a useful managerial tool of influence vs. influenceability analysis and a resource allocation model to better allocate resources across channels. the web channel has significantly reshaped shopping behavior in all the channels (kannan and li, 2017; peterson et al., 1997; verhoef et al., 2007) . studies on the addition of web to existing channels show little or no cannibalization between the channels. biyalogorsky and naik (2003) show that online activities do not significantly cannibalize offline sales for compact discs. pozzi research on the introduction of physical stores on sales in existing channels offers mixed results. avery et al. (2012) find that when an online retailer adds a physical store, it temporarily cannibalizes the catalog channel, but not the web, and over the long-run, enhances sales in both the catalog and the web channels. similarly, pauwels and neslin (2015) show that the introduction of physical stores of a retail firm dampens catalog sales without affecting web sales, resulting in a net increase in shopping frequency and revenues across channels. wang and goldfarb (2017) report that the effect of a new physical store is associated with a decline (spike) in online sales in places where the retailer has a strong (weak) presence. bell, gallino, and moreno (2018) find that eyewear showroom openings increased online and overall sales. fisher, gallino, and xu (2019) report that the opening of a distribution center to reduce delivery time enhanced online sales and spilled over to offline sales. some studies suggest that interactions among channels depend on product type. analyzing data on electronic and physical channels for products of certain quality (new cars) and uncertain quality (used cars), overby and jap (2009) find that more transactions involving low (high) quality uncertainty occur in the electronic (physical) channels, suggesting little cross-channel effect for the same product type. brynjolfsson et al. (2009) show that the internet competes fiercely with brick-and-mortar channel for mainstream products, but the effects could be complementary across the internet, catalog, and store channels for niche products. a few studies analyze the effects of shopping in a channel on shopping in other channels. based on survey responses in six product categories, verhoef et al. (2007) find significant cross-j o u r n a l p r e -p r o o f journal pre-proof channel effects within and across shopping tasks such as search. teerling (2007) also concludes that the role of informational websites on cross-channel shopping behavior may be significant. while these studies examine the effects of shopping across channels, our focus is on the effects of the firm's efforts in each channel on outcomes in the other channels. a related research stream focuses on the drivers and consequences of channel choice, switching, use, and offerings, without explicitly addressing cross-channel effects. a firm's channel efforts and customers' channel experience play significant roles in determining customers' channel selection (ansari et al., 2008) . venkatesan et al. (2007) study the timing of adoption of a new channel by a customer and find differences between the drivers of adoption of second and third channels. the results from these studies suggest that customers' channel choice, channel switching behavior, and price effects are fairly complex and depend on factors such as product category and customer shopping traits. our research extends these studies by examining the effects of the firm's marketing efforts in each channel on the outcomes in the other channels. another less directly related research stream is on communication channels that play the informative influence role, but our work differs from this stream in important ways. de haan et al. (2015) , dinner et al. (2014) , li and kannan (2014) , and naik and peters (2009) analyze the effects of efforts in communication or media channels in the earlier stages of the purchase funnel (e.g., search advertising, display/banner advertising) on outcomes in the later stages of the purchase funnel. similarly, wiesel et al. (2011) examine the effects of communication or media efforts (e.g., email, adwords, fax, flyer). dinner et al. (2014) estimate the effects of advertising spending in traditional, online display, and online search communication channels on sales in j o u r n a l p r e -p r o o f journal pre-proof online and offline channels. mark et al. (2019) find significant effects of catalogs on web and store purchases. in contrast, we examine the effects of each distribution channel's (e.g., exclusive agent, web, call center) efforts on shopping outcomes in other distribution channels in advertising's presence. thus, our research complements this research stream. these articles focus on the impact of the presence or addition of one channel on the existing channel(s), the drivers of channel choice, or communication channels. however, most firms have multiple distribution channels and seek to understand the effects of marketing efforts in one distribution channel on the outcomes in other channels. moreover, prior studies do not study cross-channel effects by controlling for the effect of advertising that is influential in many industries such as the insurance industry (guitart and hevet, 2017) . furthermore, these studies are mostly about retailers, precluding generalization to service providers, who outnumber retailers. finally, prior studies do not focus on a managerial cross-channel effort guidance tool. our study focuses on the cross-channel effects of marketing efforts in each channel on outcomes in other distribution channels after controlling for advertising effects in the same framework. our empirical application is in the services context and offers a new managerial cross-channel influence-influenceability tool (a channel management tool) and resource allocation insights. table 1 presents a review of selected relevant studies in cross-channel effects, including our research. some research relating to cross-channel effects focuses on firm outcomes when a new purchase channel is added to existing purchase channels. some studies are about channel choice, while others are on communication channels, which may also play informative roles similar to distribution channels. 2 some studies (e.g., ansari et al. 2008 ) do include email messages. 2 we recognize that research on the 'news' product category has examined cross-effects across print and online formats (e.g., chen, ho, and smith 2019; kanuri, mantrala, and thorsen 2017) . these formats are similar in spirit to channels but fundamentally differ from channels in that shoppers consume the product or service in these channels. entails convincing prospects through a rational appeal, an emotional appeal, or both. these roles add to the richness of each channel. but each channel has a primary influence role that can help us understand and predict the effects of marketing efforts in that channel on shopping outcomes in the other channels. relative to informativeness, persuasiveness differs more across channels, making it the key differentiator among channels in their impact on shopping outcomes. among the channels, the catalog channel and web channel primarily serve to inform shoppers. these leaner channels primarily allow for presentation of factual information. the web is richer than the catalog in that it allows retargeting, so it is more persuasive. mobile app, smart speaker, and kiosk are more informative as well as persuasive than the web and the catalog. the call center is richer than the web, catalog, mobile app, smart speaker, kiosk channels. it can provide information but only upon queries from the shopper, making its informative role specific. at the same time, it is interactive, so it can also serve to persuade shoppers. thus, the call center typically offers a balance of information and persuasion or moderate richness, so we call it a balanced channel. independent stores and independent agents are richer than the balanced channel. exclusive stores, exclusive agents, value added resellers, and manufacturers' representatives are also richer than the call center channel. these channels interact with shoppers more than do the other channels. during the interactions, these channels can use rational and emotional appeals to persuade prospects. thus, these channels' primary influence role is persuasion. their role goes beyond the call center's persuasive role because these channels allow face-to-face and long interactions with the shopper. a graph depicting the roles, leanness, and richness of the channels relative to one another appears in figure 1 . (avery et al., 2012; kushwaha and shankar, 2013) . for example, the web, a primarily informative channel, offers instantaneous access to product information, the ability to search, sort, and compare products, and a low pressure sales environment (balasubramanian et al., 2005) . the electronic channel offers the benefits of convenience and efficacy of information acquisition (choudhury and karahanna, 2008) . however, it is still lean in that it does not allow physical inspection, demonstration, trial, instantaneous acquisition, or instant gratification. in contrast, the call center channel, with a mix of information and persuasion, is richer, offering personal interactions, authentic answers to questions, and ubiquitous access. at the same time, it does not permit visual inspection and product and price comparisons. exclusive sales agents represent a richer channel and are primarily persuasive. they are highly knowledgeable advocates and credible stewards of the firm's products. however, prospective customers may not regard the information obtained from exclusive agents to be entirely objective and may be unable to gather accurate information about competitor brands (zweifel and ghermi, 1990) . independent sales agents, who constitute another rich and primarily persuasive channel, offer brand variety and are more unbiased than exclusive agents about the focal firm. however, prospective customers may not expect to get the most accurate information about all the competing brands and may feel that they would be unable to obtain a better deal for a brand than they could get from an exclusive agent for that brand (zweifel and ghermi, 1990 ). in addition, customers may perceive both exclusive and independent agents as pushy. from the firm's viewpoint, these distribution channels are direct or indirect based on the nature of contact and interaction with customers (sa vinhas and anderson, 2005). the web and the call center are direct channels, allowing the firm to directly interact with customers. in j o u r n a l p r e -p r o o f contrast, the exclusive agent and the independent agent channels are indirect channels as the firm relies on these intermediaries to communicate with the customers. this classification loosely maps with the categorization of informative vs. persuasive (narayanan et al., 2005) . firms use direct channels mainly to exercise an informative role and to some extent a persuasive role. in contrast, firms typically utilize indirect channels to implement a persuasive role. another way to classify the channels is as owned versus independent, consistent with dual distribution--a hybrid of vertically integrated and market governance channels (e.g., srinivasan, 2006) . viewed from this perspective, all the channels in our channel except the independent agent channel are owned channels. the main difference between these two types of channels lies in the extent of the firm's control over marketing efforts. firms can exercise greater control in their owned channels than in the independent channels. this classification corresponds with our categorization of informative versus persuasive channels in the following way. firms can use their owned channels to perform either an informative role, a persuasive role, or both. but firms typically utilize independent channels to persuade customers. the classification of channels in terms of their primary influence role as informative or persuasive is consistent with channel or information richness theory, which is quite fundamental, requiring fewer assumptions. therefore, we adopt this classification to develop the hypotheses. to understand the effects of marketing efforts in one channel on shopping outcomes in other channels, we analyze the relative richness and the primary influence roles of the channels. in cases where the channels are dissimilar in richness and if the primary influence roles of two channels are complementary, the benefits of one channel spill over to another channel, producing complementary cross-channel effects. in cases where the richness and primary influence roles of the channels are similar, the limitations of one channel extend to other channels, creating j o u r n a l p r e -p r o o f substitutional cross-channel effects. still in other cases, efforts in any one channel may have only stand-alone effects and not affect outcomes in other channels. by combining their richness and primary influence roles, some channels may complement one another. for example, efforts in the lean, primarily informative web channel can enhance outcomes in other channels such as the call center that erves both informative and persuasive roles. as discussed earlier, the addition of the web does not cannibalize sales in other channels (biyalogorsky and naik, 2003) as it acts an al informative channel for shoppers who may prefer to shop in rich, persuasive channels. furthermore, trusted websites can direct prospects to make purchases from a more persuasive channel like the call center (teerling, 2007) . similarly, call center agents can provide a high level of information and personal service to prospective customers, combining both informative and persuasive roles, empowering shoppers to complete shopping task from a primarily informative channel such as the web or the catalog. a call center agent can offer a richer human touch to a prospective customer that could increase the likelihood of buying from a primarily informative channel. by personally answering customers' questions, call center agents can enhance customer trust to effect favorable outcomes in informative channels. shoppers can make decisions in an informative channel based on the intent developed from their interactions with call center agents, consistent with the theory of reasoned action (fishbein, 1980) . after a shopper receives persuasive communication from the balanced channel, she evaluates the arguments over time. if the shopper is convinced, she could minimize the time and effort expended in making her decision by directly going to the lean, primarily informative channel (the web) and practice self-service. if she goes back to the j o u r n a l p r e -p r o o f balanced channel, she would have to wait for an agent to connect and go through social protocols before being able to make her decision. these arguments lead to the following hypotheses. a lean and primarily informative channel such as the web or the catalog can combine with a rich and primarily persuasive channel such as a brick-and-mortar store or a sales agent to positively influence shopping outcomes in each other's channel. marketing efforts in a primarily informative channel can direct shoppers to a primarily persuasive channel to complete their task. the phenomenon of research shopping in which a shopper searches for product information on the web and buys the product at the store is well documented (verhoef et al., 2007) . indeed, efforts in an online (primarily informative) channel complements outcomes in the direct sales force (primarily persuasive) channel in b2b markets (lawrence et al., 2019) . thus, we expect the effect of marketing efforts in a primarily informative channels like the web to positively influence outcomes in primarily persuasive channels such as exclusive and independent agents. a rich and primarily persuasive channel such as the store or the exclusive agent channel can also facilitate positive shopping outcomes in a lean and primarily informative channel. through extra efforts to accept and handle product returns, stores can assure customers to purchase more on the web (mahar et al., 2014) . furthermore, in banking, complementary marketing efforts in more persuasive channels such as branches facilitate customer adoption of the web channel (campbell and frei, 2010) . indeed, customers who adopt the outlet store channel increase their purchases on channels such as the web and the catalog (soysal and krishnamurthi, 2016) . similarly, exclusive agents act as strong advocates for the company's brand. although these agents expend their efforts to consummate shopping outcomes in their own channel, their overall j o u r n a l p r e -p r o o f goal is to maximize the probability of prospects buying from the company even if it is in a primarily informative channel. the role of the exclusive and independent agents is similar to that of the stores, branches, and outlet stores in that both these channels perform a persuasive role focused on the company's offerings. this reasoning leads to the next hypotheses. h 2a : marketing efforts in a primarily informative channel has a positive effect on the shopping outcome in a primarily persuasive channel. h 2b : marketing efforts in a primarily persuasive channel has a positive effect on the shopping outcome in a primarily informative channel. we expect channels with a balanced mix of information and persuasion such as the call center to have complementary cross-channel effects on outcomes in richer, primarily persuasive channels. shoppers touched by call center agents may be well informed about the focal firm's products. if the call center agent provides a high level of service, shoppers may even be inclined toward the firm's products. such shoppers can consummate their shopping tasks in a richer, primarily persuasive channel like exclusive agent or independent agent. by the same token, efforts in a richer primarily persuasive channel such as the exclusive agent or the independent agent should also facilitate outcomes in a leaner primarily balanced channel. exclusive agents and independent agents can move shoppers toward a strong purchase intent for the focal brand. such shoppers may call the call center and resolve their final questions in an interactive manner and complete their shopping task, consistent with the theory of reasoned action (fishbein, 1980) . thus, the efforts of primarily persuasive channels can complement outcomes in balanced channels. these arguments lead to the following hypotheses. in contrast, a rich, primarily informative channel when combined with another rich, primarily informative channel may not enhance the effect on the shopping outcomes in the other channel. shoppers may perceive the marketing efforts in each informative channel as substitutes for obtaining similar information. similarly, a rich, primarily persuasive channel when combined with another rich, primarily persuasive channel may not have a positive effect on the shopping outcomes in the other channel. in fact, two persuasive channels may be perceived as perfect substitutes by shoppers and may compete with each other for shopper attention. channels that are primarily lean, informative (rich, persuasive) may act as substitutes for other primarily informative (persuasive) channels. prior research offers some evidence for substitution (e.g., gong et al., 2015) . customer adoption of the primarily informative web channel in banking is associated with substitution from other primarily informative channels such as self-service channels (automated teller machines, voice response systems) (campbell and frei, 2010) . even in cases where a primarily persuasive channel opens in a location containing another primarily persuasive channel, the cross-channel effect may be substitutional. indeed, the growth in web channel sales has come at the expense of the catalog channel, another primarily informative channel (avery et al., 2012; pauwels and neslin, 2015) . the channels in these cases are similar in their benefits and limitations, so they serve as substitutes. the same logic may extend to the effects of marketing efforts in rich and primarily persuasive channels on outcomes in channels with the same primary influence role. efforts in similar channels can lead to customer perceiving the channels as substitutes because the net benefits in those channels may be similar. consequently, efforts in one channel may help outcomes in that channel but hurt outcomes in similar channels. although exclusive and independent agents serve as two primarily persuasive channels in many industries, each channel j o u r n a l p r e -p r o o f may independently promote face-to-face human persuasion (shrum et al., 2012) . consequently, customers perceive these channels as similar in net benefits and treat them as substitutes. these channels typically compete with each other to get a share of the customers' business. for example, in the auto insurance industry, the compensation of agents is commission based and both exclusive agents and independent agents do not have the incentive to refer a customer to other agents. therefore, efforts by exclusive agents may hinder shopping outcomes through independent agents and vice versa. the competition between similar channels could also lead to potential conflict (sa vinhas and anderson, 2005) . thus, we predict that these two channels may be more substitutional than complementary. therefore, we propose the following hypothesis. h 4 : marketing efforts in a channel have a negative effect on the shopping outcome in a similar (with the same primary influence role) channel. a complete list of all the predicted cross-channel effects appears in table 2 , which shows the directional effects of efforts in each channel on outcomes in other channels. we expect all the complementary (primarily informative, primarily persuasive, balanced) cross-channel effects, except those between the exclusive agent and independent channels, to be positive. because two similar (primarily informative or primarily persuasive) channels often compete for the same business, the efforts of one of these channels will likely negatively influence outcomes in the other channel. the extent to which efforts in each channel will influence outcomes in the other channels is an empirical question we will address in the subsequent sections. < table 2 about here > to test our hypotheses, we analyze data from the automobile insurance industry. a representation of how cross-channel effects occur in this industry appears in figure 2 . in this context, consumers can ask for an auto insurance brand quote from one of four channels, the exclusive agent, the independent agent, the web, and the call center channels. a consumer can get information from one channel but obtain a quote or buy a policy from another channel. thus, the informative and persuasive roles of the distribution channels are important in the channel outcomes. exclusive agents sell only the focal firm's products. in contrast, independent agents sell products from multiple competitors. in each channel, the key outcome variable is the number of quotes, a big milestone and a forerunner of policies. unlike in other product categories, a prospect in the auto insurance category needs to provide detailed and confidential personal information to elicit a quote. therefore, eliciting a quote is a strong signal of purchase intent, and quotes correlate highly with purchases. 4 quotes and policies are equivalent to sales leads and sales, respectively in most industries. marketing efforts in each channel impact the outcomes in the same channel as well as in the other channels. advertising also affects shopping outcomes. all the marketing channel and advertising efforts are mainly directed at increasing quotes. < figure 2 about here > we collect data on quotes, policies, and channel efforts for the focal brand, one of the largest brands in the industry, for 182 weeks. because channel effects occur in the presence of advertising, we need to control advertising effects. therefore, we gathered weekly advertising data for the focal brand from adtracker during the same time frame. 5 the advantages of our dataset are that it offers rich industry-specific information and captures the shopping outcomes in each channel. the operationalization of the variables in the data appears in table 3 . 4 indeed, the correlation between quotes and policies is rather high in our data (0.91). 5 we are unable to disclose the real brand names and the years to preserve confidentiality. j o u r n a l p r e -p r o o f < table 3 about here > table 4 provides summary statistics for the key variables in our model. these summary statistics suggest that the exclusive agent channel is the leading channel for the focal brand, accounting for a majority of the quotes generated, followed by the web, the call center, and the independent agent channel in that order. the focal brand also has many more exclusive agents than independent agents. this situation is common for the leading brands in the industry. the purpose of this research is to rigorously investigate the cross-channel effects through our models. < table 4 about here > table 5 presents the correlation matrix for the key variables in our data. an analysis of the correlations and collinearity diagnostics suggests that the variance inflation factors do not exceed 10, suggesting that multicollinearity is not an issue in our data. < table 5 about here > to capture the cross-channel effects, we model the number of quotes generated in each channel, that is, quotes by exclusive agents (eaq), quotes by independent agents (iaq), quotes from the web (wq), and quotes from the call center (ccq) using the following system of equations (2) 6 we do not model policies for cross-channel effects because qualification of prospects, the stage before policies are issued, could not be fully performed by the web and call center channels, so the number of policies issues through the web was negligible and through the call center was limited. furthermore, the number quotes is highly correlated with the number of policies in the channels in which they were issued (0.91). j o u r n a l p r e -p r o o f where t is week, ea and ia are the number of exclusive agents and independent agents, respectively; web is the number of unique visitors on the website; and cc is the number of call center agents. ad is the number of advertising impressions. busd is the number of business days in the week, ucar is the number of used cars registered in the country, and 1 , 2 , 3 and high correlation between web content and traffic to a website (kalhor and nikravanshalmani, 2015) . therefore, we use web as the measure of the firm's efforts in the web channel. to control for state dependence, we include a one period lagged dependent variable in each equation. we also include lagged quotes from each of the other channels. we log transform the dependent, channel effort, and advertising variables, so the coefficients serve as elasticities. while the focal brand's efforts and outcomes are temporally separated, given the high degree of autocorrelation observed in efforts, we need to account for the potential endogeneity of these efforts in equations (1) are endogenous. we use one time period lags for these variables to rule out possible reverse causality. we also use appropriate instruments for each endogenous variable. for each endogenous variable, we select instruments based on both relevance and exclusion restriction. the number of exclusive agents depends on the wage rate for brokers. if the wage rate is high (low), a firm will hire a fewer (greater) number of agents, satisfying the relevance condition. but the brokers' wage rate is not directly related to the outcome variables. thus, the exclusion restriction is met. therefore, we the wage rate for brokers is a good instrument. 8 similarly, the number of independent agents depends on the independent service agents' wage rate. as in the case of exclusive agents, the firm will hire fewer (greater) independent agents if this wage rate is higher (lower), meeting the relevance criterion. however, the wage rate of the independent service agents is not directly correlated with the outcome variables. thus, the exclusion restriction is satisfied and the wage rate of independent service agents is a good instrument for the number of independent agents. 8 we collected wage rate information from the bureau of labor statistics (https://www.bls.gov/oes/tables.htm). data on suitable instruments for web are generally difficult to obtain. good instrument candidates such as the number of unique visitors to competitors' website are hard to collect, in particular, at the weekly level. the wage rate of web developers determines the number and quality of web developers to design and run a website. although web developers' wage rate is correlated with the number of developers (satisfying the relevance condition), it is uncorrelated with the outcome variables. therefore, the exclusion restriction is satisfied and wage rate of software professionals is a good instrument for the number web development employees. the wage rate of telemarketers determines the number of telemarketers who would be hired in a call center. although telemarketers' wage rate is correlated with the number of call center employees (maintaining the relevance condition), it is uncorrelated with the outcome variables. therefore, the exclusion restriction is satisfied and wage rate of telemarketers is a good instrument for the number of call center employees. because competing brands vie for the same customers, the focal brand's advertising will be correlated with the past advertising of its closest competitor brands. thus, the instrument relevance criterion is met. however, past competitor brand advertising is not directly related with the error terms in equations 1-4 (low correlation of 0.13 to 0.24 in our data), satisfying the exclusion restriction. thus, for the focal brand's advertising efforts, we use the past advertising impressions of its three closest competitors as good instruments. to ensure that the results are consistent for different instruments, we also use another set of instruments, the average of the past four weeks of wage rates in a subsequent robustness check. j o u r n a l p r e -p r o o f we estimate equations (1-4) as a simultaneous equation system with correlated errors. because the values of the outcome variables are large numbers and are log-transformed, they are ratioscaled and normally distributed. in equations (1-4) , the system of recursive equations, the random error components are likely to be correlated, so to increase the efficiency of the system, we jointly estimate the system and model the correlated errors through a multivariate normal distribution as follows: where the equation-specific random errors that are correlated and â�� is the unconstrained variance-covariance matrix. therefore, we estimate the simultaneous equation systems using a seemingly unrelated regression (sur) procedure, consistent with mullahy (1997) . a single equation estimation would produce consistent estimates, but accounting for covariance in the error structure across equations produces more efficient estimates. we test the validity and strength of our instrumental variables. the results of the first stage regressions and the fit statistics appear in table 6 < table 6 about here > based on the results from the first stage regression and the relevant sargan-hansen test, we can conclude that the instruments are valid. we also test the strength of our instrumental variables through the staiger and stock (1997) and stock and yugo (2005) tests. the f-statistic in all cases for single instruments is greater than 10, consistent with staiger and stock (1997) . the f-statistic for the multiple instrument case (focal brand ad) is greater than 13.91, in line with stock and yogo (2005) . therefore, we conclude that the instruments are sufficiently strong. to check how well the instruments control for endogeneity, we examine the correlations between the instruments of marketing efforts and residuals from eq (1) through eq (4). the correlation between 1 and wage rate of brokers is 0.048 (p=0.40); 2 and wage rate of independent service agents is 0.054 (p=0.34); 3 and wage rate for web developers is 0.032 (p=0.57); and 4 and wage rate for telemarketers is 0.029 (p=0.90). these low and insignificant correlations suggest that these instruments control for endogeneity well. the results of the cross-channel appear in table 7 . we discuss the results in the order of the similarly, the number of independent agents has a significant but smaller negative effect on exclusive agent quotes (-0.0353; p < 0.05). taken together, the results reveal that the effects of efforts in the exclusive agent and independent agent channels on each other are asymmetric. thus, in general, efforts in a rich, primarily informative channel, a primarily persuasive channel, and a balanced channel reinforce one another and have complimentary effects on outcomes in one another's channel. however, there are some interesting asymmetries between the channels in these effects. while the effect of efforts in a primarily informative channel on shopping outcomes in a balanced channel is positive, the effect of efforts in a balanced channel on shopping outcomes in a primarily informative channel is insignificant. a possible explanation follows. a shopper typically encounters a leaner, informative channel earlier in the shopping journey. as a result, marketing efforts in such a channel will have positive effects on outcomes in richer, persuasive channels such as the call center and exclusive agent. in contrast, once a shopper moves to a balanced channel such as the call center, efforts in that channel will unlikely move the shopper back to a primarily informative channel for a desired shopping outcome. efforts in the rich, primarily persuasive, independent agent channel have some positive and significant effects on shopping outcomes in the other channels, but efforts in the other channels do not have a positive outcome in the independent agent channel. this asymmetry can be explained as follows. unlike the independent agent channel, the web, call center, and exclusive agent channels are firm-controlled channels. the efforts in these channels directly benefits the firm. in contrast, because independent agents deal with multiple brands, not all of their efforts are directed toward the focal brand. as a result, the complementary effect of marketing efforts in a dissimilar channel (primarily informative or balanced) is typically weak or insignificant. however, the substitutional effect of a similar channel (another primarily persuasive channel) on the independent agent channel is strong given the focal firm's stakes in its own exclusive agent channel. given the dedicated sales efforts of exclusive agents, the efforts of independent agents do not make much dent in the number of quotes issued through the exclusive agents. we now discuss the other effects in the models. the effects of efforts in a channel on outcomes within the same channel are positive and significant; exclusive agents (p < 0.01), independent agents (p < 0.01), website (p < 0.10), and call center (p < 0.05). the magnitudes of own channel and cross-channel elasticities reveal several surprising findings. table 8 shows the significant own-and cross-channel elasticities and within channel advertising elasticities. the confidence intervals for these elasticities, obtained by bootstrapping given 182 data points, appear in web appendix table a7 . with regard to own channel elasticities, we < table 8 about here > to ensure that the results are robust, we performed several alternative analyses. first, we use cost shifters and hausman style instruments to account for the endogeneity of cross-channel and advertising efforts, respectively. a potential concern with contemporaneous cost shifters like wage rates is that managers may be able to make their effort decisions based only on past values of cost shifters, rendering lagged values of such variables to be more appropriate instruments. as a robustness check, we reestimate our models with averages lagged wage rates as instruments. the results (web appendix table a1 ) are consistent with those of our main model. second, we focus on quotes for our cross-channel model because quote is a key metric that the industry uses analyze at the channel level and because only a limited number of policies were issued in the web and call center channels. nevertheless, we did a robustness check of the model with policies as the dependent variable primarily for exclusive and independent agents. the results (web appendix table a2 ) are substantively similar to those of our proposed model. third, to assess the value of estimating cross-channel effects in the presence of advertising, we estimated two alternative models, one without advertising and one with no channel efforts. the results appear in web appendix tables a3 and a4. the fits of these models are inferior to those of the proposed models. furthermore, the results differ from those of the proposed full models. thus, these results underscore the importance of including both channel effects and advertising in the model to accurately estimate cross-channel effects on shopping outcomes. web efforts facilitate outcomes from exclusive agents (0.07) and the call center (0.04). similarly, call center efforts enhance quotes from exclusive agents (0.02) as well as from the web (0.05). we note that var models are typically data intensive and their estimates and impulse response functions may not fully converge to stable values when the number of observations is small. furthermore, we are testing theory-driven hypotheses. for these reasons, we retain the results from the proposed set of equations that we developed to test the hypotheses. < figure 4 about here > to summarize, our results largely support our hypotheses of complementary channel effects among a primarily persuasive channel (exclusive agent), a primarily informative channel (the web), and a balanced channel (call center). however, cross-channel complementarity is asymmetric in two ways. the first asymmetry is in the direction. in the case of the web and the omnichannel management is a burgeoning area for research and practice. our research on crosschannel effects has critical implications for research and practice. the test of hypotheses offer new insights on asymmetric cross-channel effects. our results could lead to a broader theory of channel effects. in such a theory, channel efforts make significant our research also offers scope for the development of a theory about the relative magnitudes of channel effects on shopping outcomes. we find that within each channel, advertising elasticities are lower than those of channel elasticities, including cross-channel elasticities. this result suggests that within a distribution channel such as the independent agent channel and the exclusive agent channel, the role of the communication vehicle such as advertising may be muted. these findings provide a fertile ground for the development of a more nuanced theory on j o u r n a l p r e -p r o o f the roles of the communication and distribution channels and the interplay between the two channel types in effecting purchases, extending dinner et al. (2014) and li and kannan (2014) . the surprising finding that a rich, primarily persuasive independent agent channel positively affects quotes in a lean, primarily informative web channel without a reciprocating effect poses intriguing theoretical possibilities. when independent agents inform shoppers about alternative brands, some shoppers may visit their websites and indicate their interest in those brands. some of these shoppers may view the brands' websites as more accurate and perhaps unbiased sources of a brand's information than that provided by independent agents, who might be perceived as promoting slanted information on brands better aligned with their compensation. this opens up the role of channel credibility in additionally explaining asymmetric cross-channel effects. the results offer valuable implications for managerial practice. first, managers should use a rich, primarily persuasive channel such as the exclusive agent, a lean, primarily informative channel such as the web, and a balanced channel such as the call center in a coordinated way. such cross-channel integration is critical to firm's growth (cao and li, 2015) . because exclusive agents have the highest own elasticity as well as high cross-elasticity on the call center, firms should allocate more efforts to exclusive agents. moreover, the web has the second highest own elasticity and high levels of cross-channel elasticities on exclusive agents and the call center. therefore, enhancing website capability to offer better information, navigation, visibility, and direct fulfillment is critical. furthermore, because the call center has high own elasticity and high cross-elasticity on exclusive agents, investing in the call center is also advantageous. thus, allocating more resources to dissimilar channels with varying richness--such as the exclusive agent, the web, and the call center--reinforces these channels in a synergistic manner. the j o u r n a l p r e -p r o o f findings have broad implications in other contexts. consumer packaged goods and durables firms should invest in own stores, own websites and call centers to provide a seamless omnichannel experience to the customer and benefit from positive cross-channel effects. second, the result that the rich, primarily persuasive channels, the exclusive and the independent agent channels are substitutional and asymmetric offers practical guidelines. in general, firms may not desire too much competition between exclusive and independent agents lest that leads to price erosion and service variance across the agents. however, because marketing efforts in at least one of these two channels have a negative effect on the other, firms will have to act creatively. to boost overall impact, managers may want to minimize cannibalization between these two agencies by redefining their roles and incentives. they could use independent agents more for reaching new shoppers or for new product launches, and exclusive agents more to develop deep relationships and repeat purchases for core products. because exclusive agents have a stronger negative effect on independent agent quotes than vice versa, managers could reexamine the nature and magnitude of incentives to align these channels together. again, we can extend this result broadly to other contexts and suggest that firms should distribute their products judiciously between exclusive outlets and independent stores and align margins or commission levels for these channels based on the cross-channel elasticities. third, by understanding this nuanced effect of advertising relative to channel effects, capitalize on cross-channel effects, within each channel, managers should assign fewer resources to advertising than to channel efforts. finally, from the cross-channel elasticities, managers can derive greater actionable insights by using an important tool, namely, a graph of influence and influenceability of the different channels. managers can compute influence and influenceability as follows. j o u r n a l p r e -p r o o f where ince c is influence of channel c, c is the total number of channels, e ci is the elasticity of efforts in channel c on outcome in channel i, and inty c is the influenceability of channel c, and e ci is the elasticity of efforts in channel i on outcome in channel c. the influence and influenceability graph for our data appears in figure 5 . the web has the highest total influence on the other channels. in contrast, the exclusive agent is influenced most by the other channels but has the least positive influence on the other channels. the call center channel is in the middle with regard to both influence and influenceability. finally, the independent agent channel has a negative overall influence on the other channels and is also influenced lowest by other channels. < figure 5 about here > managers can use the results from this graph together with own channel elasticities to prioritize the relative importance of each channel for marketing efforts. managers can see that the exclusive agent channel is most valuable because it has the highest own channel elasticity as well as the highest influenceability. they may find the web channel to be important because it has the second highest own channel elasticity and the highest influence on the other channels. managers can also appreciate the call center's value with moderate levels of own elasticity, influence and influenceability. contrary to the efforts expended in it, the independent agent channel has the least own elasticity and cross-channel influence and influenceability. therefore, managers should deemphasize this channel. by determining the relative importance of each channel through this graph, managers can better coordinate the levels of efforts in each channel. j o u r n a l p r e -p r o o f we illustrate the managerial usefulness of our model through a resource allocation exercise. we recognize that closed form optimal marketing efforts based on the estimated model is tedious to derive. therefore, we use numerical analysis to compute the optimal levels of marketing efforts using the estimated own-and cross-elasticities, the levels of other variables, and the unit costs of efforts in each channel. the results appear in table 9 . they show the allocation percentages across the channels at different levels of marketing spending or budget relative to the actual marketing expenditures. these budget levels are akin to different spending scenarios proposed by dinner et al. (2014) . the table also reveals the profit differentials for different marketing spending levels relative to the actual marketing spending. to highlight the value of our crosschannel effects model, the table displays resource allocation outputs based on two models: model without cross-channel effects and with cross-channel effects (see web appendix table a6 ). < table 9 about here > the profit levels for resource allocation based on the model with cross-channel effects are consistently higher than those for the model without cross-channel effects for all marketing budget levels. although the profit from optimization based on a model without cross-channel effects is greater than the firm's actual profit (10.02% more), the profit from optimization based on our model with cross-channel effects is much greater (32.63% more). the incremental profits from our model is highest (37.81% more) when the marketing budget expands by 20%. the optimal reallocation of marketing budget across the channels reveals interesting insights. the firm's actual channel allocation mix was 68.07% for exclusive agents, 20.16% for independent agents, 8.61% for the web, and 3.16% for call center. a reallocation based on optimization through a model without cross-channel effects produces an optimal channel mix of j o u r n a l p r e -p r o o f 94.73% for exclusive agents, 2.02% for independent agents, 0.86% for the web, and 0.32% the call center. these numbers suggest a significant shift in resource allocation from the independent agent, the web, and the call center channels toward exclusive agents mainly because the exclusive agents have the highest own elasticity. however, a reallocation based on our proposed model with cross-channel effects suggests a channel mix of for 75.63% for exclusive agents, 2.02% for independent agents, 16.35% for the web, and 6.01% for call center. these results suggest that the firm needs to move substantially away from exclusive and independent agents toward the web and call center. these results are consistent with the influence-influenceability matrix. the cross-channel effects of a primarily informative channel and a balanced channel on primarily persuasive channels explain the shift in allocation. without considering these crosschannel effects, the firm overspent on primarily persuasive exclusive and independent agents. these findings based on cross-channel effects are consistent with the general trend in channel allocation. across industries, there is a growing shift in allocation of resources toward leaner and more efficient channels such as the web channel. banks have moved more toward the web and mobile channels. so have most retailers, in particular, post covid-19. service organizations are reallocating toward leaner and balanced channels. theoretically, the allocation findings suggest a trade-off between maximizing the high own elasticities of rich, persuasive channels and leveraging the high cross-channel elasticities of lean, informative and balanced channels. our research has limitations that future research could address. first, our models do not include competitor channel efforts because data on competitor brands were unavailable. with competitor data, a more complete model of own and competitor channel efforts could be developed. second, channel roles may vary across categories and stages in customer journey. future research could examine cross-channel effects for categories in which the catalog channel and industrial channels are relevant, extending kauferle and reinartz (2015) . with disaggregate data, future work could explore differential channel roles in customer journey and customer relationship management across channels, furthering verhoef et al. (2010) . third, with data on prices, promotions and customer satisfaction, a more comprehensive model of shopping outcomes can be developed. online and offline prices can differ (ancarani and shankar, 2004) and may result in differential cross-channel effects. furthermore, offline satisfaction can affect online purchases for services (montoya-weiss et al. 2003 ). fourth, our study could be extended to examine cross-channel integration (cao and li, 2015) . fifth, our model is at the channel level as data at the consumer level are difficult to obtain. with shopper level data, we could analyze the mechanisms behind cross-channel effects. finally, cross-channel effects could be studied in conjunction with cross-buying as the drivers and consequences could be somewhat similar (kumar, george, and pancras, 2008) . in conclusion, we sought answers to important questions relating to the growing phenomenon of cross-channel effects. our research offers fresh insights. it shows that cross-channel effects are significant and asymmetric. while channels with dissimilar primary influence roles (e.g., the exclusive agent, the web and the call center channels) are complementary, the direction and magnitude of the cross-channel effects are asymmetric. likewise, while channels with similar primary influence roles (e.g., the independent agent and exclusive agent channels) are substitutional or competitive, their cross-channel effects are also asymmetric. within each channel, own channel elasticity is greater than advertising elasticity. the findings offer important implications for theory and practice with regard to allocation of channel efforts across channels. j o u r n a l p r e -p r o o f j o u r n a l p r e -p r o o f we use hausman style instruments, i.e., advertising decisions made by competing brands as instruments for the focal brand's advertising. these advertising spending decisions may be directly influenced by common advertising cost shock but not demand shock. the sargan-hansen test for over-identifying restriction is insignificant (p > 0.05), suggesting the instruments are valid. note: * p < 0.10; ** p < 0.05; *** p < 0.01. for each instrument, f-statistic is well above 10, satisfying staiger and stock (1997) , stock and yogo (2005) tests. j o u r n a l p r e -p r o o f j o u r n a l p r e -p r o o f j o u r n a l p r e -p r o o f price levels and price dispersion within and across multiple retailer types: further evidence and extension customer channel migration adding bricks to clicks: predicting the patterns of cross-channel elasticities over time consumers in a multi-channel environment: product utility, process utility, and channel choice offline showrooms in omnichannel retail: demand and operational benefits clicks and mortar: the effect of online activities on offline sales battle of retail channels: how product selection and geography drive cross-channel competition cross-channel effects of promotions: an empirical analysis of multi-channel grocery sector the impact of cross-channel integration on retailers' sales growth cost structure, customer profitability, and retention implications of selfservice distribution channels: evidence from customer behavior in an online banking channel the impact of ebook distribution on print sales: analysis of a natural experiment the relative advantage of electronic channels: a multidimensional view organizational information requirements, media richness and structural design the effectiveness of different forms of online advertising for purchase conversion in a multiple-channel attribution framework what makes commercial web pages popular? an empirical investigation of web page effectiveness driving online and offline sales: the crosschannel effects of traditional, online display, and paid search advertising theory of reasoned action: some applications and implications in beliefs, attitudes, intentions, and behavior: an introduction to theory and research the value of rapid delivery in omnichannel retailing benchmarking retail productivity considering retail pricing and format strategy substitution or promotion? the impact of price discounts on cross-channel sales of digital movies the impact of contextual television ads on online conversions: an application in the insurance industry the economic aspects of advertising correlation between content and traffic of universities' websites digital marketing: a framework, review and research agenda optimizing a menu of multiformat subscription plans for ad-supported media platforms distributing through multiple channels in industrial wholesaling: how many and how much cross buying in retailing: drivers and consequences who are multichannel shoppers and how do they perform? correlates of multichannel shopping behavior are multichannel customers really more valuable? the moderating role of product category characteristics multichannel strategies for managing the profitability of business-to-business customers attributing conversions in a multichannel online marketing environment: an empirical model and a field experiment the own-wage elasticity of labor demand: a metaregression analysis multichannel retailing: a review and research agenda optimizing marketer costs and consumer benefits across "clicks" and "bricks catalogue as a tool for reinforcing habits: empirical evidence from a multichannel retailer can marketing campaigns induce multichannel buying and more profitable customers? a field experiment determinants of online channel and overall satisfaction with a relational, multichannel service provider instrumental-variable estimation of count data models: applications to models of cigarette smoking behavior a hierarchical marketing communications model of online and offline media synergies temporal differences in the role of marketing communication in new product categories information and consumer behavior advertising as information key issues in multichannel customer management: current knowledge and future directions its time to think omnichannel shopper, not just omnichannel electronic and physical market channels: a multiyear investigation in a market for products of uncertain quality modern (multiple) time series models: the dynamic system building with bricks and mortar: the revenue impact of opening physical stores in a multichannel environment exploring the implications of internet for consumer marketing communication and persuasion: central and peripheral routes to attitude change the effect of internet distribution on brick-and-mortar sales opportunities and challenges in multichannel marketing: an introduction to the special issue how potential conflict drives channel structure: concurrent (direct and indirect) channels persuasion in the marketplace: how theories of persuasion apply to marketing and advertising, in the persuasion handbook a study of 46,000 shoppers shows that omnichannel retailing works how does adoption of the online channel impact customers' spending in retail stores: conflict or synergy? instrumental variables regression with weak instruments testing for weak instruments in linear iv regression the economics of information home internet access and usage in the usa: trends in the socioeconomic digital divide multichannel marketing report. the direct marketing association determining the cross-channel effects of informational websites, doctoral dissertation dual distribution and intangible firm value: franchising in restaurant chains multichannel shopping: causes and consequences multichannel customer management: the research shopper phenomenon from multi-channel retailing to omni-channel retailing: introduction to the special issue on multi-channel retailing crm in data-rich multichannel retailing environments: a review and future research directions can offline stores drive online sales? the evolution of marketing channels: trends and future research directions marketing's profit impact: quantifying online and offline funnel progression exclusive vs. independent agencies: a comparison of performance. the geneva papers on risk & insurance theory in-sample and out-of-sample predictive model validation for quotes in each channel key: cord-355834-kziy850d authors: qiu, liangsheng; dong, sufen; ashour, ashraf; han, baoguo title: antimicrobial concrete for smart and durable infrastructures: a review date: 2020-11-10 journal: constr build mater doi: 10.1016/j.conbuildmat.2020.120456 sha: doc_id: 355834 cord_uid: kziy850d concrete structures in sewer systems, marine engineering, underground engineering and other humid environments are easily subjected to microbial attachment, colonization and, eventually, deterioration. with careful selection and treatment, some additives including inorganic and organic antimicrobial agents were found to be able to endow concrete with excellent antimicrobial performance. this paper reviews various types of antimicrobial concrete fabricated with different types of antimicrobial agents. the classification and methods of applying antimicrobial agents into concrete are briefly introduced. the antimicrobial and mechanical properties as well as mass/weight loss of concrete incorporating antimicrobial agents are summarized. applications reported in this field are presented and future research opportunities and challenges of antimicrobial concrete are also discussed in this review. concrete is the most widely used construction material for various infrastructures worldwide. however, concrete structures in certain aggressive environments, such as sewer systems, marine engineering, buildings exposed to high humidity and the like, are easily suffered from microbial attachment, colonization, eventually, deterioration [1] [2] [3] [4] . for example, the most typical problem faced by reinforced concrete structures in sewer systems is microbial induced corrosion, which is still commonly referred to as a sulfide (h 2 s) gas problem. the process initiated when sulfatereducing bacteria (srb) convert sulfate into hydrogen sulfide gas under anaerobic conditions, that is converted into corrosive sulfuric acid by sulfur-oxidizing bacteria (sob) of the genus thiobacillus [1, [5] [6] [7] [8] [9] [10] [11] . some fungi also participate in this activity [12, 13] . concrete structures in the tidal and splash zones of marine concrete engineering are dominantly damaged by pseudoalteromonas, along with vibrio, pseudomonas and arthrobacters, etc. [14, 15] . bio-deterioration of concrete in irrigation and hydroelectric canals [16] , spots or patches covered on concrete walls [17] , and biological decay of mortars on building facades [18] commonly result from the growth of algae and cyanobacteria. algal growth is also quite common on concrete walls of water storage and conveyance structures [19] . salmonella, an important foodborne pathogen, are easily attached and colonized on surfaces of concrete used in food industry due to their adherence forming biofilms [20] . the propagation and proliferation of microorganisms including bacteria (e.g., pathogens), fungi, and algae alone or together, on and/or in concrete structures, will affect concrete's aesthetic appearance, destroy the internal structure of concrete, degrade mechanical properties and durability of concrete, increasing the cost by rehabilitation and even replacement [2, 16, [21] [22] [23] . therefore, developing antimicrobial concrete for smart and durable infrastructures has become extremely significant and imperative. researchers have been attempting to develop antimicrobial concrete (concrete is a collective term referring to concrete, cement mortar and cement paste as well as cementitious/ cement-based materials/composites in this paper) by adding some additives having antimicrobial properties for sterilization against a specific microorganism or multiple microorganisms, meanwhile without significantly impairing concrete essential properties such as compressive strength. the last two decades have witnessed an ever-increasing growth in studies on the utilization of functionalized zeolites supporting bactericidal metal ions, such as silver, copper and zinc ions [24] [25] [26] [27] . haile et al. [28] [29] [30] reported that concrete containing silver bearing zeolite exhibited antimicrobial characteristics against acidithiobacillus thiooxidans (a.thiooxidans), as reflected by inhibition of formation of a.thiooxidans biofilm. further, xu [31] and li [32] reported that concrete added with silver-loading zeolite and polypropylene fiber exhibited obvious bactericidal effect towards escherichia coli (e. coli). moreover, it is reported that antimicrobial concrete containing zeomighty (zeolites with silver and copper ions) was introduced on the japanese market [33] . quaternary ammonium compounds (quats) have been used as antimicrobial agents for a long time, and only recently they have been reported to be effective as algaecides [11, 16, 19, 34] . intentionally, considering the severe consequences caused by microbial induced corrosion of concrete, considerable attention has been paid to find effective antimicrobial agents to admix into concrete in order to fight against thiobacill [3, 23, 35, 36] . for instance, shook and bell [37] reported that con-shield, added into concrete during the mixing stage, showed high sterilizing rate and stable bactericidal effect against thiobacillus bacteria. yamanaka et al. [38] found that calcium formate was able to completely inhibit the growth of both sulfur-oxidizing and acidophilic iron-oxidizing bacteria at concentrations above 50 mm. some investigators tried to develop antimicrobial concrete by incorporation of nickel, and tungsten specially targeting at sob which play a dominant role in biogenic corrosion of sewer systems [39] [40] [41] [42] [43] . sun et al. [44] verified the strong bactericidal effect of free nitrous acid (fna) on microorganisms due to cells in corrosion biofilms of concrete surfaces were killed. in addition, the combination of water repellents (decrease bio-receptivity) plus biocides (decrease biological activity) has been reported to be effectively inhibiting microbial growth in mortars, white concretes and autoclaved aerated concretes [45, 46] . vaquero et al. [16] proposed a novel cement-based material with biocidal activity that can be used as an overlay of mortar in existing structures, such as canals and pipes. in recent years, with the rapid development of nanotechnology, some researchers have tried to introduce some nano-particles into concrete to inhibit microbial colonization. for example, the research undertaken by singh et al. [47] indicated that cement-zno composite possesses effective antibacterial and antifungal activities under dark and solar light due to the addition of zno nano powder. wang et al. [48] demonstrated that high performance concrete (hpc) incorporated with nano zno has antibacterial ability against e. coli and staphylococcus aureus (s. aureus). concrete fabricated with titanium dioxide nanoparticles has great potential in sterilization under the light [49] . ganji et al. [50] found that cement with nano-tio 2 inhibit the growth of e. coli under uv irradiation. moreover, fonseca et al. [18] proposed that anatase can be an alternative application for preventing bio-deterioration of mortars. this paper is intended to summarize antimicrobial concrete fabricated with different types of antimicrobial agents, intuitively shown in fig. 1 . first, the classification of antimicrobial agents and their application methods into concrete are briefly introduced. then, the antimicrobial and mechanical properties as well as mass/ weight loss of concrete incorporated with antimicrobial agents are reviewed, with emphasis on antimicrobial properties. subsequently, antimicrobial mechanisms of some inorganic and organic antimicrobial agents were explicated. finally, applications of antimicrobial concrete in sewer systems, marine engineering and buildings against microbial threat are also presented. the antimicrobial property of antimicrobial concrete was attributed to the addition of antimicrobial agent, which is a collective name herein for the mentioned antimicrobial additives facilitating concrete to inhibit and/or kill various microbes including bacteria (e.g., pathogens), fungi, and algae. antimicrobial compounds including biocides, microbicides, sanitizers, antiseptics and disinfectants characterized by their ability of killing microorganisms and/or inhibiting microbial reproduction, are easily accessible [23, 34] . the antimicrobial agents reported to have been added to concrete ingredients can be classified into inorganic and organic antimicrobial agents with respect to their chemical composition as detailed below. inorganic antimicrobial agents that have been reported to be applied in concrete include heavy metals (silver, nickel, tungsten), metal compounds (silver molybdate, copper oxide, zinc oxide, sodium tungstate, sodium bromide), norganix (a silicate concrete sealer), free nitrous acid (fna), and nano inorganic antimicrobial materials. the antibacterial activity of metal or metal ions is in the order of: ag > hg > cu > cd > cr > ni > pb > co > zn > fe [22, 32, 51, 52] . although silver ion antibacterial agent series are effective but considering their high cost, few other alternatives with high bactericidal effect were explored in the literature. for example, zhang [22] found that cerium nitrate exhibited an excellent antibacterial effect in porous concrete, even with a low content of 1.25%. furthermore, the use of nanomaterials to control microbial colonization of concrete has considerably increased in recent years [53] . nanoparticles (nps) of cu 2 o, caco 3 , tio 2 , zno, cuo, al 2 o 3 , fe 3 o 4 , etc., were reported to exhibit inhibitory effects against a wide range of microorganisms in this field [3, 4, 26, 47, 48, 54, 55] . quats, phthalocyanine compound (including metal organic antimicrobial agent copper phthalocyanine), calcium formate, alkyl nitro-bromide (a ⅱ b), isothiazoline/cabamate, conshield (a highly charged cationic polymer), and conblock mic (whose active ingredient is 3-trimethoxy silyl propyl dimethyl octadecyl ammonium chloride) are various organic antimicrobial agents used in concrete. additionally, freed et al. [56] proposed that fibers incorporated with at least one antimicrobial agent, such as microban b (a phenol-based antimicrobial agent), were able to inhibit microorganisms. quats are the most representative organic antimicrobials, e.g., silane quaternary ammonium chloride(sqa) [57] , and cetyl-methyl-ammonium bromide [19] , which have been widely studied and applied by researchers [23, 51, 58] . isothiazoline/cabamate is a type of organic antifungal agents, often used to target at aspergillus niger which is easily found in the interiors and exteriors of buildings in damp environment [59] . uchida et al. [11] stated that water pollution as a result of metal eluted into sewage can be addressed by adding a phthalocyanine compound (a metal phthalocyanine, a metal-free phthalocyanine, and derivatives thereof) into concrete, which will not pollute water and a small amount of inhibitor can prevent deterioration of concrete or mortar due to sob for a long time. generally, inorganic antimicrobial agents have long service life and high temperature resistance, but have side effects like toxicity. organic antimicrobial agents possess obvious bactericidal effect in a short term and broad spectrum of killing activity but their temperature resistance is poor [22, 31, 32, 60] . moreover, most of organic biocides are ultimately ineffective at removing microbes and may eventually lead to a new wave of microbes on the affected surfaces after microbes develop a resistance [34] . the following sections will describe these antimicrobial agents and their methods of applications in detail. some antimicrobial agents use inorganic or organic cementitious materials as carriers to form protective coatings, with biocidal property on concrete surfaces [23, 35] . another method to apply antimicrobial agents into concrete is directly incorporating antimicrobials into concrete mix as functional components after pre-dispersion [23, 35] . for example, calcium formate was added in the mixture [38] , conshield was incorporated into the mix and the protection was throughout the entire thickness of concrete matrix [37] . the antimicrobial watertight admixture made of fluosilicate salts and antimicrobial compounds (ni and w) [61] is in liquid state to be homogeneously dispersed in concrete. the phthalocyanine compound [11] can be dispersed uniformly in concrete or mortar by a blending agent selected from a group consisting of an air entraining agent, a water reducing agent, and a viscosity increasing agent. liquid bactericides like dimethyl benzyl ammonium chloride can be made into powder adsorbed by carrier such as zeolite [23, 62] . in addition, heavy metal antibacterial agents are usually fixed on zeolites by means of adsorption or ion exchange [27, 51, 63] . known as crystalline pozzolanic aluminosilicate minerals with uniform molecular sized pores, zeolites can be functionalized to exhibit antimicrobial property if calcium and sodium ions in their framework are exchanged by silver, copper or zinc ions, explaining that zeolites are the most common carriers of inorganic metal ions [3, 26, 27, 29, 51, 63, 64] . agglomeration due to high activity of antimicrobial nanoparticles in cement matrix is a common concern, significantly decreasing their chemical and physical activities and, hence, affecting their efficiency in cement matrix performance and antimicrobial activity [49, 60] . a dispersion medium (most likely mixing water) and incorporation of organic admixtures and different surfactant types, e.g., plasticizers and superplasticizers, facilitate to address the issue of homogeneous dispersion in the cement matrix, as presented in fig. 2 [49, 54] . it is also reported that the application of superplasticizer in photocatalytic cement can enhance nano-tio 2 dispersion in samples by preventing agglomeration of titanium dioxide in cement pastes, which is also conducive to improve the contact between titanium dioxide and bacteria, contributing to better bacterial inactivation [50] . however, in the case of antimicrobial agents being functional components in concrete, the selection of biocide types and contents has not been systematically investigated [35, 65] . the antimicrobial property is the most important assessment factor for antimicrobial concrete, that varies with the addition of different types of antimicrobial agents, as summarized in table 1 . [30] a.thiooxidans concrete growth of planktonic and biofilm populations of a.thiooxidans was inhibited zinc and silver loading zeolite [29] a. thiooxidans concrete functionalized zeolite coated concrete specimens with epoxy to zeolite weight ratios of 2:2 and 1:3 had negligible biomass growth and acid production rates silver/copper zeolite, silver/zinc zeolite [28] a.thiooxidans mortar co-cations such as zn 2+ and cu 2+ increases antimicrobial activity of silver bearing zeolite nano-copper oxide [26] a.thiooxidans concrete higher leaching rate of copper from loosely adhered nano-copper oxide film significantly inhibited the activity of a.thiooxidans silver copper zeolites [25] e. coli, listeria monocytogenes, salmonella enterica or s. aureus centration of silver copper zeolites to obtain a bactericidal effect on mortar surfaces is required more than 3% zeomighty [33] thiobacilli n.a. a concentration of metal zeolites of 1% to cement weight is optimum for suppressing the growth of thiobacilli sodium tungstate [41] a. thiooxidans n.a. approximately 10 times more tungstate bound to the cells of a. thiooxidans at ph 3.0 than at ph 7.0 sodium tungstate [42] a. ferroxidans n.a. approximately 2 times more tungsten bound the cells of a. ferroxidans at ph 3.0 than at ph 6.0 metal (ni,w) compounds, znsif 6 [61] t.novellus mortar, concrete mortar with antimicrobial watertight admixture had higher ph(6.8) and lower concentration of sulfuric acid(3.78 â 10 -8 mol/l) compared to that (6.6 and 2.56 â 10 à7 mol/l) of plain mortar zinc oxide, sodium bromide, copper slag, ammonium chloride, cetyl-methylammonium bromide [19] algae mortar adding 20 wt% zinc oxide and 20 wt% sodium bromide exhibited the most effective algal inhibition under laboratory conditionthe addition of 20 wt% sodium bromide and 10 wt% cetyl-methyl-ammonium bromide (an organic antimicrobial agent) showed highest inhibitory effects at under field condition fna [44] n.a. concrete h 2 s uptake rate decreased by 84-92% 1-2 months and viable bacterial cells reduced from 84.6 ± 8.3% to 10.7 ± 4.3% within 39 h after fna spray. silver molybdate [52] e. coli and s. aureus concrete the residual colony count of e. coli and s. aureus is 0 cfu/ml by addition of 0.004% silver molybdate cerium nitrate [22] e. coli concrete bacterial concentration reduced drastically from 7.50 to 0.01,0,0.02 million per ml after 48 h when the content was 1.25,5.00,10.00%, respectively. nano sized tio 2 , caco 3 [4] pseudomonas, fusarium, algae, blue-green algae and manganese oxidizing bacteria mortar nano-tio 2 modified fly ash mortar and nano-sized tio 2 , caco 3 modified fly ash mortar exhibited enhanced antibacterial activities compared to nano-caco 3 modified fly ash mortar anatase [18] cyanobacteria and chlorophyta species mortar two types of mortars with different kinds of sand showed the lowest photosynthetic growth ratio (0% and 0.03%, respectively) sio 2 /tio 2 nano-composite [68] e. coli cement mortar bacteria inactivation after uv light irradiation and without illumination after 120 min was 67% and 42%, respectively. antimicrobial concrete, with the addition of diverse antimicrobial agents against microorganisms involving in microbial induced corrosion, especially in sewer systems, have been extensively studied in the literature. nickel and tungsten have been known to protect concrete from microbial corrosion owing to their antimicrobial effect towards causative bacteria, i.e., thiobacillus thiooxidans (t. thiooxidans). negishi et al. [41] found that the cell growth of a. thiooxidans, including strain nb 1-3 (isolated from corroded concrete in fukuyama, japan) was strongly inhibited by 20 lμ sodium tungstate, and completely inhibited by 50 lμ sodium tungstate. similarly, sugio et al. [42] reported that cell growth of an ironoxidizing bacterium, acidithiobacillus ferroxidans (a. ferroxidans), was strongly inhibited by 0.05 mm and completely inhibited by 0.2 mm of sodium tungstate. in the study of maeda et al. [40] , concrete containing 0.1% metal nickel and concrete with 5 mm nickel sulfate were found to completely inhibit the cell growth of strain nb 1-3 of t. thiooxidans isolated from corroded concrete. moreover, kim et al. [61] conducted an investigation to evaluate the antibiosis of antimicrobial ingredients (ni and w) of antimicrobial watertight admixture mixed in mortar and concrete on thiobacillus novellus (t. novellus). broth microdilution mic test indicated that t. novellus could not survive in the area where the admixture is dropped. as reflected in table 1 , the total colony test numerically shows that t. novellus in culture solution with mortar added with the admixture were disappeared after 24 h. the biochemical corrosion simulation test also indicated that the number of t. novellus was much lower in the case of mortar mixed with the admixture than plain mortar specimens. the results suggested that the addition of antimicrobial watertight admixture in cement mortar and concrete suppressed the growth of t. novellus. furthermore, southerland et al. [66] found that tungsten used alone is able to inhibit growth of t. novellus, whereas molybdenum, ammonium molybdate or a mixture of ammonium molybdate and tungstate activates growth of the same bacteria. likewise, it is reported that molybdenum activates growth of t. novellus but inhibits growth of t. thiooxidans, indicating sob of the same genus thiobacillus have different growth inhibitory mechanism. it is noteworthy that the antimicrobial property of antimicrobial agent ni and w is not only largely dependent on their contents, but greatly affected by ph. it is generally recognized that nickel compounds are suitable for neutral environment, while tungsten compounds are more effective in acidic environment [23, 43] . maeda et al. [40] observed that the amount of nickel contained in the strain nb 1-3 cells treated without nickel, treated with 10 mm nickel sulfate at ph 3.0 and treated with 10 mm nickel sulfate at ph 7.0 was 1.7, 35 and 160 nmol nickel per mg protein, respectively. the results indicated that nickel is able to bind to strain nb 1-3 cells, and much more nickel binds to the cells at neutral ph than at acidic ph demonstrated that nickel ions have a better inhibitory effect towards the microbe in neutral environment than in acid environment [40] . the findings of negishi et al. [41] and sugio et al. [42] , as detailed in table 1 , demonstrated that the antimicrobial property of tungsten is more effective in acidic environment than in neutral environment. furthermore, kong et al. [62, 65] conducted an investigation to evaluate the impact of adding five bactericides in concrete towards the selected bacteria (as listed in table 1 ), and to study their applicability for controlling and preventing microbial corrosion of concrete. they reported that concrete with sodium bromide and zinc oxide exhibited excellent antimicrobial property towards the tested bacteria, especially bacteroidetes, as the number of microbial populations decreased substantially. however, the antimicrobial effect of concrete with a dispersion of sodium tungstate on microbes is worst, as reflected by the lowest bactericidal rate (21.95%), it even promotes the growth and reproduction of proteobacteria. they also observed the dead and live microorganisms within biofilm with confocal scanning laser microscopy (clsm), as seen in fig. 3 . the number of live cells within the biofilm all decreased to a certain degree, indicating all the tested bactericides have a certain sterilizing effect. similarly, bao [67] obtained that the surface roughness of the control mortars and mortars with sodium tungstate and sodium bromide was 46.65, 14.3 and 9.02 lm after a 3-month immersion in intensified sewage, respectively. therefore, they concluded that the addition of sodium tungstate and sodium bromide could effectively inhibit the growth and reproduction of microorganisms attached to the surface of cement mortar. in addition, sun et al. [44] studied the bactericidal effect of fna on microbes in sewer biofilms of two concrete coupons. they observed that, as for the intact corrosion biofilm, h 2 s uptake rates (sur) were reduced markedly 15 days after fna spray and viable bacterial cells severely decreased by over 80% within 39 h (detailed in table 1 ), suggesting that biofilm cells were killed by the treatment. as for a suspended solution of corrosion biofilms scraped from the concrete coupon, atp level and ratio of viable bacterial cells were also severely decreased by the treatment, as clearly seen in fig. 4 , demonstrating that fna strongly deactivates bacteria of acidic corrosion biofilm [44] . zeolite containing metal ions has been investigated a lot to be used in concrete due to its excellent antimicrobial property. for example, haile et al. [28] evaluated the antimicrobial characteristics of mortar specimens coated with silver bearing zeolite with a. thiooxidans. they observed that biomass concentration of a. thiooxidans dry cell weight (dcw) of control specimens (236 mg tss/l and 181 mg tss/l) was as much as 2-fold higher compared to the mortars coated with silver-loaded zeolite (125 mg tss/l and 80 mg tss/l). the reduced microbial numbers evidenced that the mortar specimens coated silver bearing zeolite exerted antimicrobial characteristic on a.thiooxidans and inhibited bacterial growth. they also found that bacteria were not affected in the nutrient solution indicated that the antimicrobial characteristics of zeolitic coatings were only apparent on solid surface particles [28] . moreover, haile et al. [30] discovered that no biomass growth was observed upon exposure of the bacterium to silver-loaded zeolite coated concrete specimens, and there was no oxygen uptake measured, meaning no viability of a. thiooxidans cells for the silver-loaded zeolite coated concrete specimens. the research results confirmed that zeolite containing 5 wt% ag is inhibitory to planktonic and biofilm of a. thiooxidans [30] . similarly, de muynck et al. [69] observed that mortar specimens with silvercopper zeolites (zeolites contain 3.5% silver and 6.5% copper) obtained a 12-fold decrease of atp content after 24 h, while inhi-bition of antimicrobial fibers on bactericidal activity was limited, indicating biocidal effect towards sob was limited in the case of antimicrobial fibers and that of antimicrobial zeolites was much better. moreover, de muynck et al. [25] investigated the antimicrobial effectiveness of silver copper zeolites against e. coli, listeria monocytogenes, salmonella enterica or s. aureus in a quantitative way. a clear decrease in the total atp content was observed for mortar specimens containing silver copper zeolites, indicating the occurrence of antimicrobial activity by the presence of silver and copper ions. furthermore, they concluded that the concentration of silver copper zeolites is required to be more than 3% so as to obtain a bactericidal effect on mortar surfaces [25] . in the experiment of haile et al. [70] , cellular atp in concrete contained 2.6 wt% silver-loaded chabasite declined to zero with a corresponding dcw value of 35 mg, indicating there was no growth after bacteria were exposed to 2.6 wt% silver-loaded chabasite, whereas the biomass was 51 mg dcw and cellular atp was 0.21 mg for concrete coated 18 wt% silver-loaded chabasite. the results indicated that antibacterial characteristics of concrete specimens coated with 2.6 wt% is superior to the specimens with 18 wt % silver-loaded chabasite. the results of the experiment conducted by xu and meng [64] indicated that the content of e. coli in concrete incorporating silver-bearing zeolite and polypropylene fiber was reduced compared to the control samples, demonstrating that silver-bearing zeolite and polypropylene fiber play a bactericidal role and reduce the breeding of e. coli. likewise, li [32] discovered that concrete specimens added with 0.5% silver-loading zeolite and polypropylene fiber had the most pronounced bactericidal effect towards e. coli, as reflected by the greatest od value (the greater the od value, the lower the bacterial concentration of the concrete samples) according to antibacterial test results. while antimicrobial effect of concrete specimens incorporated with fly ash and mineral powder was not evident. researchers have paid much attention to the effect of antimicrobial nanoparticles on antimicrobial property of concrete. singh et al. [47] admixed zno nano powder into cement composite and evaluated the antimicrobial effect of the formed cement-zno composites against two bacterial strains e. coli, bacillus subtilis and fungal strain aspergillus niger. as shown in fig. 5 , the antibacterial and antifungal effects of cement-zno composite increased as the zno concentration increased in the range of 0,5, 10, 15 wt%. moreover, it was also noted that both antibacterial and antifungal activities of cement-zno composite was enhanced under sunlight compared to dark condition. in addition, wang et al. [48] conducted a research to study the antimicrobial effect of highperformance concrete (hpc) added with nano zno against e. coli and s. aureus. the results showed that the antibacterial rate of the two groups of antibacterial concrete against e. coli reached 100%, however the antibacterial rate against s. aureus was 54.61% and 99.12%, respectively. through sem observations, it is found that nano zno and its resulting compounds precipitation adhered to surface of cement hydrate, thus inhibited the growth of bacteria, accounting for the significant antibacterial effect of hpc [48] . sikora et al. [54] conducted a series of tests to evaluate the antimicrobial effect of four metal oxide nanoparticles (al 2 o 3 , cuo, fe 3 o 4 , zno) used in cement-based composites. they discovered that all the studied nanoparticles inhibited microbial growth, and the growth kinetics showed that the highest inhibitory effect on e. coli atcc 8739 tm and e. coli mg 1655 was fe 3 o 4 nanoparticles, zno nanoparticles, respectively. the biofilm formation assay indicated that the tested nanoparticles were able to reduce the formation of bacterial biofilms, e. coli atcc 8739 tm biofilms were inhibited by all nano-oxides, zno nanoparticles significantly affected the formation of p. aeruginosa and s. aureus biofilms. however, the viability of p. aeruginosa cells in sample with al 2 o 3 was significantly higher compared to the control sample. similarly, fig. 4 . levels of sur, atp and the ratio of viable bacteria measured on reactor solutions containing the suspended corrosion biofilm scraped from a concrete coupon after 40 months of exposure prior to and after fna treatment. the ratio of viable bacteria was not determined after 700 h of fna treatment as cells could not be extracted from the reactor solution [44] . note: sur means h 2 s uptake rate. dyshlyuk et al. [71] evaluated antibacterial and fungicidal properties of zno, tio 2 and sio 2 nanoparticle solution by interaction with eight types of microorganisms commonly causing bio-damage to buildings and concrete structures. they found that zno nanoparticles of 2-7 nm in size with a suspension concentration of 0.01-0.25% displayed the most noticeable antimicrobial properties against the tested strains, decreasing microorganisms by 2-3 orders of magnitude. they also revealed that zno nanoparticles interacted specifically to a microorganism type, leading to a decrease in the number of bacillus subtilis b 1448 bacterium by 2 orders of magnitude, and that of fungi of penicillium ochrochloron f 920 by 3 orders of magnitude. however, tio 2 and sio 2 nanoparticles exhibited a low antimicrobial activity. nano-tio 2 , with its excellent photocatalytic effect, has aroused much interest of many researchers in the aspect of microorganism inactivation. for instance, ganji et al. [50] investigated the antimicrobial performance of cement samples containing 1,5 and 10 wt% nano-tio 2 against e. coli under uv irradiation. they found that bacterial inactivity enhanced as the amount of tio 2 nanoparticles in cement samples increased, however, the inactivation effect was not obvious even the amount of tio 2 nanoparticles further increase to 10 wt%. therefore, 5 wt% tio 2 is proposed to be the most proper content in cement samples for inactivation of e. coli taking into account both the photocatalytic inactivation and cost. linkous et al. [72] employed nano-tio 2 in concrete to inhibit the attachment and growth of oedogonium. they discovered that concrete containing 10 wt% tio 2 nanoparticles obtained a 66% reduction in the growth of oedogonium. besides above, researchers have also investigated antimicrobial effects of antimicrobial concrete towards some other commonly microbes threatening concrete. for example, umar et al. [36] evaluated the antimicrobial activity of four types of semicircular modified cement composite specimens using serratia marcescens collected from seashore and then isolated from microbe samples. the results showed that cement composites admixed with sodium nitrite-based inhibitor performed better with the least percentage increment of total viable count at the end of 144 h as compared to the cement composite with styrene acrylate copolymer, with acrylic polymer, and cement composite without any admixture, respectively. this can infer that cement composite with sodium nitrite-based inhibitor exhibited noticeably improved ability to suppress the growth of serratia marcescens in marine environment. norganix [73] is able to endue concrete with powerful antimicrobial property to eliminate salmonella, listeria, e. coli, clostridium, and mold spores not just on the surface but deep within the concrete. moreover, antimicrobial concrete with norganix can prevent microbes from reentering concrete from any directions because norganix will hydrate with the unused portland cement within the concrete to generate new cement, thereby sealing the capillary system. paiva et al. [20] determined the antimicrobial efficiency of bio-sealed for concrete tm , a hydro-silicate catalyst in a colloidal liquid base, to prevent salmonella spp. attached on concrete bricks in food industry. they found that concrete bricks treated with biosealed for concrete tm after inoculation, before and after inoculation had immediate bactericidal effects towards the tested five strains of salmonella in contrast with bricks not treated with bio-sealed for concrete tm and bricks treated with biosealed for concrete tm before inoculation, as observed by significantly lower viable counts of salmonella. yamanaka at al. [38] studied the inhibitory effects of formats on the growth of bacteria causing concrete corrosion in sewerage systems. they found that the growth of sob isolated from corroded concrete were completely inhibited by 10 mm calcium formate for 18 days, while the growth of acidophilic iron-oxidizing bacteria was inhibited by 10 mm calcium formate during 34 days. this finding shows that even the same antimicrobial agent has different inhibitory effect on different microbes. in addition, they also observed that the formation of atp in bacterial cells was ceased after the addition of calcium formate into concrete test pieces. erbektas et al. [57] evaluated the antimicrobial efficacy of silane quaternary ammonium chloride (sqa) aqueous salt solution against planktonic halothiobacillus neapolitanus and a.thiooxidans. they found that the antimicrobial efficacy directly related to bacterial population and activity, and indirectly depends on ph. furthermore, antimicrobial effectiveness occurs when the ph is greater than 4. in the research undertaken by do et al. [59] , cement mortars with isothiazoline/cabamate exhibited a good antifungal effect against aspergillus niger, while mortars with nitrofuran did not show inhibitory effect even the content of nitrofuran was up to 5 wt%. moreover, the antifungal effect of cement mortar containing isothiazoline/cabamate on aspergillus niger enhanced almost linearly as the content increases (0%,0.3%,0.5%,1%,2% and 5% by mass to cement). according to [74] , researchers of former soviet union tested mortar samples with alkyl nitro-bromide (a ⅱ b) stored for 6 years. the results indicated that the microbial retention rate on the surface of mortar specimens was merely 0.6% and 0.1% when the content of a ⅱ b is 0.025 wt% and 0.05 wt%, respectively, after 5 h of irradiation, confirming the strong and long-lasting antimicrobial ability of a ⅱ b. it is worthwhile noting that some organic antimicrobial agents are extremely suitable to add into concrete due to their antimicrobial power to combat against diverse microbes, rather than only a single type of microbe. for example, kong et al. [62] found that concrete added with copper phthalocyanine exhibited outstanding antimicrobial effect with high bactericidal rates towards bacteroidetes (90.82%) and proteobacteria (64.25%), and the bactericidal rate towards all tested microbes is as high as 82.59%. the number of live cells within the biofilm attached to concrete added with copper phthalocyanine showed a significant drop, and the content of live cells was only 12% of that attached to plain concrete. a large number of dead microbes was observed, as seen in fig. 3 (f). vaquero et al. [16] studied the bactericidal ability of 15 commercial bactericides blended into concrete against microbial induced corrosion by culturing microbes and evaluating the antimicrobial efficiency. research results indicated that the multicomponent formulation pl-uv-h-2b was the sole formulation to succeed in all the evaluation process among all formulations. concrete samples fabricated with pl-uv-h-2b, of which the actives are 30% 2-octy l-2h-isothiazol-3-one + terbutryn and 15% 2,4,4 0 -trichloro-2 0 -hyd roxy-diphenyl ether (calcium filler as a dispersive matrix), exhibited high effectiveness in antimicrobial tests against algae (scenedesmus vaculatus and stichococcus bacillaris), fungus (aspergillus niger), and bacteria (s. aureus and e.coli), both before and after accelerated aging processes, as exhibited in fig. 6 . they also paid special attention to the reasons responsible for failure of some biocide formulations and concluded that the water-soluble bactericide showed a lower retention rate in concrete and thus plays a poor role in protecting concrete in the long term [16] . urzìet al. [45] evaluated the efficiency of three water-repellent compounds and two biocide compounds, i.e. algophase and the new water miscible formulation algophase ph 025/d having the same active ingredient 2,3,5,6-tetrachloro-4-methylsulfonyl-pyridine, against microbial colonization of mortars both in laboratory conditions and outdoors. they observed that the application of waterrepellent alone was insufficient to prevent biofilm growth on the surface, whereas the combined application of water repellents and biocides in a single step prevent microbial growth, reflecting by complete absence of bacterial colonization, absence of algal colonization, dramatically reduced colonization by fungi on the surface of mortars (seen the representative samples t4 and t5 shown in fig. 7) . single-step application of biocide and water repellent exhibits excellent performance due to biocide compound randomly distribute below, between and above the hydro-repellent film. in this way, the biocide has the ability to remove the remains of old colonization below, and stop new microbial colonization on the surface [45] . shook and bell [37] evaluated the antimicrobial effect of conshield using wafers of concrete mortar incubated with a bacterial suspension of t.thiooxidans, t. thioparus, and t. denitrificans. the results indicated that the viable bacterial count of concrete wafers treated with conshield is zero, suggesting that conshield killed all of the tested bacteria with a complete 100% kill after 24 h. moreover, it is reported that conblock mic [75], whether integrated throughout the matrix of concrete when used as an admixture and/or directly applied to concrete as a surface treatment, it inhibits the growth of bacteria, fungi, mold, and algae. freed et al. [56] evaluated the efficacy of concrete reinforced with fibers incorporating microban b. the inhibition zone of concrete treated with polypropylene fibers containing microban b towards e. coli, s. aureus, and mixed mold(fungi) was 3,4, and 2 mm, respectively, indicating fibers carrying microban b could kill microorganisms. above investigations have indicated that antimicrobial agents could endow concrete with antimicrobial property to varying degrees. antimicrobial properties of antimicrobial concrete is largely depending on respective intrinsic natures, types and contents of antimicrobial agents. however, the existing researchers paid little attention to the impact of the addition of antimicrobial agents on the microstructure of concrete. it is necessary to establish the underlying connections between different properties as well as the microstructure of concrete after adding antimicrobial agents. moreover, high retention rate of antimicrobial agents in concrete is required in order to maintain the long-lasting inhibitory or killing effect towards microbes, while the long-term retention rate of a biocide and its influence on the other properties of concrete are poorly understood [35, 65] . antimicrobial concrete exhibited different mechanical properties for various types and quantities of antimicrobial agents added. kim et al. [61] reported that compressive strength of concrete with antimicrobial watertight admixture, of which antimicrobial ingredients are nickel and tungsten compounds, was decreased at an early age but the long-term compressive strength was increased. de muynck et al. [25] observed a small decrease in compressive strength of mortar specimens added with the highest concentration of zeolites (4.65%), i.e. 41.1 ± 0.8 mpa as compared to 49.0 ± 3 .4 mpa for control specimens. kong and zhang et al. [65, 76] tested the 7 days, 28 days and 56 days compressive strength of concrete added with different types and contents of bactericide. they observed that the 28 days compressive strength of concrete adding with copper phthalocyanine (cp) was enhanced by 60% with the dosage of 0.1%, which indicated that cp not only increased the fluidity of concrete, but also accelerated the hydration of cement, thus promoted the strength development by dispersing cement. meanwhile, the enhancement of compressive strength also makes some contribution to maintain the surface ph of concrete added cp as high as 10.6. however, the strength of concrete will be impaired when the contents of zinc oxide and dodecyl dimethyl benzyl ammonium added in concrete are more than 0.05% [65, 76] . umar et al. [36] investigated the strength development of four types of cement composite modified with polymer/added inhibitor at the age of 7, 21, and 28 days. the results showed that compressive strength of cement composite admixed with sodium nitritebased inhibitor is increased by 26% (28 days) with respect to that of cement composite without any admixture, and higher than cement composite prepared with styrene acrylate copolymer and acrylic polymer, as shown in fig. 8 . vaquero et al. [16] obtained that the 28 d compressive strength of concrete mixed with multicomponent formulation pl-uv-h-2b was 37.1, 36.9, 35.7, and 34.9 mpa when the content is 0, 0.15, 0.2, and 0.3%, respectively, and the 28 d flexural strength was 9.4, 8.6, 8.2, and 8.5 mpa when the content is 0, 0.15, 0.2, and 0.3%, respectively. consequently, they concluded that the addition of pl-uv-h-2b in concrete only slightly decreased the compressive strength and flexural strength as compared to those of control samples [16] . moreover, do et al. [59] observed that compressive and flexural strengths of cement mortar containing the antifungal agent of isothiazoline/cabamate were almost equal to those of non-added cement mortar; hence, they concluded that the addition of isothiazoline/cabamate has a very little adverse impact on compressive and flexural strengths of cement mortar and is negligibly insignificant. researchers not only investigated the antimicrobial and mechanical properties of antimicrobial concrete, but also paid attention to its mass/weight loss. for example, negishi et al. [41] obtained that the weight loss of cement specimens without antimicrobial agents, with 0.075% metal nickel, and with 0.075% metal nickel plus 0.075% calcium tungstate was 10, 6, and 1%, respectively, after being exposed to a sewage treatment plant containing 28 ppm of h 2 s for 2 years. the least weight loss of nickel modified samples after adding calcium tungstate was due to the higher binding tendency of tungsten to a. thiooxidans. as it can be seen in fig. 9 , there is an apparent difference in mass losses in specimens with various bactericides and without adding any bactericide, the mass loss rate of concrete specimen with copper phthalocyanine was the lowest (4.78%) as compared to other specimens, providing evidence that copper phthalocyanine has the best effect on resistance to the microbial induced corrosion of concrete [62] . bao [67] reported that the mass loss of reference mortars and mortars added with mineral powder and fly ash was 1.26, 0.44 and 0.47% after an immersion in intensified sewage for 5 months, respectively. while the mass loss of mortar samples with antimicrobial agent sodium tungstate and sodium bromide reached 0.57% and 0.6%, which indicated that incorporation of admixture fig. 8 . comparison of compressive strength of (sar denotes styrene acrylate copolymer, ar denotes acrylic polymer and sn denotes sodium nitrite) [36] . fig. 7 . enumeration of fungi (cfu g à1 ) colonizing mortar probes after 15 months of outdoor exposure. l + s = lime + sand and p + l = pozzolana + lime. t0 represents untreated mortars; t1, t2, t3 represents mortar samples treated with different water-repellent alone; t4, t5, t6 represents mortar probes treated both with waterrepellent and biocide; t7 represents mortar probes treated with biocide alone. t4, t5, and t7 treated with algophase, and t6 treated with algophase ph 025/d [45] . has a better improvement effect than antimicrobial agents from the perspective of reduced mass loss. in addition, shook and bell [37] conducted an in-situ field test using concrete samples from concrete pipe in a sewer manhole which had evident corrosion occurring and an obviously high h 2 s concentration. they obtained that concrete samples treated without conshield had a great weight loss of 3.44%, whereas the concrete samples treated with conshield showed a significantly lower weight loss of 0.32% after 3 months. the antimicrobial mechanisms of heavy metal antibacterial agents towards microorganisms attached to and/or penetrated into concrete is generally considered to follow the reactions below. during the action of antibacterial agents, metal ions gradually dissolve and react with thiol group (-sh), amino group (-nh 2 ) and other sulfur nitrogen-containing functional groups existing in proteins and nucleic acids of bacteria, which inhibit or inactivate some necessary enzymes, and disturb the osmotic stability of the cell, thus achieving the antibacterial purpose [34, 51, 77] . more specifically, the action of silver ion released from the zeolite matrix in concrete and reactive oxygen species (ros) generated from silver within the matrix are considered as the mechanisms of bactericidal action of silver-loaded zeolites, and it has been reported that either the silver itself or the ros must interact with biological macromolecules like enzymes and dna by an electron release mechanism to maintain long-lasting antibacterial effect [63, 70] . it is assumed that nickel does not attack on bacteria themselves, but binds to an enzyme of bacteria to exhibit growth inhibitory effect [43] . nogami et al. [39] concluded that nickel ions incorporated into concrete bind to the plasma membrane and inhibit the activity of sulfur dioxygenase and sulfite oxidase of t. thiooxidans to exert its inhibitory effect. maeda et al. [40] also stated that nickel binds to t. thiooxidans cells and inhibits enzymes involved in sulfur oxidation of the bacterium, consequently inhibiting cell growth and sulfuric acid generation. similarly, tungsten exerts its antimicrobial effect on a. thiooxidans by binding to a. thiooxidans cells and inhibiting the sulfur oxidation enzyme system, such as sulfur oxidase, sulfur dioxygenase and sulfite oxidase of cells [41] . sugio et al. [42] also studied the mechanism of growth inhibition by tungsten in a. ferrooxidans, concluding that tungsten binds to cytochrome c oxidase in plasma membranes and inhibits cytochrome c oxidase activity, stopping cell growth from oxidation of fe 2+ . moreover, kim et al. [61] ascribed the antimicrobial mechanism of antimicrobial metals (ni and w) to the destruction of cell membrane or internal protein tissue of microbe by ni and w according to simulation tests. significantly increased surface area-to-volume ratio of nanoparticles contributes to greater interaction with microorganisms and enhances the release of toxic ions, assisting nanoparticles to achieve excellent antimicrobial properties [3, 78] . the multiple bactericidal mechanisms of nanomaterials, such as copper oxide and zinc oxide nanoparticles, have been attributed to damage of the cell membrane by either direct contact with nanoparticles or photocatalytic production of ros; release of toxic ions; interruption of electron transport, protein oxidation, and modification of membrane charges. degradation of dna, rna and proteins by ros, and lowering the production of atp due to acidification and ros production also accounts for bactericidal properties of nano-sized materials [3, 79] . fig. 10 illustrates a comparison of antibacterial mechanisms between antimicrobial nanomaterials and their bulk counterparts. in addition, the two major explanations to the photo-sterilization mechanism of concrete involving nano-tio 2 under light are the attack of chemical species leading to the death of microorganisms or the biological structure destruction causing the inactivation of microorganisms [55] . generally, organic antimicrobial agents inhibit the growth and reproduction of microorganisms by destroying cell membranes, fig. 10 . illustration of possible bactericidal mechanism of nanomaterials (bottom) compared to their bulk form (top) [3] . 9 . effect of various bactericides on mass loss of concrete immersed in sewage [62] . ddc, sbc, zoc, stc, cpc, and bc represent concrete incorporated with dodecyl dimethyl benzyl ammonium chloride, sodium bromide, zinc oxide, sodium tungstate, copper phthalocyanine, and plain concrete without bactericide, respectively. denaturing proteins, or disrupting metabolic processes. the phthalocyanine compound contained in concrete or mortar can be easily introduced into the cell of sob, inhibiting enzyme reaction within the cell and, eventually, killing sob [11] . in terms of copper phthalocyanine [62, 65, 76] , its high bactericidal property towards bacteria is mainly provided by copper ions. copper ions could interfere with the metabolic process of bacterial cells or interfere with the function of various enzymes, losing their biological functions and eventually leading to the death of cells [62, 65, 76] . quats, like dodecyl dimethyl benzyl ammonium chloride [62, 65] , the positively charged organic cations can be selectively adsorbed by the negatively charged bacteria contacting with concrete. they could enter into the cell membrane by permeation and diffusion, thus impede the semi-permeation action of cell membranes and then inhibit the generation of enzyme to achieve the sterilization effect [80] . mcdonnel et al. [81] proposed that quats target the cytoplasmic membrane and damage the phospholipid bilayer. additionally, the cellular membrane of bacteria will be pierced by the long molecular carbon chain of silane quaternary ammonium chloride (sqa) [57] and cell destruction will be triggered by ions exchange between the positively charged ammonium cation of sqa and ions within cell membranes, are two major hypotheses accounting for the antimicrobial working mechanisms of sqa. the antimicrobial mechanism of concrete with con-block mic [75] is the active ingredient in conblock mic 3-trimethoxy silyl propyl dimethyl octadecyl ammonium chloride has a positive charged nitrogen atom (as shown in fig. 11 ), electrostatically attracting many bacteria to the molecule. the molecular chain of carbon 18 atoms long pierces the cellular membrane of bacteria and the outer cell is punctured upon reaching the nitrogen atom. consequently, it creates an uninhabitable environment for the microbiological organisms on the surface of concrete [75] . as for conshield, it endues concrete with excellent antimicrobial effect through molecularly bonding to the ingredients of concrete mix, then, providing hundreds of microscopic spikes over an area of a single bacterium, which puncture the fragile single cell of bacteria [82, 83] . however, majority of the antimicrobial mechanisms mentioned above are relevant to inhibiting or killing bacteria, the antifungal and algaecidal mechanisms of corresponding antimicrobial agents used in concrete are rare, requiring further investigations. concrete is the most abundant material in wastewater systems but at the greatest risk for corrosion. despite most of the findings are based upon laboratory testing, there still exist some findings from practical applications of antimicrobial concrete. considering the superior antimicrobial property of concrete imparted by some typical antimicrobial agents, one of the major applications of antimicrobial concrete is to mitigate and control microbial corrosion caused by microbial metabolism in sewer systems, such as concrete sewage pipes, sewer manholes, wastewater collection systems and treatment plants, etc. for instance, in order to combat the growth and proliferation of thiobacilli in sewer systems, new sewer construction in atlanta has been utilizing concrete admixed with conshield since 1997, and rehabilitation works of concrete manholes in columbus, oh, oskaloosa co., fl, mt. prospect, il, miami, fl, and corsica, tx have adopted the same material [37] . the results shown in fig. 12 (a) and (b) clearly demonstrated the long-term protection due to the addition of conshield into concrete against microbial induced corrosion in the maline drop shaft [82] . owing to the proved high antimicrobial effectiveness, con-shield has a wide range of industrial applications in concrete structures mainly including two aspects: one is new and rebuilt concrete structures subjected to highly concentrated sulfide conditions like concrete pipe and manholes (fig. 12 c) ,wet wells, lift stations, wwtp head works, clarifiers, and the like. another one is rehabilitation of heavily corroded manholes, pipelines and tunnels in place via shotcrete (fig. 12 d) [83]. similarly, with excellent antimicrobial power and long-lasting antimicrobial effect, concrete incorporated with antimicrobial additive zeomighty (zeolitesupported silver and copper) was popular in the japanese market. the practical applications of antimicrobial concrete with zeomighty include secondary concrete products such as hume pipes, manholes, and box culverts, cast-in place concrete structures for sewer and treatment facilities and other premix mortar, etc., as shown in fig. 13 [33]. kurihara et al. [84] invented an antibacterial agent composed of a silver compound (selected from silver carbonate, silver oxide and silver phosphate), a copper compound (selected from copper carbonate, copper oxide, copper phosphate and copper hydroxide) and an ion-retaining compound, and concrete containing the antibacterial agent exhibits outstanding antibacterial effect against srb, sob, and carboxylic acidproducing bacteria particularly in sewage treatment plants. uchida et al. [11] disclosed that the addition of phthalocyanine compound (a metal phthalocyanine, a metal-free phthalocyanine, and derivatives thereof) in concrete or mortar can be easily introduced into a cell of sob, thus inhibit and/or kill sob via inhibiting enzyme reaction within the cell of sob. consequently, the deterioration inhibitor with the effective component, phthalocyanine compound, showed ability to mitigate the deterioration of concrete or mortar. antimicrobial concrete fabricated with copper phthalocyanine [62, 65] has the merits of excellent bactericidal performance, high retention rate of bactericide and low cost. moreover, the addition of copper phthalocyanine does not affect the performance of concrete. consequently, such antimicrobial concrete can be widely used in the construction of municipal sewage facilities [85] . moreover, it is stated that the antimicrobial additive conblock mic can be applied in new concrete infrastructure and cementitious infrastructure repair products, for example, concrete pipe, manhole and septic tanks, or for ready mixed concrete or cementitious mortars and liners [75] . with the advantages of long-lasting bactericidal effect on sob (one to several years), the low cost and environmentally friendly chemical (i.e. nitrite), fna spray [44] is a promising practical technology for mitigate and control of microbially induced concrete corrosion. in addition, according to [86] , concrete added with copper oxide (methyl cellulose as dispersant) and zinc oxide (fly ash as dispersant) was proved to be able to protect marine ecological engineering construction from microorganism attack. compared to the untreated concrete columns with a number of plaques found on the surface, no evidence of plaque was found on the surface of three treated concrete columns after 18 months. similarly, concrete with tio 2 , utilizing the light-induced bactericidal activity of tio 2 , can be employed to control microbiological growth on concrete surfaces, thus enhancing the durability of concrete in ocean engineering. the same concrete can be also used as exterior wall materials of buildings, achieving sterilization function by decomposing bacteria attached on surface [49, 87] . janus et al. [88] proposed that concretes admixed with modified titania, with enhanced antibacterial properties, can have a wide application in places demanding high sterilization levels, such as hospitals, institutions, school and water storage tanks. in addition, freed et al. [56] disclosed that antimicrobial concrete reinforced with fiber carrying antimicrobial agents, such as microban b, has the ability to protect concrete from biological attack. the antimicrobial agent is first incorporated into or coated onto fibers and then the treated fibers are admixed with concrete. such antimicrobial concrete, with the ability to inhibit growth and contact of microorganisms such as bacteria, fungi, mold, etc., aims to be employed in areas requiring extraordinary cleanliness such as food processing plants, hospitals, kitchens, locker rooms, and the like. microbial attachment, colonization and eventually deterioration have been a great threat to concrete structures in sewer sys-tems, marine environments, buildings exposed to high humidity and the like. antimicrobial concrete, with the addition of inorganic or organic antimicrobial agents, exhibits excellent antimicrobial effect against specific microorganism and helps to address such issues caused by microorganism metabolism. also, the appearance of antimicrobial concrete makes infrastructures smarter and more durable, prolongs the service life of infrastructures and lowers the huge cost by rehabilitation and even replacement. despite many investigations have been conducted in this area in the past decades, there still remains some key issues to be addressed. the relationship between antimicrobial property and various affecting parameters (including contents, retention rate and dispersion, etc.) should be further comprehensively investigated so as to effectively enhance the antimicrobial effect of antimicrobial concrete. combining different antimicrobial agents to form biocide formulation according to their respective intrinsic properties may be a promising strategy to boost antimicrobial efficiency. the toxicity due to the release of some active ingredients into the environment during the entire service life of inorganic antimicrobial agents such as nanoparticles and generally temporary effectiveness for organic antimicrobial agents are impediments to the widespread application of antimicrobial concrete. moreover, the resistance of microorganisms to antimicrobial agents has to be considered in developing antimicrobial concrete. currently, most researches are restricted to the laboratory stage, practical applications are few and field trails are still highly required to verify the feasibility of antimicrobial concrete with aforementioned antimicrobial agents. the development of antimicrobial concrete is based on the advancement of antimicrobial agents. in future, it is expected to provide novel, high-efficiency, long lasting, broad-spectrum and environmental-friendly antimicrobial agents for fabricating antimicrobial concrete. in addition, antimicrobial concrete with its exceptional antimicrobial performance may have an extended application in the field of fighting against viruses. especially, the world is in novel coronavirus pandemic now. countries around the world are building new hospitals or improving the facilities of existing hospitals to better treat infected patients. additionally, following its detection in the sewers in massachusetts, the novel coronavirus was also reported to be found in the non-potable water system used for cleaning streets and watering parks in paris. if the infrastructures such as hospitals and sewage systems have the ability to kill viruses, it is beneficial for preventing the spread and reproduction of viruses. furthermore, the combination of new technologies may promote the development of antimicrobial concrete, such as nanotechnology, geopolymer technology, 3d printing/digital production technology, biotechnology, self-assembly technology, damage and failure evaluation technology, organic-inorganic composite technology and multiscale simulation technology [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] . the authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. isolation and characterization of microorganisms involved in the biodeterioration of concrete in sewers reviewed of antibacterial concrete research and application status biogenic deterioration of concrete and its mitigation technologies enhancing antimicrobial properties of fly ash mortars specimens through nanophase modification microbial ecology of crown corrosion in sewers interactions of nutrients, moisture and ph on microbial corrosion of concrete sewer pipes mechanics of corrosion of concrete sewers by hydrogen sulfide microbiologically induced deterioration of concrete: a review the corrosion of concrete: 1. the isolation of a species of bacterium associated with the corrosion of concrete exposed to atmospheres containing hydrogen sulphide the corrosion of concrete: 1. the isolation of a species of bacterium associated with the corrosion of concrete exposed to atmospheres containing hydrogen sulphide microbial community structures and in situ sulfate-reducing and sulfur-oxidizing activities in biofilms developed on mortar specimens in a corroded sewer system deterioration preventive for concrete or mortar and method for preventing deterioration of concrete or mortar a newly isolated fungus participates in the corrosion of concrete sewer pipes biodeterioration of concrete by the fungus fusarium splash zone concrete of marine concrete engineering by sem and surface microorganism identification by 16s rrna microorganism identification by 16s rdna of concrete surface exposed to a tidal zone development and experimental validation of an overlay mortar with biocide activity corrosion mechanism and control technology of microorganism participation in concrete anatase as an alternative application for preventing biodeterioration of mortars: evaluation and comparison with other biocides cement-based biocide coatings for controlling algal growth in water distribution canals antimicrobial activity of commercial concrete sealant against salmonella spp: a model for poultry processing plants comparison of corrosion resistance mechanism between ordinary portland concrete and alkali-activated concrete subjected to biogenic sulfuric acid attack study on anti-bacteria concrete present and prospect of microbial corrosion prevention of concrete silver supported on natural mexican zeolite as an antibacterial material antimicrobial mortar surfaces for the improvement of hygienic conditions evaluation of the bactericidal characteristics of nano-copper oxide or functionalized zeolite coating for biocorrosion control in concrete sewer pipes zeolite-supported silver as antimicrobial agents evaluation of the resistance of mortars coated with silver bearing zeolite to bacterial-induced corrosion inhibition of microbial concrete corrosion by acidithiobacillus thiooxidans with functionalized zeolite-a coating the inhibitory effect of antimicrobial zeolite on the biofilm of acidithiobacillus thiooxidans experimental research on new type antibacterial high-performance concrete the antibacterial experiment study based on durability of concrete resistance mechanisms of bacteria to antimicrobial compounds mechanism and research approach of microbial corrosion of concrete modified cement composites for protection against microbial induced concrete corrosion of marine structures corrosion control in concrete pipe and manholes corrosion by bacteria of concrete in sewerage systems and inhibitory effects of formates on their growth inhibition of sulfur oxidizing activity by nickel ion in thiobacillus thiooxidans nb1-3 isolated from the corroded concrete nickel inhibition of the growth of a sulfur-oxidizing bacterium isolated from corroded concrete growth inhibition by tungsten in the sulfur-oxidizing bacterium acidithiobacillus thiooxidans mechanism of growth inhibition by tungsten in acidithiobacillus ferrooxidans thiobacillus thiooxidans growth inhibitor. patent us 6146666 a novel and simple treatment for control of sulfide induced sewer concrete corrosion using free nitrous acid evaluation of the efficiency of water-repellent and biocide compounds against microbial colonization of mortars evaluation of strategies to prevent algal fouling on white architectural and cellular concrete photocatalytic, hydrophobic and antimicrobial characteristics of zno nano needle embedded cement composites the experimental research on antimicrobial properties of highperformance concrete with nano-zinc oxide multifunctional cementitious composites modified with nano titanium dioxide: a review photocatalytic effect of nano-tio 2 loaded cement on dye decolorization and escherichia coli inactivation under uv irradiation progress of antibacterial materials the novel inorganic antibacterial agent silver molybdate antimicrobial properties of nanomaterials used to control microbial colonization of stone substrata antimicrobial activity of al 2 o 3 , cuo, fe 3 o 4 , and zno nanoparticles in scope of their further application in cement-based building materials concrete, mortar and plaster using titanium dioxide nanoparticles: applications in pollution control, selfcleaning and photo sterilization reinforced concrete containing anti-microbial enhanced fibers evaluating the efficacy of antimicrobial additives against biogenic acidification in simulated wastewater exposure solutions chapter 3-an introduction to microbial corrosion antifungal effects of cement mortars with two types of organic antifungal agents study on preparation and properties of silver nanocomposites doctoral thesis evaluation of properties of concrete using fluosilicate salts and metal (ni, w) compounds effect of bactericide on the deterioration of concrete against sewage research and development of zeolite antibacterial agent experimental study of antibacterial high-performance concrete study on the applicability of bactericides to prevent concrete microbial corrosion sulfite oxidase activity in thiobacillus novellus experimental and prediction study on the deterioration behavior of concrete in municipal sewage master thesis the effects of silica/titania nanocomposite on the mechanical and bactericidal properties of cement mortars effectiveness of admixtures, surface treatments and antimicrobial compounds against biogenic sulfuric acid corrosion of concrete mechanistic study of the bactericidal action of silver-loaded chabasite on acidithiobacillus thiooxidans antimicrobial potential of zno, tio 2 and sio 2 nanoparticles in protecting building materials from biodegradation photocatalytic inhibition of algae growth using tio 2 , wo 3 , and cocatalyst modifications the bio-erosion resistance of concrete application and optimization of bactericide in concrete under sewage environment master thesis in situ assessment of active thiobacillus species in corroding concrete sewers using fluorescent rna probes antimicrobial activity of metal oxide nanoparticles against gram-positive and gramnegative bacteria: a comparative study the toxic effects and mechanisms of cuo and zno nanoparticles current state and development of quaternary ammonium salt bactericides antiseptics and disinfectants: activity, action, and resistance twenty years of protecting concrete in sewers antibacterial agent for concrete, concrete compositions and concrete products a type of antimicrobial corrosion concrete concrete material and preparation method for antibacterial and anticorrosive marine ecological engineering bacterial inactivation on concrete plates loaded with modified tio2 photocatalysts under visible light irradiation performance and properties of mortar mixed with nano-cuo and rice husk ash influence of different types of nanosilica hydrosols on properties of sustainable white cement mortar effect of nano-particles and aminosilane interaction on the performances of cement-based composites: an experimental study intrinsic self-sensing concrete and structures: a review multi-phase modelling of electrochemical rehabilitation for asr and chloride affected concrete composites nano-core effect in nano-engineered cementitious composites a new index to evaluate exposure and potential damage to rc building structures in coastal areas research and practice on progressive collapse and robustness of building structures in the 21st century learning from failures in an emblematic ionic transport features in concrete composites containing various shaped aggregates: a numerical study bacterial technology-enabled cementitious composites: a review key: cord-276616-odmnvv7m authors: darcel, c. title: reflections on scrapie and related disorders, with consideration of the possibility of a viral aetiology date: 1995 journal: vet res commun doi: 10.1007/bf01839302 sha: doc_id: 276616 cord_uid: odmnvv7m the transmissible spongiform encephalopathies of domesticated animals, scrapie in-sheep and bovine spongiform encephalopathy (bse), and transmissible mink encephalopathy are more than a scientific curiosity; under certain circumstances their impact on commercial activities can be calamitous. knowledge of their causation and pathogenesis is still rudimentary, but many consider than an unconventional agent, the prion (a brain protein, prp), that is not associated with nucleic acid is involved in both. others believe that conventional viruses, which replicate by virtue of their nucleic acid-defined genes, are involved in the causation and progression of the encephalopathies but that technical problems have prevented their identification. others postulate even more exotic causative agents. while this paper will particularly address the possibility of a viral aetiology for these diseases, it is also emphasized that our knowledge of the state of the immune system in animals with encephalopathy needs broadening. there are remarkable gaps in our knowledge of the histopathology of these diseases, particularly the nature of the characteristic vacuoles. much further work is needed on the biochemical changes in the brain and the serum, particularly of the latter as it could lead to an additional means of recognizing clinical cases without waiting for the animal to die with subsequent examination of the brain for characteristic lesions and the presence of protease-k-resistant prp. in the transmissible spongiform encephalopathies (kimberlin, 1990) , a variety of neurological signs follow non-inflammatory, vacuolar, degeneration in the brain and spinal cord. accumulations of amyloid and other proteinaceous fibrils occur. these fibrillar accumulations can form plaques similar in appearance to those seen in alzheimer's disease in man. after a very long quiescent latent period following infection, the spongiform encephalopathies progress very slowly over a period of weeks to months. this group of diseases provides a good example of disastrous events in the field of veterinary medicine that have put into question all our current concepts of the genesis of disease. these events have also had the positive effect of focusing attention on enormous gaps in our knowledge and stimulating a great deal of research. reference to older literature on scrapie can be made in books edited by kimberlin (1976a) and prusiner and hadlow (1979) . the advent of bse in britain, starting in the 1980s, resulted in the appearance of numerous articles on the significance of the spongiform encephalopathies. these include a number of more recent reviews (little and thorsen, 1989; chesebro, 1991; powell, 1993) . scrapie, a disease of sheep, was the first of these diseases to attract enough attention to stimulate research into its causation and recognition. the disease had been recognized for some time but it did not become the subject of experimental work until the 1930s, when cuiu6 and chelle, cited by gordon (1946) , reported on its transmission. its accidental transmission during large-scale field trials of vaccination of sheep against louping-ill also gave impetus to research on this disease 1.. the vaccine had been prepared from formaliuized brain and spleen suspensions from louping-ill-infected yearling sheep (gordon, 1946) . starting 2½ years after vaccination, scrapie started to appear in animals of all breeds vaccinated with one lot of vaccine. this unfortunate experience prompted transmission studies designed to determine the nature of the infective agent and the pathogenesis of the disease in sheep. conclusions drawn from the vaccination trials and transmission experiments were that ~crapie, given by subcutaneous inoculation, had a latent period of 2 years and longer; that the infective agent was resistant to 0.35% formalin; that the disease appeared more quickly and in a higher percentage of recipients following intracerebral than following~subcutaneous injection; and that the causative agent was probably a filtrable virus. a similar disease to scrapie is bovine spongiform encephalopathy (bse), of which the first ca~e of a major enzootic was confirmed in november 1986 in england (bradley and matthews, 1992) . the emergence of this disease was considered to be related to the ingestion of meat and bonemeal derived from ruminant sources following countrywide changes in the rendering process (wilesmith et al, 1991) . the feeding of ruminant-derived meat and bonemeal to cattle in britain was stopped by law in 1988 . since the implementation of this ban, the occurrence of new cases of this disease in cattle has been reduced (wilesmith and ryan, 1993) . the bse outbreak drew attention to spongiform encephalopathies that occurred around the same time in other bovids, kept in zoological and wildlife park collections in britain cunningham et al, 1993) . these cases could have developed through the animals being fed similar rations to the cattle population. spongiform encephalopathy has also been found in mule deer and captive elk in north america young, 1980, 1982; wells and mcgill, 1990) , evidence that agents causing encephalopathy are present in wild ruminants in that continent. • spongiform encephalopathy has recently been demonstrated in domesticated and wild felids (wells and mcgill, 1990; willoughby et al, 1992) . *please see notes section at end of text. transmissible mink encephalopathy (tme) (marsh and hanson, 1969; marsh, 1976) , which is rarely observed in ranch-raised animals, is caused by an agent with the same physicochemical properties as the scrapie agent. in the first outbreaks of tme, it could not be excluded that the occurrence of the disease had followed feeding of meat and offal from sheep, but the feeding of sheep meat was not involved in a more recent and major outbreak of tme on a ranch in the united states (marsh et al., 1991) . three diseases are associated with spongiform encephalopathy in man: kuru, creutzfeldt-jakob disease (cjd) and gerstmann-straiissler-scheinker disease (gssd). kuru, which is very similar to scrapie, is befieved to be associated with funeral practices involving ritualistic cannibalism (gajdusek, 1977) . a very full account of this and other human spongiform encephalopathies and follow-up experimental work is given by brown and gajdusek (1991) . cjd usually occurs sporadically at a very low incidence but deslys and colleagues (1994) refer to a group of 25 cases of cjd that occurred in france among children treated between january 1984 and june 1985 with human growth hormone (hgh). the hormone had been extracted from large numbers of cadaveric hypophyses. the cases were confined to individuals with the same genetic status as sporadic cases of cjd, who were homozygous at codon 129 of the prp gene. gssd shows a clinical overlap between kuru and cjd but is familial and related to a point mutation on chromosome 20, which encodes an amyloidogenic protein prp (brown, 1990) . as in scrapie and bse, the causative agent of cjd can be transmitted to experimental animals (gibbs et al, 1979; manuelidis and manuelidis, 1979; tateishi et al, 1979) . thus spongiform encephalopathies have been described in several species, and in most of them the disease has been shown to be transmissible to animals of the same species, where this is ethically possible, and to certain laboratory animals. the purpose of this paper is to give a condensed overview of the aetiology, pathology and familial nature of some of the spongiform encephalopathies as well as aspects of diagnosis and serology; it will particularly address the possibility of a viral aetiology for these diseases. there are many difficulties in studying either the natural or experimentally induced diseases: the animals involved, the incubation period required for the emergence of the disease, the innate resistance of a proportion of the population seen as an expression of genetic influences, the differing behaviour of strains of agents isolated from a given species, the symptomatology, the pathology, the uncertain nature of the agent and its means of transmission, the perceived 'lack' of an immunological response or changes in the immune system, and the biological hazards involved in conducting experiments. none of the infective agents that cause the spongiform encephalopathies can be demonstrated except by animal inoculation. sheep were the first experimental animals used for research on this group of diseases (scrapie) and the intracerebral route of inoculation was found to be better than the subcutaneous route for obtaining the maximum incidence of disease. there are ethical, logistical and humanitarian considerations against using large animals for this type of study. the finding that the spongiform encephalopathies were transmissible to certain laboratory animals (mice, hamsters and laboratory primates) was a critically important advance. it then became possible to titrate the infectivity of spleen, brain and other tissues in small laboratory animals, which was essential if any real advance in our understanding of the causative agents was to be obtained. however, a problem with the use of laboratory animals is that they may carry other disease agents in a latent form. the presence of such agents might even determine whether or not scrapie appears following inoculation with infected material and also the progress of the disease. this possibility has attracted attention, and 'caesarian-derived, barrier-maintained colonies have been the rule for most serious work in small laboratory animals. neoplasia, however, remains a problem in some colonies of laboratory rodents' (r.m. barlow, personal communication, 1994) . the need to hold and observe animals for long periods of time is a great inconvenience in studies on natural and experimental transmission of the spongiform encephalopathies. in bse, for instance, some cases are still emerging that may reflect exposure to the agent in the early 1980s. in ovine experiments, the 4v2-year period of observation noted in the early scottish transmission experiments was not excessive, although enough data can usually be collected within a 2-year period in sheep. in mice, scrapie infection progresses more quickly, and if 4-5-week-old animals are inoculated intracerebrally with large amounts of the agent most of them will develop the disease within 10 months (chandler, 1961) ; in hamsters the disease may proceed even more rapidly (kimberlin and walker, 1977) . contributing to the incubation period is the phenomenon that has been termed the 'zero phase' (dickinson and outram, 1979) . this is a period after infection when no infectivity can be demonstrated from such organs as the brain, spleen and fiver. the severity and type of lesions vary, and a quantitative assessment of these can be used as a characteristic of the biological behaviour of scrapie isolates. transmission experiments in inbred mice have shown that there are mixtures of agents in a natural case of scrapie in a sheep, and as many as 15 distinct stains have been so identified. at least four murine strains of scrapie have been derived from the 'drowsy goat' source, the result of scrapie transmission experiments in goats (bruce and dickinson, 1979) . such a spectrum of behaviour by different isolates of field cases would be expected in studies with the spongiform encephalopathies of other species, especially if they truly represent passages of the scrapie agent. the agent causing bse is transmissible to mice and can be orally transmitted to c57b1 mice by feeding bse brain (barlow and middleton, 1991) . however, the nature and distribution of pathological lesions in experimentally infected mice were similar even when the source of the infection was different species of bovidae , and there was no evidence of strain variation. although transmissible encephalopathy is not observed naturally in pigs, it can be produced experimentally in this species ) by transmission of bse. differently located neural lesions and different signs from those seen in tme were observed in mink receiving brain tissue from bse-infected cattle. transmission was effected both parenterally and orally (robinson et al., 1994) . in transmission experiments with cjd and kuru, chimpanzees and squirrel monkeys are the most sensitive, and are uniformly susceptible, but the disease has also been transmitted to laboratory animals. guinea-pigs died quite rapidly of hydrocephalus, which did not appear to be due to the presence of any additional conventional virus in the inocuhim. hamsters seemed uniformly susceptible and developed paralysis of the hind limbs and skin hypersensitivity. mice that were infected with guinea-pig-derived cjd developed signs of skin hypersensitivity and locomotor problems after a long latent period (manuelidis and manuelidis, 1979) . not all transmissible spongiform encephalopathies are due to scrapie-like agents. mice, homozygous for the grey tremor trait, develop a transmissible encephalopathy that is not accompanied with the coat pigmentation, the seizures, tremors and ataxia characteristic of the clinical disease in the donor animals. further, brain homogenates from these mice do not show the presence of proteinase-k-resistant prpsc (bendheim et al, 1988) . 'most neurological diseases follow the same general sequence of events: excitement -, hyperaesthesia -~ dementia ~ dullness -* coma -, death; incoordination -~ ataxia -, paresis -, paralysis. it is the duration of each phase which varies with the cause and the neural centres critically damaged' (r.m. barlow, personal communication, 1994) . the signs seen in scrapie, which are not specific, reflect damage to the neural centres, following the two alternative progressions noted above. for instance, with reference to the louping-ill vaccination trials in the 1930s, louping-ill itself is also characterized by muscle tremor and incoordination of the hind limbs. there may be hypersensitivity to noise and touch. unlike scrapie, the disease progresses more rapidly. death follows within 7-10 days of signs of developing. some of the signs seen in scrapie show points of resemblance. in scrapie, there are also signs of increased skin reactions to pressure and other stimuli. there is a characteristic 'nibbling' reaction and a pruritus manifested by the sheep rubbing, leading to loss of wool over the rump, thighs and tail base. there is impairment of locomotion, starting with problems in the hind limbs (blood et al., 1983) . although scrapie is a far more chronic disease than louping-ill, it is curious that variations of hyperaesthesia are observed in both these diseases, with very different aetiologies. in bovine spongiform encephalopathy (bse), hypersensitivity to touch and sound are also common signs, but there are many other subtle changes in behaviour. however, cases of disease with these very signs, but which are not bse in terms of their histopathology, can occur in cattle (jeffrey and wilesmith, 1992) . it has also been suggested (gibbs et al, 1990 ) that bse may have occurred in the united states under the clinical guise of the 'downer cow' syndrome, even though cattle disease due to bse has not been reported in that country. these workers attempted experimental transmission of the scrapie agent in brain homogenate from sheep or goats to 10 cattle. three cattle developed neurological signs 27-48 months after inoculation, consisting of progressive difficulty in rising, stiff-legged stilted gait, incoordination, disorientation and terminal recumbency. from onset to terminal signs, the disease lasted 1-2.5 months. there were resemblances to the 'downer cow' syndrome. while autopsy revealed insufficient changes to confirm a clinical diagnosis of bse, the protease-resistant protein prp27-30 was detected in each of the 3 animals with neurological signs; none was detected in the other 7 animals. if the results of this experiment can be replicated, the consequences are serious. the 'downer cow' is a familiar one to veterinarians and, while the usual aetiology is a ca/p imbalance, successful repetition of this experiment would mean that, where that imbalance cannot be demonstrated, the presence or absence of bse will have to be verified. the signs and symptoms that follow the onset of the human spongiform encephalopathies show some differences between types: in kuru there is a fine tremor of the head, trunk and limbs, together with the onset of a progressive ataxia; in cjd the symptoms are of dementia and, instead of the fine tremor, there is a myoclonic jerking; in gssd there is an overlap with the signs of kuru and cjd and there is also a marked dementia. the type of hypersensitivity associated with scrapie and bse is absent. returning to the pruritus seen in scrapie and bse, there seems to have been no investigation of the cause of the irritation. one must question whether the hyperaesthesia is evidence solely of nerve irritation or whether it is a pointer to a development of the hypersensitivity or skin lesions seen in some virus infections (fenner et al., 1974) or even related to the local release of histamine or a related compound. there are no immunological tests to aid recognition of the diseased living animal. clinical signs are not sufficiently specific to diagnose disease and diagnosis depends on the examination, by experienced pathologists, of stained sections of brain tissue for characteristic lesions (wells and mcgiu, 1990) , electron microscopy to detect the characteristic saf fibrils ) and tests to detect prpsc, the prion protein characteristically found in the spongiform encephalopathies (cho, 1986; rubenstein et al, 1986) . this latter protein is characteristically resistant to proteinase-k. prpsc contributes to the formation of the amyloid aggregates or plaques seen in the brain of animals with scrapie. prpsc is encoded by the prp gene, which expresses a similar protein prp-c (sensitive to the action of proteinase-k) in uninfected tissues. in the past, one histological characteristic of the brain lesions seen in spongiform encephalopathies drew much interest, this being the occurrence of large vacuoles outside the outer wall of neurons. no good explanation has yet been found for these vacuoles, which invariably appeared empty by both fight and electron microscopy -the larger vacuoles appearing to have formed by a coalescence of smaller vacuoles. although interesting, these vacuoles are not pathognomonic, as similar vacuolation has been seen in the cerebrum of mice infected with a parainfluenza virus (baringer et al, 1979) and also in lymphoma-prone wild mice (gardner et al., 1979) . nevertheless, such vacuoles are a good diagnostic aid. as 'empty' inclusions, they presumably represent unstainable material, fluid, gas or volatile hydrocarbon 2. the possibility of these vacuoles representing unstainable material was addressed many years ago by darcel and colleagues (1961) . these authors were aware of two familial diseases in man: castaigne-lhermite's and wilson's diseases. in both diseases, the patient's behaviour shows some resemblance to certain signs seen with the spongiform encephalopathies, namely a motor disability in the first disease, a progressive rigidity, a coarse tremor of the limbs, and some degree of dementia in the second. in castaigne-lhermite's disease, histological examination of the nervous system reveals the accumulation of glycogen within the nerve cells, varying from small droplets to large masses. in wilson's disease, where the onset is most common in the second decade of life, copper accumulates in large amounts in the corpus striatum and in other tissues. while there are differences in the clinical signs and the histopathology shown by both diseases from those seen in scrapie, it was thought worthwhile to determine the glycogen and copper levels in the brains of sheep with experimentally induced scrapie. the levels of neither copper nor glycogen were elevated compared to the values found in the brain of normal sheep 3. with respect to the vacuoles representing fluid accumulations, this is judged unlikely because body fluids, even cerebrospinal fluid, contain protein which is stainable by routine histological procedures. we are left with the possibility that these vacuoles result from the accumulation of gas in the form of bubbles or the presence of inclusions of volatile hydrocarbons. the most likely gases to be released would be oxygen or carbon dioxide. the latter would be taken up by the buffering action of the tissues, while oxygen free radicals are released by enzymes such as the o 2-and h20-forming oxidases 4. in the blood, molecular o 2 would be quickly bound by haemoglobin in the red cells, but most locations in the brain are beyond the 'blood-brain barrier' and, there, oxygen would be poorly soluble and quite toxic. with regard to the possibility of the vacuoles being due to volatile hydrocarbons, both ethane and pentane are released from cells when there is peroxidation of polyunsaturated lipids (miiller and sies, 1984) . efforts to detect evidence of such peroxidation in brain tissue from cases of transmissible encephalopathy have yet to be made. although hunter (1979) lists a number of changes in the brain with the progress of scrapie, neither this nor a survey of the literature shows any indication that the possibility that these vacuoles are gas inclusions has been vigorously researched. resolution of this problem could have led to a more directed approach to the biochemical changes that occur in the encephalopathies. biochemical changes that have been noted include changes in amino acids in the blood plasma and brain tissue but, except for great increases in the activity of hydrolytic enzymes, evidence of other biochemical alterations in the brain, plasma and urine is scanty (kimberlin, 1976b) . liver function is normal . however, there has been the recent observation in one case of bse of a very high plasma glucose concentration (scott et al, 1988) . more data are needed as consistently high blood glucose could point to other biochemical investigations that should be made. millson and bountiff (1973) found a group of five glycosidases that showed greatly increased activity in the brain of mice with clinical signs following inoculation of the agent, four of these showing an increase before clinical signs appeared. one wonders whether such increased activity would spill over into the plasma and be usable for an early diagnostic test for the development of a spongiform encephalopathy. scrapie is transmissible, but the causative agent has not been shown to be a virus. one of the difficulties in accepting a conventional virus as the agent has been its resistance to damage with a variety of agencies. in spite of this, a viral aetiology has not yet been excluded. the possibility has recently been advanced that mollicutes (spiroplasmas) are responsible (see below), but most current thinking envisages a different kind of transmissible agent. two such agents have been postulated -'prions' and 'virinos'. the 'priori' was postulated to be an infectious protein that directs its own replication (prusiner, 1982) . the 'virino' hypothesis suggests the existence of a very small scrapie-specific nucleic acid (dickinson and outram, 1988) . evidence for the priori and virino and other possible hypotheses for the causation of the encephalopathies will now be considered. the variety in the infective agents that can cause disease is awe-inspiring, particularly with regard to the way they are propagated, retain their characters, and cause problems for the host. with bacteria and viruses, the infective agent carries its own genetic material in the form of either rna or dna, but a decade ago it was suggested that an altered protein can itself transmit the encephalopathies. this is the prion prpsc (prusiner, 1982) , or scrapie-associated fibril protein (saf), which is present as fibrils or rods, 33-35 kda in size, that can be seen by electron microscopy intracellularly and extracellularly in the brain of animals affected by scrapie. prpsc is heterogenous on electrophoresis, although this may be due to different degrees of glycosylation of this sialo-glycoprotein (somerville and ritchie, 1990 ). species differences have been found in bovine, murine and ovine prpsc (groschup and pfaff, 1993) . the homozygous prp genotype has been found to predispose to sporadic cad (palmer et al., 1991) and a prion protein missense variant is linked to the occurrence of gssd (hsiao et al., 1989) . the possibility that protein can be infectious in the absence of an association with nucleic acid is intriguing. prion protein and scrapie-associated fibrils are especially characteristic of brain lesions in scrapie (merz et al, 1981; bolton et al, 1982) . however, although hamster and mouse spleen contain only low levels of prp mrna , they can contain large amounts of the transmissible agent. this is hardly good evidence that the prion is the infective agent. kitamoto and colleagues (1989) were able to demonstrate a correlation between titres of the cjd agent in the brain and spleens of infected mice and their content of prp, while czub and colleagues (1986) were not able to do so for scrapie. unfortunately, it is difficult to resolve this question, since titration of the scrapie agent is attended with wide errors (marsh et al, 1984) and the specificity of serological methods for determining saf/prpsc is open to question, since prpsc is produced by the same gene as its homologous normal protein, differing only in its sensitivity to proteinase-k. another fact suggesting that the 'prion' is not the infective agent for scrapie is that prpsc expressed in vitro by a cloned prp gene is ineffective in inducing the disease. when the hamster gene for expressing this protein is introduced into mice, this species now has the susceptibility to scrapie characteristic of hamsters rather than the susceptibility of mice; the transgenic mice still have to be infected with the scrapie agent (scott et al, 1989) . other evidence indicating that the 'prion' or saf is not the causative agent of scrapie is given by braig and diringer (1985) . finally, bueler and colleagues (1993) showed that mice bred for lack of expression of the prp gene are resistant to the development of scrapie. after the introduction of syrian hamster prp transgenes, these mice became highly susceptible to hamster but not mouse prions. again, it has been found (xi et al, 1992; demaimay et al., 1994) that, following treatment of hamsters with amphotericin b or a derivative of this antibiotic, the accumulation of prp was not correlated with replication of the scrapie agent. race and colleagues (1988) described experiments with cloned neuroblastoma cultures with a high titre of the scrapie agent. while these workers acknowledged that a proteinase-k-resistant protein developed with scrapie infection in these experiments, they found no evidence that the pk-resistant form of prp was necessary for infectivity. they concluded that the pk-resistant prp of scrapie arose in vivo as a result of alteration of pk-sensitive prp by changes in the physiological conditions of cells as they degenerate, and that the evidence for the inseparability of the scrapie agent and pk-resistant prp was still circumstantial, a view not shared by bolton and colleagues (1991) . caughey and colleagues (1991) speculated, from further work on these neuroblastoma cultures, that the increased tendency of prp to aggregate is the origin of this resistance to pk. lewin (1982) was ready to accept that the prion is involved in the aetiology of scrapie but only as a carrier of a small protein or peptide that he called a protovirin. he suggested two alternative mechanisms by which such an entity could act either as an rna template or as an extraneous dna operon. it appears that it is very difficult to resolve the problems of whether prpsc is the causal agent of scrapie or a carrier agent, and even whether the causal agent is a protein, by the approaches that have been so far made. whatever the outcome, it is certain that prp plays an important part in the pathogenesis of the transmissible spongiform encephalopathies (weissmann, 1991a,b) . the discussion on the role of prions is taken up again in other sections of this paper. the virino hypothesis suggests that the infectious agent of scrapie is a very small piece of nucleic acid with a behaviour intermediate between that of viruses and that of viroids (dickinson and outram, 1988) . it has been further suggested that the virino is not translated and is therefore dependent on host-coded protein to form an infectious unit (kimberlin, 1990) . with the nucleic acid probe techniques now available, it would be thought that virinos would be detectable if they exist, but no report of such work has yet appeared. characteristic features of spongiform encephalopathies include their long latent period for development of overt disease and the need for the presence of 'susceptibility' genes and the prp gene in the prospective host. this does not rule out the involvement of a conventional virus in causation of these diseases. although no such virus has yet been demonstrated in infective extracts, many believe that such viruses may be involved. chesebro (1992) , reviewing the role of prp in the aetiology of scrapie, raised the possibility that a mutation of this gene increases the susceptibility to an unknown but ubiquitous conventional virus. rohwer (1991) summarized in a masterful way why conventional viruses may still be the most likely causes of the spongiform encephalopathies. for example, the long latent period needed for the spongiform encephalopathies to develop is similar to the length of time often required by lentiviruses such as the maedi-visna virus of sheep. however, no connection has been made between such viruses and scrapie. a long latent period for disease to develop is seen with other viruses, for example oncornaviruses. examples are the lymphoid leukosis viruses of chickens; these are noted for their ability to activate the formation of oncogenes. oncogenes are involved in turnout formation because their presence leads to perturbation of the normal cell physiology. scrapie is not characterized by the development of tumours but oncogenes could still be involved, tied as they are to normal cell physiology and to gene changes. studies similar to those now being made in oncology might be most fruitful. references should be made to oncogenes in the human genome, where as much as 0.1-0.6% may be derived from endogenous sequences. one such sequence is the erv-3 virus, first recognized by low-stringency hybridization to probes derived from the pol region of the chimpanzee ch2 endogenous provirus (boyd et al., 1993) . this work is mentioned as being indicative of a type of approach that should be undertaken on a major scale with bovine and ovine tissues. particular attention should be paid to the possibility of retroviruses akin to the mink-cell focus (mcf) inducing viruses being involved in the pathogenesis of the spongiform encephalopathies 5. the possible involvement of other retroviruses, particularly the lentiviruses, maedi-visna in sheep and visna-like virus in cattle (evermann, 1990) should not be overlooked as possible contributors to the development of scrapie and bse, in their respective hosts. attention is also drawn to the mouse hepatitis viruses, coronaviruses that have the ability to enter into recombination with other viruses (luytes et al., 1988) and are widely present in mouse stocks; there are also neurotropic strains (kyuwa and stohlman, 1990) . louping-ill virus is a tick-transmitted member of the virus family flaviviridae (fenner et al., 1987; gao et al., 1993) that is closely related to other european tick-borne encephalitis viruses (jiaug et al., 1993) . it was the evaluation of a vaccine designed to prevent disease due to louping-ill that led indirectly to the initial major studies on scrapie (see above). could vaccination against infection with this virus have increased the susceptibility to scrapie? another virus that has been mentioned in the context of scrapie causation is the hepadnavirus. some time ago, carl eklund stated that 'the scrapie agent is not a member of a unique group of pathogens but is a medium-sized virus whose resistance to heat is analogous to that of serum hepatitis virus' (hotchin, 1979) . this suggestion seems to have attracted little attention, but there is a strong possibility that hepatitis b viruses (tiollais and buendia, 1991) are more widespread than previously believed. viruses similar to human hepatitis b virus have been found in woodchucks and ground squirrels (minuk et al., 1986) , while antibody to this virus can be demonstrated in many other animals, including pigs, goats, cattle and especially sheep (hoofnagle et al., 1983) . of particular interest, with relation to the virino hypothesis of scrapie, is the fact that another agent is associated with hepatitis b virus. this agent is the hepatitis delta virus (hdv). hdv is a small defective viroid-like rna virus (negro et al., 1989 ) that requires another virus, hepatitis b virus, as a helper. thus, even if hepatitis b viruses are found in farm animals, the search should not stop there but continue with a search for hdv-like viruses. humphrey-smith and colleagues (1992) noted that the presence of spiroplasma-like inclusions has repeatably been noted in cases of cjd, that spiroplasmas show a number of similarities with the scrapie-associated fibrillar protein (saf), and that lesions minicking spongiform encephalopathy are produced by spiroplasma-infected rodents. spiroplasmas belong to the class mollicutes, order mycoplasmatales. like mycoplasmas, spiroplasmas, some of which are pathogenic for insects and plants, are a family within this order. humphrey-smith and colleagues (1992) suggested that spiroplasmas should not be ignored as a possible cause of spongiform encephalopathies. there are indeed immunological cross-reactions between the proteinaseresistant proteins of spiroplasma mirum and prpsc (bastian et al., 1987) and there are proteinase-k-resistant proteins in other members of the class mollicutes (butler et al., 1991) . it would have been thought that articles would have appeared by now listing attempts to demonstrate antibody to s. mirum and other spiroplasmas in the sera of sheep with scrapie and cattle with bse and attempts to correlate the titres against such agents with the onset of signs, but there do not appear to have been any such papers. scrapie is a particularly frightening disease. the transmissible agent is very resistant to heat (mould and dawson, 1970) and to other agents normally destructive to biological materials and it gives the appearance of an ability to cross species boundaries. the first clue to the resistance of the agent came from the louping-ill vaccination trials in the 1930s (see introduction), where scrapie developed following the inoculation of formalinized tissue. while later shown not to be completely resistant to the action of formalin, the agent's resistance to a number of other disinfectants is unexpectedly high for the viruses and similar agents with which we are familiar. further, studies on the critical size of the agent by x-irradiation and by exclusion chromatography give very different estimates of the limiting size of the agent. x-ray studies suggest a size of 1.5 x 102 kda, while the much larger figure of 5 × 104 kda is suggested by chromatographic studies. this difference suggested to hunter (1979) , a source for much data on the physical properties of the scrapie agent, that a two-component system was necessary for scrapie infectivity, a small informational component that provides the radiation target, and a larger component that provides a protective or carrier role. hunter (1979) also indicated that only proteases have much effect on scrapie activity and that, even for this effect to occur, the substantial amounts of lipid associated with brain tissue must be removed. part of the evidence that led to the prion concept was the finding that the proteinase pronase destroyed the activity of the agent. this pointed to a need for a protein component for its infectivity (cho, 1980 (cho, , 1983 . however, certain tissues contain proteinase inhibitors that can block the effectiveness of proteinases and, in studies on the kind of activity proteinases have on brain extracts, darcel (1990) found that pronase, highly active in control tests on sheep l';aemoglobin, had no proteolytic effect on the hirt supernatant of normal mouse brain. reasons for this lack of activity with brain extracts should be explored and extended to proteinase-k, to which prpsc is more resistant that its homologue in normal mouse brain 6. the sensitivity to this enzyme was found to va~ between three stains of the mouse-passaged scrapie agent (kascak et al., 1985) '. without data on what the proteinase-k was really accomplishing, this information cannot be convincing. there will not wittingly be repeats, on any worthwhile scale, of what i have termed the accidental 'experiments'. thus it is useful to look back at them and see whether any information was missed in the earlier analysis. the louping-ill vaccination trial provided much of the information on scrapie that we have today: the long latent period before disease appears, the resistance of the agent and the resistance of much of the population to the development of the disease. rather than implicating the acquisition of infection by ingestion of 'contaminated' feed, that event directly implicated parenteral 'infection'. no possibility should be overlooked in studies of a disease as strange as scrapie. even though louping-ill virus affects the same areas of the nervous system as evidenced by clinical signs, inactivated louping-ill virus seems to have been excluded as a contributing factor to the development of scrapie by the design of the vaccination trials. however, the possibility that the injection of a killed louping-ill virus can potentiate infection of the scrapie agent deserves further study s. other controls in attempting to understand the results of the vaccination trials of the 1930s could be to examine brains of sheep with louping-ill for the presence of protease resistant prpsc and also to test the sera of sheep used in scrapie-transmission experiments for antibodies to a wide spectrum of louping-ill virus isolates. while the above concerns spread of scrapie by parenteral injection, there has been a suggestion that natural transmission may occur from sheep eating dropped placentae (dickinson et al., 1974) , but this is contraindicated by the fact that ewes usually confine themselves to eating their own placenta. the interpretation that a change in the rendering of sheep offal led to the outbreak of bse is based on statistical evidence and on the conclusion that, in the circumstances at the time, no other explanation was possible. it was also an explanation that could be adopted for political reasons and gave some justification for the control procedures that were adopted for controlling bse. however, it has been pointed out that the human risk from eating mutton and lamb must be negligible since, in a country where scrapie occurs as a disease and where this meat has been eaten for a very long time, the incidence of encephalopathies, such as cjd, that might be expected to follow, is actually very low (0.49 cases per million) (mathews, 1990 ). finally, despite earlier concerns, a connection between the feeding of sheep offal and the incidence of tme has not been proven (see above). avery and colleagues (1960) considered that, if scrapie were a chronic disease caused by an infective agent, the level of gamma-globufin, the serum protein fraction containing antibody, might well be raised. unfortunately, the electrophoretic patterns were similar for the sera of both normal and scrapie-affected animals. nevertheless, despite the 'rather overwhelming evidence for "immunological silence"' in scrapie, stites and colleagues (1979) advanced some evidence for the development of an immunological response. they stressed that the lack of detection of an immune response does not convincingly disprove its existence. whatever the causal agent, if its tropism were entirely neural, then insufficient agent might reach the blood stream to lead to an immunological response. this explanation appears weak, but another possible explanation is that saf is too close to its equivalent produced by the prp gene of normal animals to give a response, although it would be expected that even a small difference would lead to the production of recognizable antibody. a more direct immunological approach would be to look for antibody to other disease entities; for example, with what frequency are antibodies to blv encountered in cattle in cases of bse? with what frequency are maedi-visna antibodies encountered in sheep with scrapie? these viruses are quite prevalent in dairy cattle and in sheep, respectively, and are slow viruses in their own right. both viruses have a dna phase and could well have an association with genes for both susceptibility to the encephalopathies and amyloidal fibrin formation. antibodies to hepadnaviruses, to mcf viruses and to spiroplasma should also be sought. unless there were an associated immunodeficiency, the absence of detectable antibody to these agents would largely remove them from consideration as having primary or secondary roles in the causation of the spongiform encephalopathies. evidence of an immunological deficiency should also be sought using current technology similar to that used by rao and colleagues (1990) to study the immune deficiency produced by the avian retrovirus, avian erythroblastosis virus. stites and colleagues (1979) affirmed that current information could accommodate the possibility that negative regulation or active immunosuppression play an important role in the pathogenesis of scrapie. they noted that mice show a distinct splenomegaly after intracerebral inoculation with the agent, that the agent at first appears in higher concentrations in the spleen than in the brain, and that splenectomy prolongs the incubation period. they also recorded altered responses of spleen cells to mitogens. as in other studies with laboratory animals, they could not entirely rule out the possibility that a passenger virus or microorganism was responsible for these changes. this possibility of the presence of passenger viruses and microorganisms is a constant obstacle to understanding the results of agent transmission experiments in laboratory animals. is spiroplasma mirum, proposed as a possible cause of scrapie, merely a passenger responding to a suppressed immune state in the same way as pneumocystis carini in patients with aids (redfield and burke, 1988) ? any passenger agents involved in the spongiform encephalopathies might well vary according to the host species. because of the widespread consumption of products obtained from cattle by most of the human population, the recent outbreak of bse has caused great alarm and has been the subject of a considerable correspondence in the british medical press. it is not known whether cattle are the final hosts, nor whether the infection can spread to humans by the consumption of beef, milk or products, such as bovine albumin used in various forms of therapy 9. with the ability for agents causing the spongiform encephalopathies to cross specimen boundaries, the fear that people might acquire human forms of these diseases from cattle has to be very real. it might be assumed, for instance, that, as a result of the epidemic of bse through which british cattle have passed, slaughterhouse workers will have inadvertently come into contact with preclinical cases (howard and castle, 1990; deauer and lacey, 1991; coyle and harvey, 1992) . scrimgeour and brown (1991) pointed out that the most efficient means for transmission of scrapie, other than the parenteral routes used experimentally, are via the conjunctival and nasal mucosae and cutaneous abrasions. they suggested that slaughterhouse workers would be prone to this type of contact with microbial zoonoses. the wearing of protective clothing by these individuals when they are extracting the cranial contents from cattle was also advised. however, these fears of the human population acquiring bse should be tempered by realism (collee, 1991) . cjd often occurs in clusters with a history of familial relationships, and these seem to have more relevance to the incidence of the disease than do eating habits (taylor, 1989) . if bse presents a hazard to the human population, especially to veterinarians or to slaughterhouse personnel, a statistical study of the occurrence of spongiform encephalopathy over the next several years should show whether the incidence of cjd has increased as a result of the intimate exposure to cattle. the potential risks to which the british meat-eating population has been exposed have led to funding of the surveillance of cases of cjd that arise in the united kingdom, with particular reference to the dietary and occupational history and other possible risks . however, fear and devakumar (1990) have questioned the accuracy of such an ongoing survey, given the variety of symptoms exhibited by patients with cjd. a major problem with experimental work on the encephalopathies is the resistance of the agent. in performing experiments on scrapie, etc., great care has to be taken to avoid using contaminated laboratory equipment. illustrations of the care that has to be taken are the technical details given in papers on the oral transmission of bse to mice (barlow and middleton, 1991) and in the transmission studies with alzheimer's disease (manuelidis and manuelidis, 1991) . the innate resistance of a proportion of the population to experimental transmission of the spongiform encephalopathies is seen as the result of genetic influences. some animals within a species will prove resistant to experimental infection following infection. however, the degree of resistance will vary according to the route of infection. in most species, the natural route of infection appears to be the oral route and innate resistance is then likely to be more important than when animals are infected intracerebrally. offspring from infected ewes are prone to develop scrapie. this may be evidence of a familial trait and the practice has been to remove affected family lines in attempts to control the disease 1°. however, it is uncertain whether this predisposition of sheep to scrapie is due to increased exposure at time of parturition or to gene-dependent susceptibility. there are specific genes for susceptibility to scrapie: sip, which controls the incubation period of the disease in sheep, and the sinc gene in mice (bradley and matthews, 1992; kimberlin, 1990) . sinc has also been found genetically linked to the prn protein gene (hunter et al., 1987) . the finding by deslys and colleagues (1994) that all the cases of cjd acquired by injections of human growth hormone in 25 french children occurred in individuals who were homozygous at codon 129 of prp is seen as support for the importance of the prp gene in the development of encephalopathies. a complication in the british outbreak of bse is that the national gene pool among cattle will have changed over the same time-frame as the epidemic owing to the widespread adoption of artificial insemination (ai) and embryo transfer (et). perhaps a new study of the data that have been accumulated on this outbreak might point to certain bulls or certain cows that have had an unexpectedly high incidence of bse in their progeny. if such an effect could be demonstrated, it would mean either that these cases have followed the introduction of genes for susceptibility into previously unaffected herds or that actual infection occurred through the use of these breeding techniques. wrathall and brown (1991) stated that there is no evidence to suggest that vertical transmission of bse occurs and discussed et and ai studies with scrapie-infected sheep. it has been thought that et and ai might provide a means for breaking the disease transmission cycle in scrapie and bse, but such studies will be complicated by the existence of the genes for susceptibility. the research done on the prion theory has greatly expanded our knowledge of these diseases, but the most active promoters of the prion ask us to believe that it, and its gene, can cause the encephalopathies, that it is responsible for the predisposition to and the pathogenesis of these diseases, and that it can ensure its own replication. much more controlled work on the prion is necessary before conclusions of this kind can be justified. for instance, can the material of which prions are a component (saf) act as a viral promoter of other neurotropic viruses? an investigation of such a possibility would seem to be essential in understanding the role that the prion plays. studies with known neurotropic viruses, with the viruses given in the presence and absence of saf, should show whether this material can have a promoter effect. tile quotation at the beginning of this article describes workers in the field of the encephalopathies very well, each interested in their own particular area. in the author's case, the bias is towards viruses being involved in the causation of these diseases. in considering the experimental work that has been done, the differences in the biological activities of the isolates of the transmissible encephalopathies are impressive. if bse were really a manifestation of transmitted scrapie, one would expect to find that its isolates behaved with the variability shown by scrapie isolates. this does not occur. not knowing the cause of the transmissible encephalopathies makes the approach to controlling these diseases very difficult. the outbreak of bse in britain has disrupted the export of pedigree cattle, and there were other ramifications. there is the concern, not substantiated, that once transmitted to a new species, such as bovine, the agent might persist in the new species. there are also fears that the disease might become established in other countries and cross the species boundary to man. the expenditure of money for research on the encephalopathies is therefore justified from the agricultural standpoint alone, and the present paper has tried to suggest areas that are suitable for further exploration. one interpretation of the cause of the outbreaks of scrapie, tme and bse noted in this paper is that material in the feed or in the inoculum triggered the growth of a quiescent but unrecognized virus. is such triggering of a quiescent virus possible? we know that contact with certain compounds can lead to the multiplication of latent oncornaviruses (robinson et al, 1976) and that corticosteroids can provoke the emergence of viral activity with latent herpesvirus infections (darcel and dorward, 1975; fenner et al, 1987) . other mechanisms may also exist for the stimulation of virus growth. it is peculiar how refractory the transmissible spongiform encephalopathies are to investigation, and a real understanding of their causation is still elusive. yet, with research, there is a chance of discovering new types of disease agents and learning much about the nature of important neurological diseases. 2. autophagic vacuoles in mice, with sequestered cytoplasmic areas containing ribosomes and occasionally mitochondria and small secondary vacuoles, have been described in neuronal perikarya and neuronal processes of mice with experimental scrapie and cjd (liberski et al, 1992) . there seems to be a renewed interest in these neuropil and neuronal vacuolations in circumscribed areas of the brain, with both scrapie and bse (shinagawa et al, 1985; sasaki et al, 1986; wells et al, 1987; wells and wilesmith, 1989; liberski, 1990; scott et al, 1990) . 3. further references to lhermite's and wilson's diseases are given in the paper by 4. to study the possibility that these characteristic vacuoles are due to oxygen would mean the application of a variety of biochemical and histochemical techniques, including a qualitative and quantitative study of the enzymes just mentioned (brunori and rotilio, 1984) . r.m. barlow (personal communication, 1994) has suggested that marginal copper deficiency, which is very common in sheep, would affect the superoxide dismutases and facilitate accumulations of oxygen free radicals and toxic peroxides, but the result from studies on copper levels in the brains of sheep with scrapie and those of normal sheep do not support this idea. mcf murine viruses are formed de novo by recombination between exogenous, infectious ecotropic virus and endogenous sequences related to xenotropic virus. members of this group of viruses have been isolated from spontaneous mouse lymphomas and can infect both murine and xenogenic cells (cells of other species). special attention is drawn to work of gardner and colleagues (1979) on a routine retrovirus motor disease. mcf-like viruses have also been recovered from feline (neil et al, 1991) and avian oncogene-murine chimeric systems (feuer et al, i993) . other references that provide a useful entry to this field include adachi et al. (1984) , chattopadhyay et al (1991) , di fronzo and holland (1993), khan et al (1987) , koch et al (1984) , levy et al. (1985) , makino et al. (1990) , oliff et al. (1984) , quint et al. (1984) and shields et al. (1991) . 6. if scrapie-infected brain had been used instead of normal brain and loss of activity of the scrapie agent had been observed, then the effect would probably have been due to non-proteolytic factors in the pronase. the suggestion was offered (darcel, 1990 ) that the brain extracts contained a proteinase inhibitor. this should be looked for, and, if found, removed from the system before applying the proteinase treatment. these authors' protocols show that brain homogenates were homogenized with a solution containing proteinase-k, and the supernatant resulting from centrifugation was incubated for i hour. the action of the enzyme was then blocked and the solution was then treated with micrococcal nuclease. many other steps followed before titres were determined in mice and hamsters and compared with those in the original homogenates. the name 'berry-dedrick phenomenon' was coined to describe the results that followed the appearance of myxomatosis after inoculation of rabbits with a heat-inactivated preparation of myxoma virus inoculated into rabbits together with live fibroma virus. this was later explained as resulting from the presence of live, residual myxomatosis virus (fenner et al., 1974) . however, cross-reactivation is a known phenomenon by which infectious virions containing active genes from both viruses can result from genetic recombination between an infectious virus and a related inactivated virus (fenner et al., 1987) my late colleague robert kasting became ill some years after we concluded our studies of amino acids in scrapie and normal brain. he felt that the signs of neurological disease that he exhibited meant he had acquired scrapie from aerosols produced by grinding the brains. his brain was sent to the rocky mountain laboratory after his death for detailed histology and for transmission experiments. the results did not confirm his suspicions. i0. six of the 26 lambs born to recipients of embryos transplanted from donor ewes that had been experimentally infected with scrapie developed this disease (foster et al., 1992) . no correlation was found between susceptibility and the phenotypes for haemoglobin, blood potassium levels, lactoglobulins, albumin, pre-albumin, esterase, haemoglobin, transferrin, reduced glutathione or a-mannosidase hunter, 1991) . characterization of the env gene and long terminal repeat of molecularly cloned friend mink cell focus-inducing virus a note on serum proteins in normal and scrapie infected sheep fnrther observations on the neuropathology of experimental scrapie in mouse and hamster oral transmission studies of bse to mice antiserum to scrapie-associated fibril-protein cross-reacts with spiroplasma mirum fibril proteins scrapie agent proteins do not accumulate in grey tumour mice purification of scrapie agents: how far have we come? identification of a protein that purifies with the scrapie prion copurification of sp33-37 and scrapie agent from hamster brain prior to detectable histopathology and clinical disease the human endogenous retrovirus erv-3 is unregulated in differentiating placental trophoblast cells sub-acute, transmissible spongiform encephalopathies: current concepts and future needs scrapie, concept of a virus-induced amyloidosis of the brain transmissible spongifolan encephalopathies in human. kuru, creutzfeldt-jakob disease and gerstmann-str~iussler-scheinker disease the human spongiform encephalopathies: kuru, creutzfeldt-jakob disease and the gerstmann-str~iussler-scheinker syndrome biological stability of different classes of scrapie agent biochemistry of oxygen radical species mice devoid of prp are resistant to scrapie identification and characterization of proteinase-k resistant proteins in members of the class mollicutes detection of prion protein mrna in normal and scrapie-infected tissues and cell lines n-terminal truncation of the scrapie-associated form of prp by lysomal protease(s): implications regarding the site of conversion of prp to the protease-resistant state encephalopathy in mice produced by inoculation of scrapie brain characteristics and contributions of defective, ecotropic, and mink-cell focus-inducing viruses involved in a retrovirus-induced immunodeficiency syndrome of mice spongiform encephalopathies: the transmissible agents spongiform encephalopathies: prp and the scrapie agent requirement of a protein component for scrapie infectivity inactivation of the scrapie agent by pronase antibody to a scrapie-associated fibril protein identifies a cellular antigen bovine spongiforna encephalopathy and man bovine spongiform encephalopathy and risk to health transmissible spongiform cncephalopathy in greater kudu (tragelaphus strepsiceos) the major protein of saf is absent from spleen and thus not an essential part of the scrapie agent biochemical aids to the genetic typing of sheep -a further note pronase does not reduce the protein content of hirt supernatant of normal mouse brain potassium and hemoglobin types in sheep with special reference to the transmission of scrapie recovery of infectious bovine rhinotracheitis virus following corticosteroid treatment of vaccinated animals brain glycogen-and copper-levels in normal sheep and sheep affected with serapie scrapie: i. transmission and pathogenesis primary parental transmission of bovine spongiform encephalopathy to the pig bovine spongiform encephalopathy and man pharmacological studies of a new derivative of amphotericin b, ms-8209 in mouse and hamster scrapie similar genetic susceptibility in iatrogenie and sporadic creutzfeldt-jakob disease a direct demonstration of recombination between injected virus and endogenous viral sequences, resulting in the generation of mink cell-focus-inducing viruses in akr mice the scrapie replication site hypothesis genetic aspects of unconventional virus infections: the basis of the virino hypothesis maternal and lateral transmission of scrapie in sheep comparative features of retroviral infections of livestock bovine spongiform encephalopathy the biology of animal viruses veterinary virology specificity of disease induced by m-mulv chimeric retroviruses containing v-myc or v-src is not determined by the ltr studies on the natural transmission of scrapie in sheep by embryo transfer transmission of bovine spongiform encephalopathy and scrapie to mice unconventional viruses and the origin and disappearance of kurn sequencing and antigenic studies of a norwegian virus isolated from encephalomyelitic sheep confirna the existence of louping-ill virus outside great britain and ireland murine retrovirus motor neuron disease strain variation in the viruses of creutzfeldt-jakob disease and kuru experimental transmission of scrapie to cattle advances in veterinary research: louping-ill, tick-borne fever and scrapie studies on a species-specific epitope in routine, ovine and bovine prion protein antibody to hepatitis b surface antigen in nonprimate animal species scrapie as a slow and latent virus concern about bovine spongiform encephalopathy linkage of a prion protein missense variant to gerstmann-str~iussler syndrome spiroplasma and spongiform encephalopathies the enigma of the scrapie agent: biochemical approaches and the involvement of membranes and nucleic acids natural transmission and genetic control of susceptibility of sheep to scrapie linkage of the scrapie-associated fibril protein (prp) gene and sinc using congenic mice and restriction fragment length polymorphism analysis idiopathic brainstem neuronal ehromatolysis and hippocampal sclerosis: a novel encephalopathy in clinically suspect cases of bovine spongiform encephalopathy ultrastructural features of spongiform encephalopathy transmitted to mice from three species of bovidae single amino acid codon changes detected in louping411 virus antibody-resistant mutants with reduced virulence biochemical differences among scrapie-associated fibrils support the biological diversity of scrapie agents comparisons by paper chromatography of extracts of brain, urine and plasma from normal and scrapie sheep expression of mink cell~focus forming murine leukemia virus-related transcripts in akr mice slow virus diseases of animals and man biochemical and behavioural changes in scrapie transmissible encephalopathies in animals characteristics of a short incubation model of scrapie in the golden hamster organ distribution of proteinase-resistant prion protein in humans and mice with creutzfeldt-jakob disease molecular analysis of the envelope gene and long terminal repeat of friend mink cell focus-inducing virus: implications for the functions of these sequences pathogenesis of a neurotropic murine coronavirus, strain jhm in the central nervous system of mice normal expression of polynlorphic endogenous retroviral rna containing segments identical to mink cell focus-forming virus infectious peptides: postulated mechanisms of protovirus replication in scrapie ultrastructural neuropathologic features of bovine spongiform encephalopathy neuronal autophagic vacuoles in experimental scrapie and creutzfeldt-jakob disease bovine spongiform encephalopathy sequence of mouse hepatitis virus a59 mrna: 2. indications for rna recombination between corona-viruses and influenza c virus h-2-associated and background genes influence the development of a routine retrovirus-induced immunodeficiency syndrome observations on creutzfeldt-jakob disease propagated in small rodents search for a transmissible agent in alzheimer's disease: studies of human buffy coat the subacute spongiform encephalopathies physical and chemical properties of the transmissible mink encephalopathy agent equilibrium density gradient centrifugation of the scrapie agent in nycodenz epidemiological and experimental studies on a new incident of transmissible mink encephalopathy bovine spongiform encephalopathy abnormal fibrils from scrapie-infected brain glycosidases in normal and scrapie mouse brain ground squirrel hepatitis virus (gshv) and hepatocellular carcinoma in the canadian richardson ground squirrel (spermophilus richardsonii) the response in mice to heat treated scrapie agent assay of ethane and pentane from isolated organs and cells basis of hepatitis delta virus disease? feline leukaemia virus: generation of pathogenic and oncogenic variants the envelope gene and long terminal repeat sequences contribute to the pathogenic phenotype of helper-independent friend viruses homozygous prion protein genotype predisposes to sporadic creutzfeldt-jakob disease spongy brain: the mystery of mad cow disease. toronto globe and mail novel proteinaceous infectious particles cause scrapie pathogenesis, immunology, virology and molecular biology of the spongiform ence_phalopathies generation of akr mink cell focus-forming xenotropic-like provirns provides recombinant long terminal repeat sequences analyses of frequency of infection, specific infectivity, and prion protein biosynthesis in scrapie-infected neuroblastoma cell clones immune abnormalities in avian erythroblastosis virns-infected chickens hiv infection: the clinical picture production of unique c-type viruses by chicken cells grown in bromodeoxyuridine experimental infection of mink with bovine spongiform encephalopathy the scrapie agent: 'a virus by any other name detection of scrapie-associated fibril proteins using anti-saf antibody in non-purified tissue preparations spongiform encephalopathy in sheep scrapie: electron microscopic observations bovine spongiform encephalopathy: detection and quantitation of fibrils, fibril protein (prp) and vacuolation in brain transgenic mice expressing hamster priori protein produce species-specific scrapie infectivity and amyloid plaques bovine spongiform encephalopathy in an adult british friesian cow bse and potential risk to slaughtermen calorie restriction suppresses subgenomic mink cytopathic focus-forming routine leukemia virus transcription and frequency of genomic expression while impairing lymphoma formation characterization of the scrapie agent isolated from sheep in japan differential glycosylation of the protein (prp) forming scrapie associated fibrils the immunology of scrapie a transmissible variant of creutzfeldt-jakob disease with kuru plaques bovine spongiform encephalopathy and human health hepatitis b virus the prion's progress a 'unified theory' of priori propagation recently described scrapie-like encephalopathies of animals: case definitions the distribution of neuronal vacuolation in bovine spongiform encephalopathy (bse) is constant a novel progressive spongiform encephalopathy in cattle bovine spongiform encephalopathy: diagnostic significance of vacuolar changes in sgleeted nuclei of the medulla oblongata bovine spongiform encephalopathy: observations on the incidence during 1992 bovine spongiform encephalopathy: epidemiological studies on the origin bovine spongiform encephalopathy: case-control studies of calf feeding practices and meat and bone meal inclusion in proprietary concentrates bovine spongiform encephalopathy and risk to health chronic wasting disease of captive mule deer: a spongiform encephalopathy spongiform encephalopathy of rocky mountain elk spongiform encephalopathy in a captive puma (felis concolor) embryo transfer, semen, scrapie amphotericin b treatment dissociates in vitro replication of the scrapie agent from prp accumulation i am deeply appreciative of the help given me by dr h.j. cho and dr r.m. barlow in the preparation of this manuscript. lisa kalischuk-timensen helped collect references by computer searches. this essay is dedicated to the memory of my friend robert kasting, for his assistance many years ago. notes 1. louping-ill is a tick-borne virus-caused disease of lambs, characterized by incoordination of the hind limbs and by hypersensitivity to noise and touch (blood et al., 1983) , seen principally in great britain and russia. key: cord-022034-o27mh4wz authors: olano, juan p.; peters, c.j.; walker, david h. title: distinguishing tropical infectious diseases from bioterrorism date: 2009-05-15 journal: tropical infectious diseases doi: 10.1016/b978-0-443-06668-9.50124-1 sha: doc_id: 22034 cord_uid: o27mh4wz nan bioterrorism can be defined as the intentional use of infectious agents or microbial toxins with the purpose of causing illness and death leading to fear in human populations. the dissemination of infectious agents with the purpose of attacking livestock and agricultural resources has similar motives. many of the agents that could potentially be used in bioterror (bt) attacks are also responsible for naturally occurring infectious diseases in the tropics. as such, naturally occurring outbreaks must be differentiated from bt attacks for public health, forensic, and security reasons. if a bt attack occurs in tropical underdeveloped countries, owing to their weak public health infrastructure, the public health implications would be even more dramatic than in developed countries. an outbreak of smallpox due to a bt attack would probably require vaccination and mandatory quarantine of millions of people in order to control the outbreak and quell global public unrest. this chapter will concentrate on selected infectious agents that have the potential to be used as bioterror agents in human populations. the first step in managing the damage from a covert biological dissemination is recognition of the attack and the organism(s). as in most emerging infections, we predict that in bioterrorist attacks the etiological diagnosis will be made by a clinician or pathologist and the recognition of a bioterrorist event will be through geographical and epidemiological anomalies. we have very limited environmental detection capability at this time, and there are no comprehensive pointof-care diagnostics for most of the high-impact bt agents. some diseases such as inhalational anthrax or smallpox may be relatively readily recognized by an alert clinician because of their very distinctive presentation in many cases. however, the leading edge of a bt epidemic may arrive on a pathologist' s doorstep without prior suspicion. for example, individual cases of pneumonic plague as the earliest harbingers of an attack will presumably present as community-acquired pneumonia and probably die without clinical diagnosis. given the short window available for successful treatment, the recognition of these earliest cases is paramount. sartwell 1 has demonstrated empirically that incubation periods follow a log-normal distribution, which results in "front-loading" of cases ( fig. 119-1 ). delay in recognizing the epidemic through reliance on syndromic surveillance or other surrogates will likely result in most of the cases of diseases such as plague and tularemia being well into their disease course and perhaps unsalvageable. 2 bioterrorist events will enlarge our knowledge of tropical diseases. for example, inhalational anthrax and several viral hemorrhagic fevers (vhf) thought to be transmitted mainly by aerosol 3 are under-represented in naturally occurring case series, and a bt attack would provide an opportunity to answer questions about the underlying host factors and pathogenesis. indeed, the extension of the risk population to include children, the elderly, and the immunosuppressed is likely to provide considerable insight into these oftenunderstudied groups. it is also likely that our lack of information about them will challenge our current diagnostic algorithms. in october 2001, anthrax spores were distributed covertly in the u.s. postal service, leading to 22 cases of human anthrax and billions of dollars spent on controlling the potentially devastating effects of a small inhalational anthrax epidemic. 4, 5 this attack was by no means the first intentional attempt to use infectious agents as weapons of terror. ever since the times of the ancient greeks and romans, humans have tried to inflict damage by the use of contagion on other populations. 6, 7 less than 4% of the people or groups responsible for terrorist attacks on human populations take responsibility for their actions. 8 therefore, the use of biological weapons is ideal to conduct covert attacks. in addition, it has been estimated that to kill the same number of human beings with biological weapons as compared to chemical or nuclear weapons, the cost is far less with biological weapons ($2/human casualty) compared with chemical ($2000/ human casualty) and nuclear ($2,000,000/human casualty) weapons. 6 hypothetical bt attacks would range from an overt attack of a large city with a bomb containing several kilograms of an agent (weaponized bacteria, viruses, or toxins) to discrete or covert intentional release of the infectious agent through a delivery system, such as spray devices, postal service, ventilation ducts, water supplies, and food supply. based on transmissibility, severity of morbidity and mortality, and likelihood of use (availability, stability, weaponization), potential bt agents are divided into three categories (table 119-1) . this chapter will concentrate on selected agents from categories a and b and on the diagnostic challenges posed by illnesses caused by such agents. table 119 -1 are capable of producing illness under natural circumstances. therefore, the first challenge is to identify the infectious agent responsible for a certain disease correctly, followed by a thorough epidemiologic and microbiologic analysis of the epidemic or outbreak. in some circumstances, the identification of a bt attack would be obvious. a case of smallpox in any human population is an international emergency that would trigger a massive response of the public health systems around the world. sophisticated epidemiological investigations would follow in order to characterize the outbreak, identify the source, and possibly label it "intentional." in other cases, the identification of the outbreak as secondary to intentional dissemination of an infectious agent will require the use of sophisticated epidemiological and molecular tools, especially for diseases endemic to the area where the outbreak occurs. the need to use genetic sequences as markers has spawned a new discipline referred to as microbial forensics, sister to phylogenetics and "molecular epidemiology." differentiation between natural infections and a biological warfare attack rests firstly on disease patterns given by several epidemiological clues. they include presence of disease outbreaks of the same illness in noncontiguous areas, disease outbreaks with zoonotic impact, different attack rates in different environments (indoor versus outdoor), presence of large epidemics in small populations, increased number of unexplained deaths, unusually high severity of a disease for a particular pathogen, unusual clinical manifestations owing to route of transmission for a given pathogen, presence of a disease (vector-borne or not) in an area not endemic for that particular disease, multiple epidemics with different diseases in the same population, a case of a disease by an uncommon agent (smallpox, viral hemorrhagic fevers, inhalational anthrax), unusual strains of microorganisms when compared to conventional strains circulating in the same affected areas, and genetically homogenous organisms isolated from different locations. 9, 10 these are a few guidelines that could prove helpful when investigating an outbreak, but it has to be kept in mind that the deduction will not be based on any single finding but rather the pattern seen in its totality. first and foremost, the possibility of an attack must be ever in mind, or differentiation of a covert bt attack and a natural outbreak of an infectious disease may not be made. in fact, the outbreak of salmonellosis in oregon in 1984 was due to a covert attack planned by the rajneeshee leadership and accompanied by distinctive epidemiological clues. it was not labeled as intentional until somebody came forward with the information leading to the responsible group; as in most of medicine, the unsuspected diagnosis is the easiest to miss. 11 an increasing number of public health departments are now acquiring the technology necessary to perform syndromic surveillance. this new method of surveillance is based on syndromic disease rates such as respiratory, gastrointestinal, and neurological syndromes or analysis of other health-related activities such as laboratory test requests and results, purchasing rates for certain pharmaceutical agents, unexplained death rates, and veterinary surveillance. 2, 10, 11 the purpose of syndromic surveillance is to detect a bt attack as early as possible by analyzing the previously mentioned variables by extracting and analyzing data through computer networks. the rationale behind syndromic surveillance is the nonspecific nature of early signs and symptoms of many of the illnesses caused by bt agents. examples of proposed syndromes are as follows: gastroenteritis of any apparent infectious etiology, pneumonia with the sudden death of a previously healthy adult, widened mediastinum in a febrile patient, acute neurologic illness with fever, and advancing cranial nerve impairment with weakness. 12 a key component factors affecting syndromic surveillance include selection of data sources, definition of syndrome categories, selection of statistical detection thresholds, availability of resources for follow-up, recent experiences with false alarms, and criteria for initiating investigations. it must be emphasized that these systems are experimental and not yet of proven value in managing bt attacks. they are expensive, require follow-up confirmation, have unproven sensitivity and specificity, and ultimately depend on the clinician. 2 they may prove to be more useful in managing an event than in expeditiously detecting one. conventional epidemiological investigations are by no means obsolete with the availability of more sophisticated methods to study possible bt attacks. they include the confirmation of an outbreak once it is suspected. confirmation is based in many cases on laboratory analysis of patients' samples or autopsy material. a case definition is constructed to increase objectivity of the data analyzed and to enable determination of the attack rate. other variables are included in the analysis, such as time and place, and an epidemiological curve can be constructed. 10 epidemiological curves are an important tool to analyze epidemics and suggest the mode of transmission and propagation. a point source epidemic curve is classically log-normal in distribution 1 and would suggest a common exposure of a population to an infectious agent. of course, there can be variations depending on the presence of susceptible subpopulations (e.g., children, immunosuppressed, aged) and on varying doses of the agent. propagative curves are more characteristic of highly communicable agents such as smallpox. a short description of selected category a and b agents follows. all these pathogens are addressed as naturally occurring disease agents in other chapters of this book. bacillus anthracis (anthrax) b. anthracis (see chapter 39) is without a doubt the microorganism that has received the most attention as a bt agent due to its high lethality (inhalational form), ease of propagation, and high environmental stability. fortunately, the disease is not transmitted from person to person. however, the first three characteristics make it one of the ideal bioweapons. anthrax presents in humans as four different clinical syndromes, depending on the portal of entry: cutaneous (the most common form of the disease resulting from contact with infectious animal products), gastrointestinal and oral/oropharyngeal (both secondary to ingestion of contaminated meat), and inhalational (woolsorter' s disease), secondary to inhalation of spores from the environment. in the event of a bioterror attack, either overt or covert, the clinical presentation of the patients affected by the attack would be that of inhalational anthrax. this form of anthrax is so rare that a single case of inhalational anthrax should raise immediate suspicion, as dramatically demonstrated during the bt attacks in the fall of 2001. [13] [14] [15] during those attacks, 50% of cases were cutaneous anthrax thought to be secondary to handling of anthraxlaced mail envelopes or environmental surface contamination in the presence of minor cutaneous lesions, providing a portal of entry for the spores. 5 an outbreak of inhalational anthrax also took place in sverdlovsk (former soviet union) as a result of an accidental release into the air of b. anthracis spores from a facility producing anthrax for the bioweapons program in the ussr. 5, [16] [17] [18] inhalational anthrax should be suspected clinically in any individual presenting with fever and a widened mediastinum on chest radiograph (due to hemorrhagic mediastinitis). 19, 20 the incubation period is normally 3 to 5 days, but in some cases it can be as short as 2 days and as long as 60 days depending on inoculum and the time of germination of the spore. 17 based on research performed on rhesus monkeys, the ld 50 is estimated to be 8000 to 10,000 spores. [21] [22] [23] however, as few as 1 to 3 spores may be capable of producing a fatal outcome in approximately 1% of those exposed to these quantities. 24 the initial symptoms are nonspecific and consist of fever, malaise, anorexia, fatigue, and dry cough. these symptoms are followed in 3 to 4 days by an abrupt onset of respiratory insufficiency, stridor, diaphoresis, and cyanosis. the subsequent clinical course is rapid, and patients usually die within 24 to 36 hours after clinical deterioration. mortality is 100% without antibiotic therapy. 20, [25] [26] [27] early diagnosis, aggressive treatment with antimicrobial agents to which the bacteria are susceptible, and aggressive supportive therapy decreased the mortality to 40% in the 2001 attacks. 5 pathologic studies performed on the sverdlovsk victims confirmed some of the findings in animal models of inhalational anthrax, such as hemorrhagic lymphadenitis and mediastinitis. however, many patients also developed hematogenous hemorrhagic pneumonia. pleural effusions were usually large and frequently led to severe lung atelectasis. in about half of cases, hemorrhagic meningitis developed, leading rapidly to central nervous system (cns) manifestations terminating in coma and death. 16, 28, 29 yersinia pestis (plague) y. pestis (see chapter 42) is a gram-negative, aerobic, nonsporulating coccobacillus, member of the enterobacteriaceae with a wide host range, including rodents, felines, and humans. 30 the most important reservoirs are urban rats, and its main vector is the rat flea. in rural epizootics, reservoirs include prairie dogs and squirrels in the united states. 31 y. pestis has been responsible for some of the most devastating pandemics in human history in the preantibiotic era (6th, 14th, and 19th centuries). 32 public health measures have made this disease a rarity in the united states (around 20 cases/year) and around the world, although approximately 1000 cases are reported to the world health organization (who) every year (countries reporting plague include madagascar, tanzania, and peru, among others). clinical presentation in naturally acquired infections takes five forms, namely bubonic, septicemic, pneumonic, cutaneous, and meningeal. the pneumonic form is the most likely presentation in a case of plague due to a bt attack. it is worth mentioning that plague has already been used as a bt agent when japan dropped thousands of y. pestis-infected fleas over china leading to small outbreaks of bubonic plague in continental china during world war ii. 33, 34 the incubation period for pneumonic plague is short, ranging from 2 to 3 days. it is the rarest form in natural infections (1% or less) but has the highest mortality, reaching 100% in untreated patients. the initial presentation is nonspecific and consists of cough, fever, and dyspnea. cough may be productive (bloody, purulent, or watery in the initial phases). this is followed by a rapid clinical course leading to respiratory failure and the patient' s demise if not treated with antibiotics early in the course of the disease. 30, 31, 35 the factors that led to the severe manchurian pneumonic plague outbreaks in the early 20th century are unknown, but weather, hygiene, and crowding were important factors. more recent outbreaks worldwide and particularly in the united states have been much smaller and readily controlled. pneumonic cases are common in the united states, but secondary transmission has been rare in the last 50 years. modeling of pneumonic transmission using eight small outbreaks to derive the parameters find average of secondary cases per primary case (ro) to be approximately 1.3 prior to any control measures. 36 this is one of the most scientifically neglected microorganisms with bt potential. tularemia is a zoonotic infection caused by a strictly aerobic, gram-negative, nonsporulating small coccobacillus. two subspecies are recognized, namely f. tularensis subspecies holarctica (jellison type b) and f. tularensis subspecies tularensis (jellison type a). 37 type a is by far the more virulent and is present only in north america. of the bacteria with potential as bt agents, f. tularensis has by far the widest host range, including wild and domestic animals, humans, fish, reptiles, and birds. vectors are also numerous and include ticks, fleas, mosquitoes, and biting flies. 37, 38 this is an impressive range for any human pathogen. in contrast to other diseases described in this chapter, tularemia does not have the remarkable history that some of the other pathogens have. in europe, tularemia was first described in 1532; in the united states, it was first described in 1911 in california in the aftermath of the san francisco earthquake. 38 in natural infections, the most common source of infection is a tick bite and manipulation of infected animals such as wild rabbits. six different clinical syndromes have been described as follows: ulceroglandular, glandular, oculoglandular, pharyngeal, pneumonic, and typhoidal. marked overlap exists among all these forms, and for practical purposes two syndromes (ulceroglandular and typhoidal) have been proposed. [39] [40] [41] as a bt agent, f. tularensis will most likely cause a disease with a primary pulmonary component with secondary dissemination (typhoidal/systemic). in natural infections, both ulceroglandular and typhoidal forms can have a hematogenous pulmonary component, although it is more common in typhoidal forms. pulmonary features include cough, pleural effusions, and multifocal bronchopneumonic infiltrates. if not treated promptly, patients usually develop adult respiratory distress syndrome leading to respiratory insufficiency and the patient' s demise. case-fatality rate approaches 30% if not treated with appropriate antibiotics. 41 smallpox eradication remains the single most important victory in the war against infectious diseases. smallpox (see chapter 58) is the only disease so far eradicated from the face of the earth due to human intervention. the who declared smallpox eradicated in 1980 after the last case of natural disease was diagnosed in somalia in 1977, 42 and vaccination ceased around the world, rendering humankind vulnerable to reintroduction of the virus. [43] [44] [45] a laboratory accident was responsible for two more cases in 1978 in england. this accident prompted the who to restrict the frozen virus to two places in the world: the cdc in atlanta, georgia, and the institute for polyomyelitis and viral encephalitides in moscow, later moved to npo vector, novosibirisk, russia. however, it is suspected that secret military repositories exist after the fragmentation of the soviet union and the subsequent exodus of scientists involved in its bioweapons program (biopreparat). 46, 47 the agent responsible for this disease is an orthopox virus with no known animal reservoir, but high aerosol infectivity, stability, and mortality. although not a category a agent, monkeypox is responsible for outbreaks in africa and is the only other member of the orthopox genus capable of producing systemic disease in humans. the clinical disease is potentially indistinguishable from smallpox, where mortality rates in tropical africa are around 10% to 15%. in may and june 2003, an outbreak of monkeypox occurred in the united states. 48 thirty-seven infections were laboratory-documented and involved humans exposed to infected prairie dogs that had become infected because of contact with infected gambian rats and dormice, two animal species shipped from africa earlier that year. infected humans included veterinarians, exotic pet dealers, and pet owners. the clinical spectrum in this outbreak ranged from asymptomatic seroconversions to febrile illness with papulovesicular rash. no deaths were associated with this outbreak. however, phylogenetic analysis of the virus placed it in the west africa clade as opposed to the central africa clade which carries the previously mentioned case-fatality rate of 10% to 15%. a single case of smallpox would trigger a massive public health response in order to contain the outbreak. an outbreak in germany in 1970 resulted in 19 cases with 100,000 people vaccinated to contain the infection. in 1972, yugoslavia underwent an epidemic with a total of 175 cases (35 deaths) and a vaccination program that included 20 million people in order to contain the outbreak and obtain international confidence. vaccination with the vaccinia virus (a related orthopox virus) is the most effective way to prevent the disease and can be administered up to 4 days after contact with ill patients. strict quarantine with respiratory isolation for 17 days is also mandatory. the newer generation of antivirals that have been developed after the disease was eradicated has never been tested in human populations, but in vitro data and experiments in animal models of poxvirus disease suggest some antiviral activity for the acyclic nucleoside phosphonates such as cidofovir. 49 the only vaccine available in the united states is dryvax, and sufficient doses have been manufactured to cover the entire u.s. population. however, newer vaccines that may have fewer side effects are being developed. the clinical presentation is characteristic. the incubation period ranges from 10 to 12 days. the initial phase is nonspecific, common to other viral syndromes, and is characterized by abrupt onset of fever, fatigue, malaise, and headaches. during this prodromal phase in 10% of patients with fair complexion, a discrete erythematous rash appears on the face, forearms, and hands. the typical smallpox rash has a centrifugal distribution (that is, more abundant on the face and extremities than on the trunk and abdomen). an enanthem also develops with presence of oral ulcerations by the time the exanthem appears. systemic manisfestations begin to subside once the rash appears and can reappear with superinfection of skin lesions or superimposed bacterial bronchopneumonia. progression of the lesions is synchronous (maculopapules, vesicles, pustules). after pustules rupture, scabs form and detach in 2 to 3 weeks, leaving depigmented, scarred areas. this form of the disease, called variola major, is fatal in up to 30% of unvaccinated patients and 3% of vaccinated individuals. various hemorrhagic forms exist. in some cases, the rash progresses very slowly and hemorrhage develops into the base of the lesions, which remain flat and soft instead of tense, carrying a bad prognosis. in some other cases, the disease is hemorrhagic from the beginning, leading to death 5 to 7 days after the initial symptoms appear (case-fatality rate: 100%). finally, in some cases, a severe and overwhelming illness is followed by dusky skin lesions; these patients have a large quantity of virus and are extremely dangerous epidemiologically. previously vaccinated individuals usually develop a milder disease that consists of a mild pre-eruptive phase followed by few skin lesions that appear more superficial, evolve more rapidly, and are not as synchronous as the classical type. 50 viral hemorrhagic fever (vhf; see chapter 65) is caused by a heterogenous group of rna viruses that belong to several different families. the cdc identified filoviruses (ebola and marburg viruses), arenaviruses (lassa, junin, machupo, guanarito, and sabia), and bunyaviruses (crimean-congo hemorrhagic fever [cchf] and rift valley fever [rvf]). [51] [52] [53] the common denominator in these infections is the increased vascular permeability in the microcirculation leading to hemorrhagic diathesis and systemic manifestations such as pulmonary edema and cerebral edema related to leaky capillaries. 54 these viruses usually have a very narrow geographic range determined by their natural reservoirs and vectors. humans are accidental hosts. these diseases have caught great public attention due to their high mortality. this, combined with their aerosol infectivity, has led to the use of biosafety level 4 laboratories in their study. clinical presentation is usually nonspecific and consists of fever and malaise, followed by signs of increased vascular permeability and circulatory compromise. vhf usually terminates in shock, generalized mucocutaneous hemorrhages, and multiorgan failure. differences exist among the clinical details and pathogenesis of the different viruses (see chapter 65 for an overview and the individual chapters for details). for example, vhf due to filoviruses usually have prominent hemorrhagic manifestations and disseminated intravascular coagulation (dic) as a terminal event. rvf virus leads to liver damage, dic, and hemorrhagic manifestations in approximately 1% of patients with severe disease. cchf also behaves like the filoviral infections with prominent hemorrhagic manifestations. lassa fever has few neurologic or hemorrhagic manifestations. the south american arenaviral hemorrhagic fevers usually have hemorrhagic and neurologic components. toxins in the context of bt agents are substances of biologic origin that are capable of producing human illness. toxins are usually proteins synthesized by living bacteria, fungi, or plants. toxins are generally less dangerous than infectious agents. the most potent biological toxin is that from clostridium botulinum and it is 10-fold or more less lethal than anthrax on a weight basis. other toxins such as ricin are more than a 1000-fold less toxic than botulinum toxin and sarin is 30-fold less toxic than ricin. there are seven similar toxins produced by seven different serotypes of c. botulinum (a to g), all leading to the same clinical manifestations and with the same lethality. the toxins have a molecular weight of approximately 150 kda and block neurotransmission at the presynaptic level in cholinergic neurons including the neuromuscular junction, leading to progressive palsies of cranial nerves and skeletal muscle. botulinal toxins are among the most lethal substances known to mankind with ld 50 of 0.001 î¼g/g of body weight when administered parenterally. 25, 55, 56 the aerosol route decreases its lethality 80 to 100 times. both aerosol attacks and contamination of food supplies are potential bt scenarios. clinical manifestations consist of progressive bulbar and skeletal paralysis in the absence of fever, including diplopia, dysphagia, blurred vision, ptosis, dysarthria, dysphonia, mydriasis, dry mucosae, and descending paralysis. 25, 56 the cause of death in lethal cases is respiratory insufficiency due to paralysis of respiratory muscles. onset of symptoms is variable and depends on the inoculum, ranging from 24 hours to several days after exposure. most cases of naturally occurring intoxication are related to consumption of improperly sterilized canned food or ingestion of preserved fish. rare cases of inhalational botulism were documented in germany in the early 1960s due to accidental laboratory exposure. the rapid absorption through the respiratory tract may offer a different pathogenesis and it is not known if antitoxin is useful in therapy, although animal models show efficacy in prophylaxis. all the agents in category a are generally recognized as serious threats for causing extensive casualties. categories b and c are much more heterogeneous. they were considered to provide significant threat potential but there are continuing reassessments. these conditions are caused by the genus alphavirus, family togaviridae (eastern, western, and venezuelan equine encephalitis [vee] viruses; see chapter 74). natural infections are usually transmitted by mosquitoes, but aerosol transmission is the notorious cause of numerous laboratory infections and is the basis of its historic weaponization. 52, 57 most of these viruses cause systemic illness characterized by fever, myalgias, and prostration. clinically apparent involvement of the central nervous system is present in some cases and varies among the different viruses. eastern equine encephalitis (eee) is by far the most virulent, leading to case-fatality rates of 50% to 75%, and survivors usually have severe neurologic sequelae. 58, 59 vee, in contrast, leads to cns manifestations in no more than 4% of cases and almost all vee infections are symptomatic even in the absence of cns involvement. [60] [61] [62] rickettsia prowazekii (epidemic typhus) and r. rickettsii (rocky mountain spotted fever) typhus (see chapter 51) is another disease that has played a historic role in human populations. [63] [64] [65] [66] millions of people perished in world war i and world war ii due to epidemic, louse-borne typhus. large outbreaks of the disease still occur in tropical regions around the world in areas stricken by war, famine, and poverty. rocky mountain spotted fever (rmsf), on the other hand, is transmitted by tick bites and occurs endemically in south and central america as well as north america. rickettsiae target the microvascular endothelium leading to leaky capillaries systemically. 67 the main causes of morbidity and mortality are noncardiogenic pulmonary edema and cerebral edema leading to diffuse alveolar damage and meningoencephalitis. clinical manifestations are nonspecific and include fever, malaise, headache, myalgias/arthralgias, cough, nausea, vomiting, confusion, stupor, and coma in severe cases. skin rash ranges from maculopapular to petechial, depending on the severity, and is observed in around 90% of patients with rmsf and 2% to 100% of cases of epidemic typhus, depending on the darkness of cutaneous pigmentation. rickettsiae are remarkably underestimated biothreats as they are highly infectious by low-dose aerosol exposure, possess a stable extracellular form, and are resistant to most empirically administered antibiotics, including î²-lactams, aminoglycosides, and macrolides, and are exacerbated by sulfonamides. case-fatality rates can be as high as 40% to 50% without antibiotic therapy and 3% to 5% with adequate antibiotic coverage. lethal cases are usually due to delayed diagnosis. 64, 65, 68 these rickettsiae are highly infectious by aerosol and are potent bt agents. they are often discounted because of their susceptibility to tetracycline and chloramphenicol. however, the severity of the illness, the exhaustion of antibiotics in the face of a mass attack, and the existence of antibiotic-resistant organisms suggest they are still formidable players. this gram-negative, obligately intracellular bacterium has a high degree of infectivity (one organism is capable of causing infection by inhalation) and low lethality. [69] [70] [71] [72] the distribution of q fever is worldwide and results from exposure to animals such as sheep, cattle, goats, cats, rabbits, and others. c. burnetii has spore-like characteristics that can withstand harsh environmental conditions and be transported by wind to other places. in natural infections, 60% of cases are asymptomatic and are diagnosed by seroconversion. in symptomatic cases, the presentation is nonspecific and includes malaise, fever, myalgias, cough, chills, headaches, anorexia, weight loss, and in some cases pleuritic chest pain. hepatomegaly and splenomegaly are sometimes observed, although not frequently. transmission occurs by exposure to infected animal products (meat, milk). less common routes of infection are inhalational and cutaneous. the clinical presentation of brucellosis is highly variable, even after inhalational exposure. the clinical spectrum ranges from asymptomatic seroconversion to severe acute systemic disease. intermediate forms include undulant fever or chronic disease, characterized by presence of brucella in virtually any organ. acute systemic disease is highly incapacitating with high fever, headache, nausea, vomiting, chills, severe sweating, and, in very severe cases, delirium, coma, and death. undulant fever is characterized by relapses of fever, weakness, generalized aching, and headache. chronic infections may have manifestations related to several organ systems such as the gastrointestinal and genitourinary tracts, cns, joints, and bones. [73] [74] [75] developing countries with insufficient water treatment and food security are more vulnerable to enteric bt attack. these agents include shigella dysenteriae, salmonella spp., enterohemorrhagic e. coli, vibrio cholerae, and cryptosporidium parvum. shigella and salmonella have in fact already been used as agents of biorevenge or biopolitics in small-scale attacks: one (shigella) in an office setting by a disgruntled employee and one in oregon by a religious sect that led to nearly 1000 cases of salmonella-related gastroenteritis. 11, 76 these agents are indeed ideal for small-scale attacks since large-scale attacks would require contamination of large water supplies which, because of enormous dilution factors and susceptibility of all these agents (except for c. parvum) to standard chlorinating procedures, would decrease the number of bacteria to below that required to infect large numbers of people. 69 occasional outbreaks of nontyphoidal salmonella and shigella infections occur in the united states. shigella is a highly infectious organism that requires very low numbers (10 2 -10 3 organisms) to provoke clinical disease. the illness caused by shigella and enterohemorrhagic e. coli is explosive and starts with fever, vomiting, severe abdominal cramping, bloody diarrhea, and systemic manifestations such as hypotension, and circulatory collapse if not treated rapidly. both microorganisms produce an exotoxin responsible for most of the systemic manifestations associated. a distinct complication, hemolytic uremic syndrome, occurs in a small percentage of cases, being more common in children younger than 10 years of age, leading to renal failure and hemolysis. salmonella is less infectious and less explosive than shigella, and leads to fever, vomiting, diarrhea, abdominal cramping, and in some cases to typhoidal manifestations. imported cases of v. cholerae have been diagnosed in the united states in the past. however, the disease occurs in southern asia and latin america as large outbreaks. the clinical illness is characterized by explosive watery diarrhea that leads to rapid dehydration and circulatory collapse. c. parvum infections are characterized by watery diarrhea and abdominal cramping for 2 to 3 weeks. the disease is self-limited except in patients with acquired immunodeficiency syndrome (aids) or other conditions of compromise, in whom illness can last for months or years if immune function is not restored. c. parvum is resistant to standard chlorine concentrations in water supplies. 77 the largest outbreak in this country occurred in milwaukee in the early 1990s and was responsible for thousands of cases and increased mortality among those with aids. 69, 78, 79 this section addresses other toxins considered of potential bt use, such as staphylococcal enterotoxin b (seb) and ricin toxin (derived from castor beans, which in turn are the fruit of the ricinus communis plant). the ricin toxin is composed of two glycoproteins of approximately 66,000 kda. 80 the toxin inhibits protein synthesis by blocking elongation factor 2 (ef2) at the ribosomal level. ricin toxin is not a weapon of mass destruction since its lethal dose in humans is much higher than previously believed. however, the use of the toxin in small bt attacks is possible in the tropics because of its ready availability and relatively easy extraction from the beans. clinical presentation depends on the route of administration as does the ld 50 . in cases where large amounts of the toxin are ingested, the manifestations include nausea, vomiting, severe abdominal cramping, rectal hemorrhage, and diarrhea. as the clinical course progresses, anuria, mydriasis, severe headaches, and shock supervene leading to the patient' s demise in 2 to 3 days. clinical manifestations usually appear within 10 hours after ingestion of the toxin. inhalational exposure leads to prominent pulmonary manifestations 8 to 28 hours after exposure and fever, dyspnea, progressive cough, cyanosis, and death. histologically, there is widespread necrosis of pulmonary parenchyma and pulmonary edema. a single case of parenteral intoxication was documented. a defector from bulgaria was injected with a pellet containing ricin from a weapon disguised in an umbrella, resulting in local necrosis, regional lymphadenopathy, gastrointestinal hemorrhage, liver necrosis, nephritis, and dic. 81 staphylococcus aureus enterotoxin b (seb) is a 28-kda, heatstable exotoxin produced by certain strains of s. aureus and is responsible for food poisoning after ingestion of the preformed exotoxin in improperly handled food. in bt scenarios, exposure can occur either by inhalation or ingestion leading to seb food poisoning or seb respiratory syndrome. the toxin is highly incapacitating and not very lethal. the dose that causes symptoms in half of exposed persons and ld 50 differ by a magnitude of 5 log scales for inhalational exposure. 82 thus, it is thought of as an incapacitating agent. incubation time after ingestion is short (4-12 hours) followed by explosive vomiting that persists for several hours. weaponization of the toxin as an aerosol is possible due to its high stability. manifestations after inhalation of the seb are related to the respiratory system and consist of fever, cough, chills, myalgias, chest pain, and pulmonary insufficiency due to alveolar edema. general symptoms and signs are universal and consist of multiorgan failure secondary to a cytokine storm. 25 these toxins are superantigens due to their ability to bind to major histocompatibility complex (mhc) class ii molecules on large numbers of lymphocytes and macrophages, leading to a hyperactivation of the immune system and massive cytokine release including interferon-gamma (ifn-î³), tumor necrosis factor-alpha (tnf-î±), interleukin (il-6), and other mediators such as leukotrienes and histamine. 82 the role of the clinical laboratory in the diagnosis of possible cases related to a bt attack is of utmost importance. 83, 84 on the one hand, standard clinical microbiology laboratories will be receiving specimens for diagnostic purposes, and communication with clinicians regarding their suspicions is critical. certain isolates in the laboratory are not pursued further (bacillus spp. is a classic example) unless specifically requested due to the frequent isolation of contaminants with similar characteristics. in addition, handling of certain specimens will require added biosafety level requirements due to their infectivity (table 119-2) . certain samples will have to be shipped to highly specialized laboratories for initial or further work-up. environmental testing is challenging due to the complexity of the samples to be analyzed. 85, 86 this type of testing takes place in highly specialized laboratories and is not undertaken by the standard clinical microbiology laboratory. the bacterial diseases caused by the bt agents outlined in this chapter, with the exception of c. burnetii and rickettsia spp., can be diagnosed by standard isolation techniques in clinical microbiology laboratories. isolation of rickettsiae and the bt viruses requires specialized laboratories with bsl-3 or bsl-4 biocontainment. 87 serological assays are available for detection of antibodies against all bt agents. however, for many organisms serological assays require the presence of rising antibody titers, and therefore the serologic diagnosis is usually retrospective in nature. for some viral diseases, a reliable diagnosis can be established based on elevation of immunoglobulin m (igm) titers in the acute phase of the disease. with the advent of molecular techniques, rapid and sensitive diagnostic tests are becoming available for bt agents during the acute phase of the disease. [88] [89] [90] this is of utmost importance in a bt event since identification of the first cases would be critical for a rapid and effective public health response. in addition, treatment and prophylactic measures can also be initiated as quickly as possible. molecular diagnostic techniques can be applied to potential bt agents in an additional setting: as part of the epidemiological and forensic investigations that a bt attack would immediately trigger. postmortem diagnosis is also possible by analysis of frozen or paraffin-embedded tissues by immunohistology or nucleic acid-based amplification techniques. rapid diagnosis of the initial case (cases) in a bt event requires a high degree of clinical suspicion from the physicians having contact with such patients in the emergency room or outpatient setting. the clinical laboratories would then play a critical role in detecting the suspected agent and/or referring the appropriate specimens to higher level laboratories for specialized testing (table 119-3) . 83, 85, 91 several of the agents discussed in this chapter are zoonotic diseases. therefore, diagnosis of certain zoonotic diseases in animals may be important in identifying some bt attacks. in such situations, animals could be seen as either direct victims of the attack or as sentinel events in a human outbreak. there are currently efforts to establish a network of laboratories dedicated to diagnosis of veterinary agents. 85 bsl-1 suitable for work involving well-characterized agents not known to cause disease in healthy bacillus subtilis adult humans and of minimal potential hazard to laboratory personnel and the environment. naegleria gruberi canine hepatitis virus bsl-2 suitable for work involving agents of moderate potential hazard to personnel and the measles virus environment. laboratory personnel have specific training in handling pathogenic agents salmonella spp. and are directed by competent scientists; access to the laboratory is limited when work toxoplasma spp. is being conducted; extreme precautions are taken with contaminated sharp items; and hepatitis b virus certain procedures in which infectious aerosols or splashes may be created are conducted in biological safety cabinets or other physical containment equipment. bsl-3 suitable for work with infectious agents which may cause serious or potentially lethal coxiella burnetii disease as a result of exposure by the inhalation route. in addition to the requirements rickettsia spp. described for work in bsl-2 environment, all procedures are conducted within biological m. tuberculosis safety cabinets, or other physical containment devices, and by personnel wearing alphaviruses appropriate personal protective clothing and equipment. laboratory should be located in a separate building or an isolated zone within a building. laboratories are equipped with double door entry, directional inward flow, and single-pass air. bsl-4 required for work with dangerous and exotic agents that pose a high individual risk of filoviruses aerosol-transmitted laboratory infections and life-threatening disease. members of the arenaviruses laboratory staff have specific and thorough training in handling extremely hazardous infectious agents. they are supervised by competent scientists who are trained and experienced in working with these agents. access to the laboratory is strictly controlled by the laboratory director. the facility is either in a separate building or in a controlled area within a building, which is completely isolated from all other areas of the building. all activities are confined to class iii biological safety cabinets, or class ii biological safety cabinets used with one-piece positive pressure personnel suits ventilated by a life support system. the biosafety level 4 laboratory has special engineering and design features to prevent microorganisms from being disseminated into the environment. the diagnosis of inhalational anthrax is based on isolation and identification of b. anthracis from a clinical specimen collected from an ill patient. in cases of inhalational anthrax, samples of sputum, blood, or cerebrospinal fluid (csf) may yield growth of the agent. demonstration of b. anthracis from nasal swabs has more epidemiological and prophylactic implications than clinical importance. standard diagnostic techniques are based on visualization and isolation in the clinical microbiology laboratory and serological demonstration of antibodies against b. anthracis. [92] [93] [94] [95] [96] visualization of b. anthracis from clinical specimens (blood cultures, csf, and cutaneous lesions) by gram stains is not difficult. b. anthracis appears as large gram-positive, spore-forming rods with a bamboo appearance. isolation is achieved by inoculating standard sheep blood agar plates, and colonies appear as small, gray-white, nonhemolytic colonies. a selective medium (polymyxin-lysozyme-edta-thallous acetate agar) is available mostly for environmental samples and inhibits the growth of other bacillus spp., such as b. cereus. growth is rapid (24-48 hours) . 93 confirmatory tests include î³-phage lysis, detection of specific cell wall and capsular antigens, and polymerase chain reaction (pcr) amplification of dna followed by sequencing. 90 serological tests available for clinical diagnosis are based on detection of antibodies directed against protective antigen (pa). cross-reactive antibodies decrease the specificity of this test. assays based on toxin detection are available in specialized centers and are based on capture of anthrax toxins by using antibodies. antibody-coated immunomagnetic beads are then analyzed by electrochemiluminescence technology. the analytical sensitivity of this technique for detection of anthrax toxin is at the picogram to femtogram level (10 â��12 to 10 â��15 ). 97, 98 immunoliposomal technology combined with real-time pcr (for a dna reporter sequence) is also in the early stages of development for several toxins (ricin, cholera, and botulinum) and appears promising with analytical sensitivity in the attomolar to zeptomolar (10 â��18 to 10 â��21 ) range for cholera toxin. 99 the specificity of this assay is given by the toxin-capturing antibody. nucleic acid amplification techniques (pcr) are also available both in standard format and real-time format. extraction of dna from spores is challenging and requires modification of dna extraction protocols in order to facilitate release of dna from spores or induction of germination prior to dna extraction. 90 real-time pcr tests have been developed by applied biosystems (taqman 5' nuclease assay) and roche applied science (lightcycler). [100] [101] [102] the analytical sensitivity of both techniques is extremely high, and testing times have been decreased to 1 to 2 hours. portable pcr instruments are being developed for rapid deployment to the field. 103 examples include the rugged advanced pathogen identification device (rapid), 100 the smartcycler (cepheid, ca), 101 and the miniature analytical thermal cycler instrument (matci) developed by the department of energy' s lawrence livermore national laboratory. 104 this instrument later evolved into the advanced nucleic acid analyzer (anaa) and handheld advanced nucleic acid analyzer (hanaa). 105 molecular subtyping of b. anthracis is also possible by using the 16s ribosomal rna (rrna) subunit gene, multiplelocus vari-able number tandem repeat analysis of eight genetic loci, and amplified fragment length polymorphism (aflp) techniques. 106, 107 environmental testing also plays a role in the investigation of a bt event. in this setting, detection of b. anthracis relies heavily on molecular techniques for confirmation of potentially contaminated samples (e.g., surfaces, air). 108, 109 postmortem diagnosis is also possible by using gram stains on paraffin-based tissues or immunohistochemical procedures using polyclonal or monoclonal antibodies against various anthrax antigens. diagnosis of y. pestis is based on demonstration of the bacillus in blood or sputa from patients. standard diagnostic techniques in the laboratory include visualization of gramnegative coccobacilli, which by giemsa, wright, or wayson stains reveal a "safety pin" appearance. isolation is performed in blood and mcconkey agar plates on which colonies appear as nonlactose fermentors. the organisms are identified preliminarily by direct immunofluorescent assay with y. pestisspecific antibodies, with final identification based on biochemical profiles in clinical microbiology laboratories. 110 molecular diagnostic techniques based on real-time pcr have become available in recent years and involve detection of y. pestis genes such as plasminogen activator (pla), genes coding for the yop proteins and the capsular f1 antigen, and the 23s rrna gene, which allows distinction from other yersinia spp. [111] [112] [113] assays have been developed to detect resistance to particular antibiotics. the importance of these diagnostic techniques in a disease such as plague is evident. the log-normal epidemic curve with a narrow dispersion of the incubation periods (see fig. 119 -1) and the short interval for successful antibiotic therapy mandate recognition of the earliest cases if the bulk of the exposed are to be saved. molecular subtyping of y. pestis is also possible by analyzing polymorphic sites in order to identify the origin of strains in the event of a bt attack. diagnosis is made in the clinical laboratory by demonstration of the microorganisms in secretions (sputa, exudates) by direct immunofluorescence or immunohistochemically in biopsy specimens. isolation in the clinical laboratory may be achieved by using regular blood agar plates, posing a risk to laboratory personnel not employing bsl-3 facilities and procedures. the procedure for isolation of f. tularensis in the laboratory is very similar to that described for y. pestis. final identification in the clinical laboratory is based on the biochemical profile. 114 molecular diagnostic techniques are based on pcr detection of f. tularensis by using primers for different genes such as outer membrane protein (fop) or tul4 and real-time detection systems. 90, 115, 116 smallpox virus diagnosis of variola major is suggested by its clinical presentation and the visualization of guarnieri bodies in skin biopsy samples. preliminary confirmation requires visualization of the typical brick-shaped orthopox virus by electron microscopy, followed by isolation from clinical specimens and accurate molecular identification to differentiate it from the morphologically (and sometimes clinically) similar monkeypox virus. confirmation of this diagnosis is performed only under bsl-4 containment facilities at the cdc. 47 molecular techniques are based on pcr amplification using real-time or standard technology followed by sequencing or use of restriction fragment length polymorphism (rflp) for accurate identification. 117 technologies so far developed for smallpox molecular testing include taqmanand lightcycler-based assays with primers designed for the hemagglutinin gene and a-type inclusion body proteins. [118] [119] [120] [121] sequencing of the smallpox genome has been completed for some asian strains of variola major and one of variola minor. other strains are being sequenced and will provide more information for probe design and treatment targets. 90 diagnosis of these diseases is performed in highly specialized centers in the united states because special isolation procedures and highly contained laboratories are required. initial diagnosis of these diseases is suspected on clinical and epidemiologic grounds. laboratory diagnosis involves isolation, electron microscopy, and serological assays. immunohistochemical detection of hemorrhagic fever viral antigens in paraffin-embedded tissues is also performed in highly specialized centers such as the cdc. [122] [123] [124] [125] [126] molecular diagnostic techniques have also improved dramatically during the last few years. serum or blood is the most common specimen used for reverse transcriptase-pcr amplification of viral nucleic acids. both standard and realtime techniques are available. design of primers for this heterogenous group of rna viruses that are highly variable is one of the limitations. 90 therefore, multiplex pcr techniques are required to detect as many targets as possible in a single assay. 127, 128 real-time pcr based on detection of the target sequence using fluorescent probes therefore limits the number of targets that can be identified because of the limited wavelength range for fluorescent applications (usually only four different wavelengths can be detected at the same time). [128] [129] [130] the use of microchips containing several thousands of oligonucleotides from all viruses known to be pathogenic to humans is an encouraging development. in fact, the rapid identification and characterization of the novel human coronavirus responsible for the sars outbreak in 2003 is an excellent example of the power of hybridization-based microchips. the creation of an automated and easily deployable instrument capable of detecting all possible potential bt agents based on highly sensitive techniques such as electrochemoluminescence (ecl) or pcr would be ideal. the nonspecific nature of presenting symptoms is a major problem with several of the agents. the rapid recruitment of cases into the infected cohort requires that an early diagnosis of the epidemic be established, particularly for organisms such as y. pestis in which there is only a short window for successful treatment. in fact, such projects are already in the making. an example of this system is the automated biological agent testing system (abats) that combines the techniques mentioned previously. 86 the system is the result of integrating several commercially available technologies into a single automated and robotized instrument for detection of viruses, bacteria, and parasites considered potential bt agents. the technologies incorporated into this "super system" include automated specimen preparation (both nucleic acid-based and protein-based such as immunodiagnostics), thermocyclers for pcr detection, chemiluminescent detectors for immunobased assays, sequencers, and software programs for sequence analysis. rickettsia prowazekii (epidemic typhus) and r. rickettsii (rocky mountain spotted fever) diagnosis of these infections in the clinical microbiology laboratory currently rests on the identification of antibodies in serum during the acute and convalescent period in order to demonstrate seroconversion or rising titers. the diagnosis is therefore retrospective. 131, 132 detection of rickettsial dna from blood or skin samples during the acute phase of the disease is possible via pcr assays. however, these assays are not standardized and are not commercially available. primers have been designed for amplification of several rickettsial genes including citrate synthase, 17-kda protein gene, ompa, and ompb. [132] [133] [134] [135] [136] the clinical sensitivity and specificity of standard or real-time pcr techniques have not been determined. most likely real-time pcr is superior due to the higher analytical sensitivity of this technique and low risk of sample contamination with dna amplicons when compared to standard pcr amplification methods. isolation of rickettsiae from clinical specimens is performed in very few specialized laboratories in the nation and requires the use of cell monolayers, embryonated eggs, or animals. detection of rickettsial antigens or whole bacteria in blood specimens is theoretically possible by using ultrasensitive methods, but such assays are currently only in the early phases of development. immunohistochemical detection of rickettsiae in paraffin-embedded tissue has also been applied to tissue samples obtained pre-or postmortem. [137] [138] [139] salmonella spp., shigella dysenteriae, vibrio cholerae, and cryptosporidium parvum (acute enteric syndromes) diagnosis of salmonella, shigella, and vibrio infections is based on isolation of the offending agent on standard microbiological media in the clinical laboratory, followed by specialized confirmatory tests to identify the specific serotype involved. 140 diagnosis of c. parvum is based on visual identification of the protozoan in fecal specimens by using modified trichrome stain. 140 the diagnosis rests on serological demonstration of antibodies by immunofluorescent assay (ifa) or enzyme-linked immunosorbent assay (elisa). antibodies remain elevated for years after the acute infection, and therefore a fourfold rise in titers is the gold standard for diagnosis. pcr detection of c. burnetii dna from blood or tissues also yields a diagnosis of q fever. 88 brucella spp. diagnosis of brucellosis requires a high degree of clinical suspicion due to the protean manifestations related to this disease. laboratory diagnosis is based on isolation of the microorganism from blood, bone marrow, or other tissue samples. isolation is not easy due to the slow-growth of brucella spp. colonies usually appear after 4 to 6 weeks, and therefore communication with the clinical laboratory is important so that appropriate media will be used and the cultures will be held long enough for colonies to be detected. 90 serologic assays for demonstration of rising antibody titers are available, although the diagnosis is retrospective. pcr detection is promising, but it is not standardized. [141] [142] [143] alphaviruses (encephalitic syndromes: venezuelan, eastern, and western equine encephalomyelitis) diagnosis is based on isolation of the virus from serum or brain (postmortem specimens) in a bsl-3 environment. pcr detection of viral sequences is also possible. serologic diagnosis is based on demonstration of antibodies in acute and convalescent sera. [144] [145] [146] botulinum toxins the diagnosis of botulism relies heavily on clinical parameters. an afebrile patient with signs and symptoms of progressive bulbar palsies and descending neuromuscular paralysis is highly suspected of having botulism. demonstration of the toxin in cases of botulism due to ingestion of contaminated food is made from gastric samples, feces, blood, and urine. however, detection of minute amounts of toxin (and contacts with samples from cases may prove fatal due to the toxin' s potency) would be difficult by current immunoassay systems such as elisa platforms. 146 detection techniques based on electrochemiluminescence and immunoliposomes are currently under development. 99, 147 pcr assays can be performed in cases of ingestion of contaminated food in order to detect the genetic material present in c. botulinum. if weaponized toxin is used in the absence of c. botulinum organisms, detection of the genetic material would be difficult and would rely on the presence of residual dna after toxin purification procedures. if inhalational botulism is suspected, respiratory secretions and nasal swabs should be obtained as early as possible. postmortem samples of liver and spleen can be used for detection of botulinum toxins. diagnosis is also based on clinical presentation and requires a high index of suspicion due to the nonspecific nature of the signs and symptoms. laboratory diagnosis rests on detection of the toxin in body fluids by immunoassays (capture elisa and igg elisa). 146 a new generation of tests using more sensitive detection methods is under development (see preceding discussion). diagnosis is also suspected on clinical grounds and confirmed by demonstration of the toxin in nasal swabs early in the disease process, feces, and, in fatal cases, from kidney and lung tissue. serum can be analyzed by elisa, and pcr can be performed for detection of toxin genes of s. aureus if present. 146 the distribution and incubation periods of infectious diseases syndromic surveillance and bioterrorism-related epidemics viral hemorrhagic fevers anthrax bioterrorism: lessons learned and future directions investigation of bioterrorism-related anthrax history of the development and use of biological weapons biological warfare. a historical perspective [see comment laboratory aspects of biowarfare. philadelphia, wb saunders mass casualty management of a large-scale bioterrorist event: an epidemiological approach that shapes triage decisions a large community outbreak of salmonellosis caused by intentional contamination of restaurant salad bars a systems overview of the electronic surveillance system for the early notification of community-based epidemics (essence ii) epidemiologic response to anthrax outbreaks: field investigations bioterrorism-related inhalational anthrax: the first 10 cases reported in the united states first case of bioterrorism-related inhalational anthrax in the united states quantitative pathology of inhalational anthrax i: quantitative microscopic findings the sverdlovsk anthrax outbreak of 1979 death at sverdlovsk: what have we learned? anthrax as a biological weapon: medical and public health management. working group on civilian biodefense updated recommendations for management pathology of experimental inhalation anthrax in the rhesus monkey the pathology of experimental anthrax in rabbits exposed by inhalation and subcutaneous inoculation pathology of inhalation anthrax in cynomolgus monkeys (macaca fascicularis) fatal inhalational anthrax in a 94-year-old connecticut woman [see comment clinical recognition and management of patients exposed to biological warfare agents clinical presentation of inhalational anthrax following bioterrorism exposure: report of 2 surviving patients [see comment recognition and treatment of anthrax pathology of inhalational anthrax in 42 cases from the sverdlovsk outbreak of 1979 pathology and pathogenesis of bioterrorism-related inhalational anthrax yersinia species, including plague yersinia infections: centennial of the discovery of the plague bacillus the bubonic plague factories of death clinical recognition and management of patients exposed to biological warfare agents plague as a biological weapon: medical and public health management. working group on civilian biodefense [see comment epidemiologic determinants for modeling pneumonic plague outbreaks manual of clinical microbiology clinicopathologic aspects of bacterial agents tularemia as a biological weapon: medical and public health management [see comment francisella tularensis tularemia: a 30-year experience with 88 cases the world' s last endemic case of smallpox: surveillance and containment measures virological evidence for the success of the smallpox eradication programme eradication of infectious diseases: its concept, then and now smallpox and its eradication. geneva, world health organization smallpox: an attack scenario smallpox as a biological weapon: medical and public health management. working group on civilian biodefense monkeypox transmission and pathogenesis in prairie dogs potential antiviral therapeutics for smallpox, monkeypox, and other orthopox virus infections viral hemorrhagic fevers including hantavirus pulmonary syndrome in the americas viral agents as biological weapons and agents of bioterrorism hemorrhagic fever viruses as biological weapons: medical and public health management [see comment role of the endothelium in viral hemorrhagic fevers the action of botulinum toxin at the neuromuscular junction botulinum toxins venezuelan equine encephalitis an outbreak of eastern equine encephalomyelitis in jamaica, west indies. i: description of human cases eastern and western equine encephalitis an outbreak of venezuelan equine encephalomeylitis in central america. evidence for exogenous source of a virulent virus subtype an epidemiologic study of venezuelan equine encephalomyelitis in costa rica venezuelan equine encephalitis zinsser h: rats, lice and history rickettsial infections. in lack ee rocky mountain spotted fever and other rickettsioses typhus and its control in russia rickettsia conorii infection of c3h/hen mice. a model of endothelial-target rickettsiosis rickettsioses as paradigms of new or emerging infectious diseases threats in bioterrorism ii: cdc category b and c agents q fever 1985-1998. clinical and epidemiologic features of 1,383 infections epidemiologic features and clinical presentation of acute q fever in hospitalized patients: 323 french cases human brucellosis an overview of human brucellosis medical aspects of chemical and biological warfare an outbreak of shigella dysenteriae type 2 among laboratory workers due to intentional food contamination viability of cryptosporidium parvum oocysts in natural waters surveillance for waterborne-disease outbreaks-united states cryptosporidiosis in children during a massive waterborne outbreak in milwaukee, wisconsin: clinical, laboratory and epidemiologic findings medical aspects of chemical and biological warfare georgi markor--death in a pellet staphylococcal enterotoxin b and related pyrogenic toxins bioterrorism: implications for the clinical microbiologist the role of the clinical laboratory in managing chemical or biological terrorism diagnostic analyses of biological agent-caused syndromes: laboratory and technical assistance automated biological agent testing systems department of health and human services applying molecular biological techniques to detecting biological agents current laboratory methods for biological threat agent identification molecular diagnostic techniques for use in response to bioterrorism a national laboratory network for bioterrorism: evolution from a prototype network of laboratories performing routine surveillance definitive identification of bacillus anthracis-a review pc bacillus and other aerobic endospore-forming bacteria mabs to bacillus anthracis capsular antigen for immunoprotection in anthrax and detection of antigenemia specific, sensitive, and quantitative enzyme-linked immunosorbent assay for human immunoglobulin g antibodies to anthrax toxin protective antigen comparison of a multiplexed fluorescent covalent microsphere immunoassay and an enzyme-linked immunosorbent assay for measurement of human immunoglobulin g antibodies to anthrax toxins an enzymatic electrochemiluminescence assay for the lethal factor of anthrax comparative studies of magnetic particle-based solid phase fluorogenic and electrochemiluminescent immunoassay high-sensitivity detection of biological toxins a field investigation of bacillus anthracis contamination of u.s. department of agriculture and other washington, dc, buildings during the anthrax attack of sensitive and rapid identification of biological threat agents detection of bacillus anthracis dna by lightcycler pcr a handheld real time thermal cycler for bacterial pathogen detection real-time microchip pcr for detecting single-base differences in viral and human dna rapid pathogen detection using a microchip pcr array instrument genetic sleuths rush to identify anthrax strains in mail attacks sequencing of 16s rrna gene: a rapid tool for identification of bacillus anthracis environmental sampling for spores of bacillus anthracis bacillus anthracis contamination and inhalational anthrax in a mail processing and distribution center manual of clinical microbiology development of rrnatargeted pcr and in situ hybridization with fluorescently labelled oligonucleotides for detection of yersinia species 5' nuclease pcr assay to detect yersinia pestis yersinia pestis-etiologic agent of plague francisella and brucella detection of francisella tularensis in infected mammals and vectors using a probe-based polymerase chain reaction detection of francisella tularensis within infected mouse tissues by using a hand-held pcr thermocycler pcr strategy for identification and differentiation of small pox and other orthopoxviruses real-time pcr system for detection of orthopoxviruses and simultaneous identification of smallpox virus gene for a-type inclusion body protein is useful for a polymerase chain reaction assay to differentiate orthopoxviruses the potential of 5â�² nuclease pcr for detecting a single-base polymorphism in orthopoxvirus detection of smallpox virus dna by lightcycler pcr comparative pathology of the diseases caused by hendra and nipah viruses a novel immunohistochemical assay for the detection of ebola virus in skin: implications for diagnosis, spread, and surveillance of ebola hemorrhagic fever. commission de lutte contre les epidemies a kikwit immunohistochemical and in situ localization of crimean-congo hemorrhagic fever (cchf) virus in human tissues and implications for cchf pathogenesis retrospective diagnosis of hantavirus pulmonary syndrome, 1978-1993: implications for emerging infectious diseases hantavirus pulmonary syndrome. pathogenesis of an emerging infectious disease molecular diagnostics of viral hemorrhagic fevers rapid detection and quantification of rna of ebola and marburg viruses, lassa virus, crimean-congo hemorrhagic fever virus, rift valley fever virus, dengue virus, and yellow fever virus by real-time reverse transcription-pcr quantitative real-time pcr detection of rift valley fever virus and its application to evaluation of antiviral compounds development and evaluation of a fluorogenic 5' nuclease assay to detect and differentiate between ebola virus subtypes zaire and sudan rickettsioses as paradigms of new or emerging infectious diseases laboratory diagnosis of rickettsioses: current approaches to diagnosis of old and new rickettsial diseases citrate synthase gene comparison, a new tool for phylogenetic analysis, and its application for the rickettsiae diagnosis of mediterranean spotted fever by cultivation of rickettsia conorii from blood and skin samples using the centrifugation-shell vial technique and by detection of r. conorii in circulating endothelial cells: a 6-year follow-up differentiation of spotted fever group rickettsiae by sequencing and analysis of restriction fragment length polymorphism of pcr-amplified dna of the gene encoding the protein rompa differentiation among spotted fever group rickettsiae species by analysis of restriction fragment length polymorphism of pcr-amplified dna diagnostic tests for rocky mountain spotted fever and other rickettsial diseases immunohistochemical diagnosis of typhus rickettsioses using an anti-lipopolysaccharide monoclonal antibody monoclonal antibody-based immunohistochemical diagnosis of rickettsialpox: the macrophage is the principal target investigation of foodborne and waterborne disease outbreaks rapid laboratory confirmation of human brucellosis by pcr analysis of a target sequence on the 31-kilodalton brucella antigen dna the 18-kda cytoplasmic protein of brucella species-an antigen useful for diagnosis-is a lumazine synthase characterization of an 18-kilodalton brucella cytoplasmic protein which appears to be a serological marker of active infection of both human and bovine brucellosis genus-specific detection of alphaviruses by a semi-nested reverse transcription-polymerase chain reaction standardization of immunoglobulin m capture enzyme-linked immunosorbent assays for routine diagnosis of arboviral infections toxins as weapons of mass destruction: a comparison and contrast with biological warfare and chemical warfare agents sensitive detection of biotoxoids and bacterial spores using an immunomagnetic electrochemiluminescence sensor key: cord-018463-a6qu0cuv authors: wimmer, eckard title: synthetic biology, dual use research, and possibilities for control date: 2018-03-23 journal: defence against bioterrorism doi: 10.1007/978-94-024-1263-5_2 sha: doc_id: 18463 cord_uid: a6qu0cuv the anthrax attack on the human population in the united states in 2001/2002 may be considered the naissance of modern bioterrorism. this attack, e.g. the planned killing by means of deadly microorganisms (bacillus anthracis) caused enormous public concern, because, numerous other deadly agents, now known as “select agents”, occur in nature and are available for misuse. the anthrax attack coincided with the first report in 2002 of the de novo synthesis in the test tube of a pathogenic human virus, poliovirus, that was equally shocking because it indicated that dangerous infectious agents could be produced in laboratories outside of government control. these events were synchronous with the advent of a new discipline, synthetic biology, which was an emerging area of research that can broadly be described “as the design and construction of novel artificial biological pathways, organisms or devices, or the redesign of existing natural biological systems.” the synthesis of viruses, or more broadly expressed: each experiment in synthetic biology, fits the definition of “dual use research” – the dual use dilemma in which the same technologies can be used for the good of humans and misused for bioterrorism. in view of these threats the us government has formulated rules that can lower the chances of misuse of biological research. that includes all research with select agents or the modification of agents to acquire dangerous traits (“gain of function”). it also calls for the continuous education of all generations entering research: to be aware that results of research can be dangerous, if not immediately then possibly at later times. on september 11, 2001, the unites states suffered the deadliest (non-military) terrorist attack of all times when high jacked airline jets crashed into the world trade center towers in new york city, into the pentagon in washington, dc, and into a field in rural pennsylvania. about seven days after the 9/11 attack, anonymous letters laced with deadly spores of the bacillus anthracis (anthrax) arrived in offices of media companies and the us congress. during the following 2 months at least ten letters, anonymously mailed, were identified causing at least 17 anthrax infections and altogether 5 casualties. this was the first incidence of what seemed to be the planned killing by means of deadly organisms and, as such, may be considered the emergence of modern bioterrorism. the attack triggered a nine-year investigation that consumed many millions of dollars and was extraordinary for its magnitude [1] . the 9/11 terrorist attack claimed >3000 casualties and changed life in the us. the anthrax attack, in contrast, caused "only" 5 casualties but it was a most frightening wake-up call for the public. bioterrorism executed with an anonymous, highly dangerous, self-replicating infectious agent, which cannot be detected even with high-powered light microscopes (e.g. viruses), suddenly presented to the society a new vulnerability that could cause deadly chaos. to lower the risk of such scenario all dangerous microorganisms, known now as "select agents", are currently controlled for research by the government and their use is subject to strictest rules. the publication in july 2002 of the first chemical synthesis of poliovirus [2, 3] , an organism harmful for humans, struck a sensitized and anxious american public after 9/11. it became instantaneously clear that viruses like poliovirus no longer "exist" only in nature but also in computers. viruses, therefore, can be synthesized using the information stored in computers. these synthetic viruses have the computer as parent -a natural isolate is no longer required. this scenario immediately brings us to an ancient dilemma in science called "dual use research." every result produced in science has the propensity to be useful and/or harmful for humankind. in addition, every progress in biomedical technology yields the possibility to achieve new "unheard-of" results, like the chemical synthesis of poliovirus. currently, there are the genomes of >2,500 viruses available in public databases. all of these sequences could be used to synthesize the corresponding viruses, including small pox virus. that is the dark side of progress. however, true to the tenet of "dual use research": the de novo synthesis of poliovirus has also lead to the development of very promising new strategies to develop vaccines [4, 5] that will hopefully aid humankind in the future. since the first ground breaking synthesis of poliovirus in 2002, some very dangerous viruses have been "re-created" in the test tube in the absence of natural isolates [6] . most notably was the resurrection of the type a strain of influenza virus 1918/1919 that killed an estimated 20-50 million humans [7] . the virus vanished after the catastrophic pandemic, leaving no trace for investigations that would explain its extraordinary virulence. studies carried out after its de novo synthesis in 2005, however, have identified the murderous armor of this virus and, thus, prepared medical scientists for its possible recurrence. these results justified the enormous efforts to "revive" this terrible virus. the revival of another lost virus in the absence of the natural template was described in a recent, unpublished report [8] . briefly, horsepox virus, an orthopoxvirus , thought to be extinct from nature, has been synthesized de novo. this virus, which is related to the human smallpox virus, may have been part of the original vaccine against human smallpox and, therefore, it may be highly useful for improving the current smallpox vaccine. please note that in the history of humankind smallpox virus is considered the most vicious killer of all time. it was globally eradicated in 1980. not surprisingly then, the report that the related horsepox virus was rebuild by genetic engineering fueled again the debate about the benefits and risks in biomedical research [8] . in the mid 2010s these events coincided with the advent of a new discipline, synthetic biology. synthetic biology is an emerging area of research that can broadly be described "as the design and construction of novel artificial biological pathways, organisms or devices, or the redesign of existing natural biological systems" [9] . about 12 years ago, the idea of combining molecular genetics, genetic engineering, cell biology, mathematics and engineering has been heralded as the dawn of a new science with unlimited possibilities to create new artificial biological systems useful for humans and the environment. it has been greeted with skepticism by a nervous general public as these goals could lead also to new bioterrorist agents. are there sure measures for preventing misuse by modern biomedical techniques leading to results harmful for humans? an answer provided by joshua lederberg (1998 nobel prize laureate in medicine) is sobering: "there is no technical solution to the problem of biological weapons. it needs an ethical, human and moral solution -if it is going to happen at all" [10] . it is worth noting that this crucial statement was made just 3 years before the anthrax attack. listed below are some constraints that show how in the us the development of dangerous infectious agents, referred to as "select agents", is controlled -perhaps misuse even prevented -through technical and administrative hurdles: i. re-creating an already existing dangerous virus for malicious intent is a complex scientific endeavor. (i) it requires considerable scientific knowledge and experience and, more importantly, considerable financial support. that support usually comes from government and private agencies (nih, naf, etc.), organizations that carefully screen at multiple levels all applications for funding of all biological research. (ii) it requires an environment suitable for experimenting with dangerous infectious agents (containment facilities). any work in containment facilities is also carefully regulated. ii. genetic engineering to synthesize or modify organisms relies on chemical synthesis of dna. synthesizing dna is automated and carried out with sophisticated, expensive instruments. the major problem of dna synthesis, however, is that the product is not error-free. any single mistake in the sequence of small dna segments (30-60 nucleotides) or large segments (>500 nucleotides) can ruin the experiment. companies have developed strategies to produce and deliver error free, synthetic dna, which investigators can order electronically from vendors, such as integrated dna technologies (us), genscript (us) or geneart (germany). this offers a superb and easy way to control experimental procedures carried out in any laboratory: the companies will automatically scan ordered sequences in extensive data banks to monitor relationship to sequences of a select agent. if so, the order will be stalled until sufficient evidence has been provided by the investigator that she/he is carrying out experiments approved by the authorities. the entire complex issue of protecting society from the misuse of select agents has been discussed in two outstanding studies [11, 12] . iii. engineering a virus such that it will be more harmful (more contagious, more pathogenic) is generally difficult because, in principle, viruses have evolved to proliferating maximally in their natural environment. that is, genetic manipulations of a virus often lead to loss of fitness that, in turn, is unwanted in the bioterrorist agent. a special case, however, is research leading to "gain-of-function" of microorganisms, an objective broadly defined by the us government as follows: gain-of-function studies, or research that improves the ability of a pathogen to cause disease, help define the fundamental nature of human-pathogen interactions, thereby enabling assessment of the pandemic potential of emerging infectious agents, informing public health and preparedness efforts, and furthering medical countermeasure development. gain-of-function studies may entail biosafety and biosecurity risks; therefore, the risks and benefits of gain-of-function research must be evaluated, both in the context of recent u.s. biosafety incidents and to keep pace with new technological developments, in order to determine which types of studies should go forward and under what conditions. us. government, october 17, 2014. the debate about gain-of-function research became particularly intense in 2013 when benefits vs risks were broadly assessed in studies of the host range change, increased transmissibility and pathogenicity of non-human infectious agents such as sars, mers or animal influenza virus strains. indeed, following an intense debate, all experiments aimed at "gain-of-function" of infectious agents were put on hold but have since been re-assessed [13] . if large, secret organizations (like the "aum shinrikyo" in japan) or governments of hostile countries would plan to develop and apply agents of bioterrorism the opportunities to interfere with such activities would be limited. therefore, it is imperative to train young scientists to be aware of molecular techniques to modifying existing dangerous organisms or generate an infectious agent with properties of a select agent. this training must also include education in the ethics of science -preparing young scientists to share responsibilities to work only for the good of humankind. obviously, this is a long-term goal. in the short-term, vigilance and communication are our best defense [14] [15] [16] . amerithrax or anthrax investigation (2010) fbi.gov chemical synthesis of poliovirus cdna: generation of infectious virus in the absence of natural template the test-tube synthesis of poliovirus: the simple synthesis of a virus has far reaching societal implications virus attenuation by genome-scale changes in codon pair bias large-scale recoding of an arbovirus genome to rebalance its insect versus mammalian preference the future of synthetic virology characterization of the reconstructed 1918 spanish influenza pandemic virus how canadian researchers reconstituted an extinct poxvirus for $100,000 using mail-order dna the fourth international meeting on synthetic biology (sb4.0) sequence-based classification of select agents: a brighter line synthetic genomics: options for governance national academy of sciences. see <gain-of-function research nasp 2016> case studies of dual use research presidential-commission-for-the-study-of-bioethical-issues-03.10.10.pdf 16. proposed framework for the oversight of dual use life sciences research: strategies for minimizing the potential misuse of research information key: cord-324656-6xq5rs0u authors: bellika, johan gustav; hasvold, toralf; hartvigsen, gunnar title: propagation of program control: a tool for distributed disease surveillance date: 2006-04-18 journal: int j med inform doi: 10.1016/j.ijmedinf.2006.02.007 sha: doc_id: 324656 cord_uid: 6xq5rs0u purpose: the purpose of the study was (1) to identify the requirements for syndromic, disease surveillance and epidemiology systems arising from events such as the sars outbreak in march 2003, and the deliberate spread of bacillus anthracis, or anthrax, in the us in 2001; and (2) to use these specifications as input to the construction of a system intended to meet these requirements. an important goal was to provide information about the diffusion of a communicable disease without being dependent on centralised storage of information about individual patients or revealing patient-identifiable information. methods: the method applied is rooted in the engineering paradigm involving phases of analysis, system specification, design, implementation, and testing. the requirements were established from earlier projects’ conclusions and analysis of disease outbreaks. the requirements were validated by a literature study of syndromic and disease surveillance systems. the system was tested on simulated ehr databases generated from microbiology laboratory data. results: a requirements list that a syndromic and disease surveillance system should meet, and an open source system, “the snow agent system”, has been developed. the snow agent system is a distributed system for monitoring the status of a population's health by distributing processes to, and extracting epidemiological data directly from, the electronic health records (ehr) system in a geographic area. conclusions: syndromic and disease surveillance tools should be able to operate at all levels in the health systems and across national borders. such systems should avoid transferring patient identifiable data, support two-way communications and be able to define and incorporate new and unknown diseases and syndrome definitions that should be reported by the system. the initial tests of the snow agent system shows that it will easily scale to national level in norway. the severe acute respiratory syndrome (sars) outbreak in 2003 showed that existing electronic health record (ehr) syndromic or disease surveillance system should therefore serve all levels, from the local healthcare service to the national and global level [2, 3] . syndromic or disease surveillance has two distinct phases: a detection phase, the goal of which must be to detect an outbreak as early as possible, and a monitoring phase, where the actual disease is known and the goal is to prevent the disease from spreading. in the latter case, it is important to detect new cases and to track existing cases of the disease. to enable early detection of disease outbreaks and the monitoring of the diffusion of a disease through the identification of abnormal disease patterns, "sensors" in the form of a software system should be distributed to all the points in the healthcare system where patients present, or where they leave traces of evidence of illness [4] . lober et al. [5] provide a good description of the information sources that can be monitored to detect an outbreak as early as possible. where affected people present may depend on the type of disease and on the organisation of the healthcare service. in norway, which has a system of personal general practitioners (gps) with gatekeeper functions to secondary and tertiary levels of care, patients will typically consult their gp or the casualty clinic in the first instance. patients are only likely to present at a hospital emergency unit if they are acutely ill (for example, if they have been injured in an accident, or referred to the casualty clinic by the gp). a telephone survey in the usa, in new york, showed that 29.1% of persons with influenzalike illness (ili) visited a physician, 21.4% called a physician for advice, 8.8% visited the emergency department, and 3.8% called a nurse or health hotline [6] . the ehr systems used by gps are therefore likely to be very good sources for detecting disease outbreaks in a healthcare context where most people normally consult their gp in the first instance if they get ill. however, using patient data stored in the ehr system is difficult because of the privacy issues involved in exporting data from the ehr systems [7] [8] [9] [10] [11] . the ehr systems used by gps in norway use the international classification of primary care (icpc), which is more specifically symptom-related than icd-9 and icd-10, making it more applicable for syndromic and disease surveillance use. unfortunately, there is no messaging standard such as hl7 available for primary care services or hospitals in norway, which could enable the use of an available open source solution for syndromic or disease surveillance [2, 11, 12] . syndromic or disease surveillance is relatively easy to implement for all well-defined and existing diseases. if a new disease appears, as in the case of sars, or as result of a deliberate spread, disease surveillance may become more difficult. first, we do not have codes for specifying confirmed, probable or possible cases of the new disease. this may easily be solved if it is sufficient merely to update the coding system. how fast such an update can be performed depends on a number of factors: how fast can the coding system be updated, how fast can the new version be disseminated, how fast can the updated code-set be employed, etc. updating a disease classification is normally done very infrequently and the reliability of such classifications is therefore problematic when the goal is to count occurrences of a new disease. the creation of an additional coding system for contagious diseases, which enables quick updating, may provide a solution to this prob-lem. however, this is also problematic, because duplication of coding systems adds to the quantity of information, which must be entered into the ehr systems. basing surveillance of new diseases on existing classification categories is therefore problematic and adds to the complexity of the problem. however, without such disease codes we cannot automatically count the number of cases and get a picture of the geographical diffusion of the disease. another solution may be to use a list of prodromes or early symptoms (syndrome definition) as basis for definitions of possible and probable cases, but we cannot guarantee that a new disease will be covered by such predefined lists of symptoms. conditions with similar symptom patterns may be interpreted as possible or probable cases, thereby polluting the diffusion pattern. in the case of sars, a visit or contact with a person who had visited a geographic area where there were reported cases was a main indicator of probable or possible cases. such conditions are not known beforehand [13] . in the case of sars, the definition of possible and probable cases evolved over the weeks following the outbreak from initially being very vague to becoming a very specific set of symptoms. it may therefore be necessary to distribute new and updated syndrome definitions if local classification of cases based on the syndrome definition is required. this dynamic behaviour of a disease or syndrome definition makes it hard to build software that automatically identifies and distinguishes between possible, probable and confirmed cases. there are at least two approaches to building a surveillance system: a centralised and a distributed approach. all of the systems reported by lober et al. [5] and kun et al. [14] are based on the centralised storage and analysis of data. these systems have been classified as using first-or second-generation architectures for data integration [9] . the real-time outbreak and disease surveillance system (rods) from the university of pittsburgh [12] also uses a centralised approach. the emergency departments covered by the system send hl7 messages containing icd-9 codes in real time to a centralised inference system where outbreaks are detected. when a disease outbreak is detected by rods, the appropriate authorities are informed about the situation. the outbreak detection process is run periodically on the data available to the system. the bio-surveillance, analysis, feedback, evaluation and response (b-safer) system [2, 15] is based on a federated approach involving the open source openemed infrastructure. openemed [16] (formerly known as telemed [17] ) is based on the clinical observation access service (coas) [18] and patient identification data service (pids) [19] corbamed specification of the object management group. b-safer and openemed take a more interoperable approach that demonstrates the value of the federated approach to syndromic surveillance, which makes it possible to view and manage information at several levels ranging from local to regional, national and global levels [2] . we have constructed a system, which we have named "the snow agent system", after the famous dr. john snow. in the snow agent system we also take a distributed approach, where all contributors of data to the system also have the opportunity to access information about the current situation in a geographic area directly from the system. in that sense, the snow agent system is a true "peer-to-peer" net-work among general practitioners. according to lober's classification [9] , this system would be classified as using thirdgeneration data integration. the snow agent system does not depend on the centralised storage of patient information, and may therefore be used by any group of health institutions or general practitioners that want to share data about the health of the population, without exposing any patientsensitive information. it is designed for everyday use by gps who want to share information about contagious diseases and disease outbreaks in their local community. it also avoids the problematic privacy issues, because no patient-identifiable information ever leaves the ehr system. it also supports the requirement of regular usage, which is important in being prepared for an emergency [3, 20] . the snow agent system has a number of other applications besides syndromic or disease surveillance and epidemiology. most important is the extraction, collection and visualisation of a patient's health record from distributed ehr systems for authorised persons. the system also makes it possible to generate statistics about the activity in the health institutions covered by the system. the messaging system used by the snow agent system may also be used for transferring any kind of data between health institutions, which is a valuable spin-off. however, in this paper we focus on the requirements that we argue syndromic, disease surveillance and epidemiology systems may need to meet when facing a local, national or a pandemic outbreak of a communicable disease. some of these requirements also correspond to the requirements for extraction, collection and visualisation of a patient's health record. the paper presents the core principle utilised in building a distributed system for monitoring and interrogating the status of the population's health, and discusses the results of the initial system tests. the work reported here has been conceived within the engineering paradigm. the processes involved in building a system can be divided into the distinctive phases of analysis, system specification, design, implementation and testing. the system described here has gone through several iterations of analysis, system specification, design, implementation, and testing. the requirements listed in the section below are derived from the following sources: (1) a project that evaluated the needs and conditions for access to patient information from geographically distributed electronic health records [21] ; (2) an analysis of the sars outbreak in march 2003; (3) an analysis of a whooping cough outbreak in troms county in north norway in october/november 2003. the requirements were validated and augmented by a literature study of syndromic and disease surveillance systems and solutions directed towards early detections of bio-terrorist attacks [2] [3] [4] [5] [8] [9] [10] [11] [12] [13] [14] [15] 20, [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] . the snow agent system is a redesign and a complete reimplementation of the virtual secretary system [41] in terms of the requirements reported here. the initial virtual secretary system was developed at the department of computer science at the university of tromsø. the current version is built on top of the jabber instant messaging and presence system. (see details in section 4, "propagation of program control" below.) to test the system components, we implemented a distributed epidemiology service which extracts epidemiological data from simulated ehr databases. the test data used for constructing the ehr databases is described in the next section. the hardware and software used for system testing is described in section 2.2 below. getting access to production data and system is problematic from both a privacy and resource point of view. we therefore chose to create fictitious ehr databases based on data from the microbiology laboratory at the university hospital of north norway. this laboratory serves a large geographical area and many gp clinics, and therefore has data that is well suited to our needs. the data used in the system test was laboratory test data from the microbiology lab. from the pertussis lab dataset we constructed a very limited ehr database by using a number of assumptions: 1. we assume that it takes 1 day to deliver the test sample to the hospital laboratory. the consultation date with the gp is therefore test received date minus 1 day, and consultation time equals test received time. 2. we assume that all gps have the same behaviour regarding entering data into the electronic patient record (ehr) system. 3. they all create a consultation entry in the ehr system using the date and time values specified above. 4. they all do a test from the posterior nasopharynx using darcon or calcium alginate swabs, which they send to the hospital laboratory and order a "b. pertussis pcr" (bordertella pertussis) laboratory test (icpc2 process code-33 microbiological/immunological test) (code 6232 and 6192 in the lab system). 5. if the lab report is negative, the gp does nothing. if the lab test is positive, the doctor creates a telephone consultation with the patient using "r71 whooping cough" as the diagnosis, and prescribes erythromycin. five days later the patient is assumed not to be contagious to others. (pertussis is most contagious during the first week. three weeks after the onset of symptoms, the patient is seldom contagious.) 6. the "x days since onset" value for the whooping cough patients will probably vary widely. we assume the distribution of x is gaussian. we assume that most people will not see their gp before the second stage of the disease has started (7-14 days after onset of symptoms), when the severe cough with whooping and vomiting begins. this value may also vary with the age and gender of the patient. we assume it is likely that babies under the age of one (below 1 year in the test dataset) probably see their doctor earlier than older children do. we also assume male patients wait longer to see their gp than female patients do. knowledge about an outbreak probably also affects how early a patient sees their gp. we assume that children <2 years see their gp 2-5 days after onset of symptoms, children ≥2, <8 years see their gp 3-7 days after onset, and that children ≥8, <18 years see their gp 5-10 days after onset. female ≥18 years see their gp 10-15 days after onset and male >18 years see their gp 12-20 days after onset. from negative laboratory reports, the following ehr documents were created: first consultation, lab request and lab report. from positive laboratory reports, the following ehr documents were created: first consultation, lab request, lab report, second consultation (telephone), drug prescription (erythromycin). the dataset contained 7939 rows of lab results from 3997 different patients. the dataset contained 6463 negative lab result records and 198 lab result records with "pos" results, 794 patients with test result = "neg" (non-negative blood sample test results) and 194 patients with a "pos" test result. the dataset contained 487 test requesters, including hospital internal requesters. of these test requesters, 459 were from outside the hospital and 28 from inside the hospital. the dataset covered 372 postcodes, 153 postcodes with test result = "neg" (non-negative blood sample test results) and 54 postcodes with test result = "pos". the dataset contained patients from 138 municipalities, where 62 municipalities had test result = "neg" (non-negative blood sample test results) and 25 municipalities with test result = "pos". there were 4528 female patients and 3411 male patients in the data set, including 103 female patients and 95 male patients with a "pos" test result, and 779 female patients and 697 male patients with test result = "neg" (non-negative blood sample test results). in the system test, we only used confirmed cases as test data (test result = "pos"). in addition to the laboratory dataset, a map was needed, on which to plot the data. the map was created from bitmaps made available in the norgeshelsa system [42] . the bitmaps were first converted to svg [43] polygons. then the map polygons were inserted into a mysql database [44] in a convenient format for map generation, and coded with geographical information. in our experiment, we wanted to determine the performance of the system in a close to reality configuration of the system. the results of the analysis described in section 2 above identified the requirements for the tools to be constructed for distributed syndromic or disease surveillance. we have divided the requirements which distributed surveillance and epidemiology systems should fulfil, are shown below. while hospitals can afford a real-time connection to a centralised system such as rods [12] , this solution is not very suitable for all gp clinics. for some gp clinics a more suitable solution may be connection to a shared interconnecting network on a regular basis for reading emails, receiving electronic lab results and electronic hospital discharge letters. this has been the most common solution for gp clinics connected to the north norwegian regional health network. however, the gp clinics are now migrating to an alternative solution, in which they are constantly connected to the health network. to be able to cover all connection alternatives, a distributed solution for epidemiology and healthcare surveillance must support periodic connection to a shared interconnecting network, minimise communication costs, and tolerate narrow bandwidth. coverage is perhaps the most important aspect of a distributed epidemiology and surveillance system. the best approach to achieving coverage is through developing a free and opensource solution that has a minimum of hardware and software costs. the rods [12] system and b-safer [2] system, and several other tools used in outbreak detection, use this approach. the heterogeneity of ehr solutions for general practices is problematic because of the range of different ways in which they store medical data. existing systems may have problems converting information stored in the internal system to a standardised format [9, [47] [48] [49] . they may also lack important information items such as the onset date of a communicable disease. a distributed epidemiology and disease surveillance system would benefit from using a standardised data format internally. a standardised data format will make querying of data easier. a syndromic or disease surveillance system must be automatic (independent of manual data entry for surveillance purposes) to be sustainable [22, 37] . the primary task of healthcare workers is to treat patients, not to perform computer system configuration or system maintenance. however, many institutions in primary care are small, and do not normally employ professional system administrators. in some areas, there is a shortage of available staff with the skills for system administration. a system should therefore ideally require zero configuration and maintenance. this requirement may be hard to meet because most systems need at least a minimum of configuration and maintenance to operate smoothly. this can be difficult to achieve because of the need for manually specifying and ensuring a security policy (see section 4.7 below). a system intended to enable the computation of global epidemiological queries must be asynchronous. it is difficult to anticipate how a globally distributed computation can be performed while a user is waiting for a system response. if we allow participating servers to connect periodically to a shared network, we must ensure that the computations performed on the servers are independent of the computation requester being available and online. asynchronous computations are also more autonomous. a surveillance system must be dynamic in terms of when and how often calculations are performed. the need for reporting frequency may span from running a query every week to several times a day. if an outbreak is detected, the frequency of calculating the diffusion of the disease may need to be increased. the responsibility for switching a gp clinic system into online mode (constantly connected to the shared interconnecting network) may need to be external to the gp clinic. this decision should only be made after a local authentication and authorisation of the request. to be able to meet our objective of not revealing patient identification, and of not being dependent on centralised storage of information about individual patients, the system needs to be distributed. classification of cases into disease or syndrome groups must therefore be done locally, within each ehr sys-tem. many of the syndromic surveillance systems in the usa utilise the 'chief complaints' free text field, and icd-9 or icd-10 codes in the hl7 messages to perform classification of events into syndrome groups. in these systems, the classification of cases against disease and syndrome groups is done centrally, which makes it easy to define and put into use new disease definitions. in norway and several other european countries, the gps use the icpc classification. icpc is more directly related to symptom recording than either icd-9 and icd-10, which makes it more applicable for syndromic and disease surveillance. however, recent travel to affected areas or contact with infected persons is not part of the classification. nor are explicit values of biometrical measurements such as body temperature >38.4 • c. these features were among the criteria for identifying a possible sars case. the most difficult requirement for existing ehr systems to meet is therefore the capacity to incorporate new syndrome definitions that need to be reported. it is important that these syndrome definitions are distributed and automatically incorporated into the ehr system, because of the limitations which paper-based forms have in an outbreak situation, as identified by foldy [13] . for some ehr systems, a new disease or syndrome definition may require that new software is added to the existing ehr system. an example that illustrates this issue is the use of biometrical measurements such as body temperature, pulse, or blood pressure as part of the disease classification. this problem does not apply to all ehr systems, as we will see below. constructing, distributing and updating ehr system software are all time-consuming activities that preferably should not be necessary at the point when an outbreak of a new and dangerous disease is detected. in the case of sars, we saw that the definition of the disease evolved and became more specific over time. as new knowledge about a disease becomes available, it needs to be disseminated to all systems participating in surveillance. this may require that the ehr system is revised to include the new definition of the disease. experience from milwaukee, wisconsin [13] shows that manual routines for updating screening forms are problematic. the "counting" part of a distributed disease surveillance system also needs to be highly dynamic and to allow for inclusion of new disease definitions "on the fly". to be able to monitor the diffusion of a new disease, we argue that we need systems able to include new definitions of diseases in real time. communication in such systems therefore must be two-way, enabling the flow of information about diseases or syndrome definitions in one direction, and information about occurrences and diffusion of the condition in the other direction. the requirements above seem to present a perfect case for mobile code, also known as mobile software agent solutions [50, 51] . many implementations of mobile software agent solutions already exist. mobile software agent technology makes it possible to build highly dynamic distributed systems that support asynchronous and autonomous computations. new functionality in ehr systems may be defined and disseminated as source code to the target systems. however, are mobile-code-based solutions imaginable in healthcare? the risks of mobile code are that it has the capacity to bring down the ehr system in a gp's office due to programming errors, deliberately or unintentionally dump patient data on the internet, take all computing resources away from the gp, etc. even if the source code is certified, it should be asked whether it is appropriate to trust software providers which assert that they "will not be held responsible", if patient data finds its way onto the internet. yes, indeed, it may be difficult to accept mobile-code-based solutions in healthcare information management. is the mobile-code-based solution the only solution able to meet the requirements? mobile code is not suitable, because the physician responsible for the patient data should be able to know exactly what the software that accesses the patient data do. if not, it is unlikely that a physician will authorise the use of such a system. authorisation is closely connected to the tasks performed. if the physician does not know beforehand what the software does, it will be inappropriate to provide authorisation. instead of moving code, we can get some of the flexibility inherent in mobile-code-based approaches by moving program control and data. propagation of program control allows program control to be moved between hosts, in the same manner as mobile code, by transferring a task specification and data between the hosts. the principle is located somewhere between process migration and remote procedure call [52] . instead of letting a process or software agent wander autonomously at its own will between hosts, we ensure that the software agent is under the control of several authorities at all times. these authorities may terminate the software agent at any time. by applying this principle, we gain control over the code executed, but lose some of the flexibility of mobile-code-based solutions. as described, the snow agent system would not appear to provide a simple solution to the hard requirements outlined above. the system may however be used as a dissemination tool for technology that is able to meet the hard requirements above. we will come back to this issue in section 4.8, below. the current version of the snow agent system is a redesign and a complete reimplementation of the virtual secretary system developed at the department of computer science at the university of tromsø [41] in 1993-1998. the first version of the system was initially a mobile-code-based solution for mobile software agents. however, due to the security issues with mobile code [53] [54] [55] , we have moved away from this approach. the current version of the snow agent system does not support mobile code, but allows program control to propagate between hosts using two daemon processes as the mobility enabling system services. these two services are the agent daemon and the mission controller (mc). their tasks and responsibilities are explained in more detail below. the agent daemon and the mission controller are built as extensions to the jabber extensible open source server version 1.4.2 [56] , as shown in fig. 1 . jabber implements the extensible messaging and presence protocol (xmpp) (ietf rfcs) [57, 58] , which allows users to show their presence, chat, and exchange instant messages (xmpp messages). the xmpp messages are expressed using xml. the agent daemon, the mission controller, snow agents and users of the snow agent system also communicate using xmpp messages. the jabber server communicates with jabber clients using the "c2s" (client-to-server) component. jabber talks to other jabber servers using the "s2s" (server-toserver) component. we define an agent as a software entity that performs a series of operations on one or more host computers on behalf of a user. an agent may be mobile and able to migrate or propagate between host computers. an agent has a mission that it seeks to accomplish. this mission is specified by the user (or "mission requester") and is expressed as a mission specification using xml. we use the term "mission agent" for the incarnation of an agent that runs on a single computer. the mission agent is instantiated by an agent daemon that creates the mission agent processes. an agent (or agent mission) may therefore consist of several mission agents that run in parallel on multiple hosts or serially as a relay. a mission agent may also employ sub-missions by sending a mission specification to a mission controller. the agent daemon has more or less the same responsibility as traditional server technology and middleware such as corba, j2ee etc., and can easily be replaced by such technology. the agent daemon manages local resources such as processor, disk, network, and agent applications. its main responsibility is to instantiate, evict and restore mission agents. it negotiates agent missions with remote mcs and performs authentication of instantiated mission agents. in fig. 1 , above, the topmost agent daemon has instantiated two mission agents. these mission agents may have been requested by agent 1 or by a user using a jabber client. the agent daemon also provides information about running mission agents to system administrators, local and remote agents and users if allowed by the local system administrator and each agent. the mission controller functions as an intermediary between the user client and the agent daemon which instantiate the processes based on the specification provided by the user. by performing mission control on behalf of the user, the user becomes independent of the agent execution and the agent becomes more autonomous and independent of the user. the mc knows the set of applications (or mission types) available for each user based on user identity. the mc performs mission control by receiving mission specifications from end users, negotiating mission agent instantiation with remote agent daemons, keeps track of the whereabouts of the agent and notifies the mission requester when a mission is finished, when mission results are available, or if an error occurred. the mission controller performs the migration service for agents executing on a host by receiving the mission specification from the local agent. the mission controller also provides mission control to agents that want to use sub missions on remote hosts. a mission agent is a process that is instantiated by the agent daemon based on a specification received from a mc. the mission specification is expressed in xml and contains the name and version of the application to run, user identity, certificates and signatures proving the identity and authorisation of the user, public keys for encryption of mission results, time and mobility constraints, list of hosts to visit and data used during the mission. the code used for constructing the mission agent must exist on the target host before the mission agent can be instantiated. the system owner needs to consent to supporting the agent type before the agent type can be instantiated. a snow agent has the capability to communicate with any entity using xmpp messages. it can ask either the mission controller or the agent daemon, or both, for a list of local or remote agents. it can participate in a conference with users or other agents or even obtain presence information about local or remote agents and users. the agents may send mission results directly to the mission requester using ordinary xmpp messages with encrypted or unencrypted contents, or to any other valid recipient. the mission controller is able to control two kinds of agent distribution schemes: spread mode missions and jump mode missions, as shown in fig. 2 . in fig. 2 an agent, labelled the "parent agent", requests a sub-mission to three hosts by sending a mission specification to the local mc (not shown). in fig. 2a , the mission specification specifies a spread mode mission, in fig. 2b it specifies a jump mode mission. fig. 3 shows the phases of the mission agent's life cycle on the hosts in a spread mode mission. in spread mode missions, the mission controller forwards the mission specification to the remote agent daemons, indicated by the arrows labelled "1" in fig. 2a . the mission controller negotiates with several remote agent daemons simultaneously. this is illustrated by the negotiation phase in fig. 3 . the agent daemons may respond with a waiting time for execution. this is illustrated by the waiting phase shown in fig. 3 . after the potential waiting phase, the mission agents are instantiated by the agent daemon and the mission specification is transferred (through the connection to the c2s component). this stage is shown as the activation phase in fig. 3 . when the complete mission specification (including the task specification and data) is transferred, the mission agents become operational and start performing their tasks. this is illustrated by the working phase in fig. 3 . mission agents are potentially instantiated on each host covered by the mission and run in parallel until the mission has ended. when a mission agent has finished its task, the mission result can be transferred to the mission requester as illustrated by the arrows labelled "2" in fig. 2a . in jump mode (fig. 2b) , the mission controller sends the mission specification to the agent daemon on the first target host. this is illustrated by the arrow labelled "1" in fig. 2b . a negotiation phase may follow while the agent daemon and the mission controller try to reach an agreement on the mission specification. this is illustrated by the negotiation phase (t1-t2) in fig. 4 . after the negotiation phase, a waiting, activation and working phase may follow, as in the spread mode case. when execution on that site has finished, the mission agent asks the local mission controller for migration service to a new host. this is illustrated by the migration phase (t5-t6) shown in fig. 4 . the local mission controller negotiates the migration with the agent daemon on a remote site by transferring and negotiating the contents of the mission specification. the main mission controller (on the initiating host in fig. 2b ) regains control when the agent has migrated to the new host. the first migration is illustrated by the arrow labelled "2" in fig. 2b . this scheme continues until all hosts in the agent's itinerary have been visited as illustrated by the arrows labelled "3" and "4" in fig. 2b. the users of the snow agent system may use any jabber client that supports sending raw xml messages or specialised plugins that produces the mission specifications. jabber clients such as exodus [59] and jeti [60] provide a plug-in interface. specialised graphical user interfaces (gui) to the snow agent system applications may be implemented as exodus or jeti plug-ins and provide gui for specifying missions and interacting with the snow agent system. the snow agent system clients communicate with mission controllers, create missions, obtain mission results, terminate, restore and remove agent missions, and show incoming or stored mission results. snow agent system missions are not dependent on the jeti or exodus jabber clients. any program or jabber client (or even telnet) capable of communicating with a jabber server may be used to deploy snow agent missions and receive mission results. in this section we describe how the snow agent system meets the requirements stated in section 3 above. the snow agent system may use several connection or organisation models. peer-to-peer model where the snow agent servers must be constantly connected and available for mission requests. each snow agent server is responsible for one gp clinic or patient clinic. specialised snow agent servers, supporting agent types that coordinate large agent missions, may also be used. communication between the servers is done using the standard "s2s" jabber component. in this model, one snow agent server serves as post office for a number of snow agent servers that poll the post office for messages and processing requests. a post office may serve all gps in a county or a single municipality. if the ehr installation needs to be protected by a firewall that blocks incoming connections from the network into the ehr system, the hierarchical model is needed. the snow agent system uses two additional jabber extensions to enable periodical connections to the centralised snow agent server (jabber + snow agent system extensions). in this organisation, the snow agent servers have two distinct roles, the "poller" and the "post office" roles. the "post office" server provides offline storage and relay of xmpp messages on behalf of the "poller" which is the ehr installation. the snow agent server behind the firewall, the poller, must take the initiative to connect to the post office (outgoing connection). when the connection is established, and authentication is performed, the poller will receive all messages and processing requests stored by the post office server while the poller was offline (see fig. 1 above.) the poller may also choose to stay constantly connected to the post office. these two components make it possible to treat a hierarchically organised network as a peerto-peer network. the hierarchical model enables gp clinics protected by firewalls to periodically receive computing requests (in the form of a mission specification). it also satisfies the norwegian data inspectorate's requirements for how small health institutions and gp clinics can connect to a common health network [61] . these requirements state that it should be impossible to open connections from the network into the ehr system protected by the firewall. the connection frequency of the poller component is configurable. if an outbreak is detected, an external authority may request the component to stay permanently connected. the poller component is also capable of connecting to several post office servers, which enables participation in several organisation models or "networks" simultaneously. the snow agent system is designed to be an automatic epidemiology and surveillance service. as long as healthcare workers continue to enter data into the ehr system, the system will be able to collect statistical data, given that the necessary consent for collecting the data is obtained. contents and collection frequency are however dependent on the configuration of the participating hosts. the system can only collect information about diseases explicitly supported by the gps. the reporting frequency depends on the connection model supported by the gp clinics. if all gp clinics are constantly connected (using the peer-to-peer or hierarchical model), then the system supports a high reporting frequency. if the gp clinics connect periodically, then the gp clinic with the lowest connection frequency decides the maximum reporting frequency if 100% coverage and up-to-date data is necessary. the snow agent system epidemiology service does however cache epidemiological data (mission results) locally and on the "post office" servers, which makes it possible to have a higher reporting frequency if the data "freshness" requirement can be relaxed. the computations performed by the mission agents within the snow agent system operate independently of its requester. the requester may however terminate or query the state of the mission agent at any time. the important issue is that the computation is independent of an online mission requester. a snow agent mission may run for days or weeks without the requester being available and online. a snow agent system server capable of connecting to and extracting data from an ehr system needs to operate under a security policy. the security policy states which agent applications may be allowed to run by which user. snow agent applications necessarily generate network traffic and the associated communication costs. such issues may also influence the policy for hosting snow agent applications. in terms of the hard requirements outlined above, it may be necessary to distribute new disease or syndrome definitions to be able to automatically count occurrences of possible, probable and confirmed cases of a new disease. such definitions may be built using archetypes as used in the openehr approach [62] . archetypes are explicit definitions of the structure and content of clinical data and of the constraints on the data. the archetypes may be built by an epidemiologist using an archetype editor. the archetype (disease or syndrome definition) may be read by a form generator component to build a graphical user interface for the input of data according to the structure, contents and constraints in the archetype. the screen representation may also be built manually, as in the proper system [63] , which is based on openemed [16] . the data produced by the gui may be validated by a validator component that uses the archetype definition as input, to ensure that the data stored by the input system is valid. to count cases of a new disease such as sars, specialised archetypes may be used to define concepts like "possible sars", "probable sars" and "confirmed sars" cases. the counting of cases may be done by evaluating stored data against these archetypes. the specialisation mechanisms in the openehr archetype definition language (adl) [64] allow for revision of a disease definition and may ensure that the definition of a new disease evolves without losing the cases defined by earlier definitions of the disease. the snow agent system and the openehr or the proper approaches fit perfectly together because the snow agent system is able to distribute a disease or syndrome definition on a global scale. the definition, represented by explicit archetypes, can be used to: (1) generate gui for data input, (2) validate user-supplied data, and (3) provide definitions or screening forms of concepts such as "possible disease x", "probable disease x" and "confirmed disease x". these definitions may be used by snow agents monitoring the diffusion of disease x in a geographical area. a first version of the system has been developed at the norwegian centre for telemedicine and at the distributed system technology centre (dstc) at the university of queensland, in brisbane, australia. the major parts of the system will be open source and available for download and use from the project web page at http://www.telemed.no/opensource/snow/ or http://mit.cs.uit.no/snow/. to learn more about the strengths, weaknesses and performance of a system based on propagation of program control, we performed a series of tests and experiments to see how fast a distributed epidemiology and surveillance system can provide useful information to the system's users. this question is interesting because we expect that a distributed system based on propagation of program control would need to have considerably slower response time than a centralised system. the reason for this is that such a system is more complex and may involve low-performance servers using low-bandwidth network connections. we also wanted to test the scalability of the mission controller component. this is necessary because mc is the component that handles most of the communication in the system, mainly because of its role as an intermediary between users, agents and agent daemons. the task for the epidemiology system was to plot the pertussis occurrences on an interactive map implemented in svg [43] . the pertussis occurrences were extracted from a set of simulated ehr databases. the ehr databases were generated from microbiology laboratory data as described in section 2.1 above. the map polygons were created from bitmaps made available in the norgeshelsa system [42] and converted to svg polygons. then the map polygons were inserted into a mysql database [44] in a convenient format for map generation and coded with geographical information. the interactive map was based on the vienna example provided by carto.net [65] . in our experiments, we wanted to determine the performance of the system in a close to reality configuration of the system. in norway, this means that we needed to use the hierarchical connection model with one post office server and several gp installations polling for processing requests. we used two mission agents, the "main-epidemio" mission agent, which ran on the post office server, collecting and merging epidemiological data from the participating hosts, and the "ehr-epidemio" agent, which computed on our gp clinics servers. the ehrepidemio mission agent queried the local ehr database for pertussis cases in the given time period, and reported the result to the main-epidemio agent. in our experimental setup, a computer located in tromsø, norway, acted as post office server for the gp clinic snow agent servers. the poller components in the five gp clinic servers were configured with a permanent connection to the post office. this configuration provides the fastest system response possible. (see section 2 above for description of the hardware and software used on the snow agent servers.) in our experiments we first tested the computation time for each participating host. in this case the hosts acted as both post offices and gp clinic servers. the databases and mission specification were identical for all hosts. we measured the amount of time used from receiving the mission or submission request until the mission result notification message was sent by the mission controller. our results are shown in table 2 . the table shows the average processing time from 10 execution runs on each target. in fig. 5 we have plotted the processing time for an identical mission specification on each of our target hosts (shown as points on the line), which shows that the processing time is dependent on processor speed. the time difference between the ehr agent and the main agent (line labelled diff) consists of processing time used to prepare the sub-mission specification and process sub-mission result (from ehr-epidemio agents) and producing the final svg file (113 kb in this case). in our next experiment, we wanted to see how the cost of administrating several target hosts in the sub-mission affected the total mission duration. to test this, we deployed the main-epidemio agent on our post office server. we deployed the ehr-epidemio mission to an increasing number of target hosts, starting with the slowest host first. again, we averaged the processing time over ten consecutive runs. we used the same database and mission specification as in experiment 1. this should make the computation needs for the sub fig. 5 -processing time as a function of processor speed. missions similar to experiment 1. however, the administration cost of the sub mission has increased because the mission specification was transferred over the internet to an increasing number of participating hosts. this is likely to cause an increased delay. in table 3 , we see that the increase of total processing time, compared to that in table 2 , stems from an increase in the ehr mission processing time measured by the mission controller on host 1. we believe the variations in processing time result from varying network conditions, as the participating hosts were spread around the globe. there is an increase in processing needs required by the mission controller to administrate the sub mission, but these processing costs are probably less than the variation in network conditions we experienced. some evidence for this may be found in the standard deviation columns for the processing time. the good news from the second experiment is that the total processing time seems to be minimally affected by increasing the geographical area covered by the epidemiological query. this result is according to our expectations. however, we have only tested the system on a minimal number of hosts. the processing time reported in table 3 shows a wait of about 25 s for getting updated information directly from the ehr systems. a 25-s wait may be a bit too long for busy gps. with the expectation of slow response times in mind, we designed the epidemiology service with a cache. use of the cache reduces the processing time to an average of 1.53 s (over 10 execution runs with s.d. 0.57) when interrogating the server in tromsø, norway from brisbane, australia. by relaxing the requirement for freshness of data, and scheduling periodic epidemiology missions to slow hosts, it seems to be possible to achieve an acceptable response time from the epidemiology service even if servers with 1997 technology (server 6) are used. in fig. 6 we see that the processing times involving six hosts are shorter than those involving five hosts. this effect is explained by the varying bandwidth conditions on the internet during the experiments, and represents a worst-case scenario as intended in the design of the experiment. in fig. 7 we see the uml sequence diagram for an agent mission involving one target ehr host. the diagram shows that the mission controller is the component, which handles most messages. the scalability of the system is therefore partly dependent on the ability of this component to scale when the number of target hosts in a spread mode mission grows. in the second experiment the "service provider host" and the "post office" was the same host, which means that a single mission controller handled both agent missions shown in the sequence diagram. to establish a rough estimate of the scalability of the snow agent server and the mission controller we computed the processing requirements and then performed two tests: (1) message throughput test for the snow agent server and, (2) mission controller and the snow agent server's ability to send outgoing messages. in norway, there are approximately 1800 gp clinics [66] . the distribution of gp clinics by county is shown in table 4 . a hypothetical deployment of the snow agent system could be configured with one snow agent post office server in each county, to which all gp clinics in that county connect. in this configuration, the snow agent post office servers need to be able to support 180-200 ehr system connections simultaneously. the mission specification used in the above experiment is about 1 k of data for the ehr epidemiology mission. for each additional host participating in the mission, the mission specification grows by about 100 b. the output requirement for the mission controller therefore has an n squared growth, or o n 2 in big-oh notation. the output requirement is shown in fig. 8 . the input requirement, shown in fig. 9 , is linear. to achieve this we had to modify the data format returned by the ehr agent running on the ehr hosts, about 2 kb for 15 postcodes for a period of 34 days. the system internal messages shown in fig. 7 , the "ok", "alive" and "finished" status messages require 179, 179 and 180 b, respectively. first, we tested the message throughput of the snow agent server. in this test we used server 2 (a pentium 4, 2.4 ghz processor) and 3 (a pentium iii 800 mhz processor) connected through a 100 mbit/s network. we measured the time taken to transmit messages (4 kb) from a client program on server 3 over the network to server 2 and back to a second client on server 3. because server 3 had a slower processor than server 2, we only observed a 30% load of the snow agent server running on server 2. the throughput in the test stabilised at 0.8 mb/s or 197 messages per second. this throughput may actually reflect the maximum performance of server 3, which both sent and received the messages. in the second test we measured how fast the mission controller and the jabber server were able to output messages. the mission controller on server 2 sent 1000 messages to a client on server 3. we measured the time from when the mission controller sent the message, until the client received the message (using synchronized clocks). because of limitations in the test software, the maximum message size was 4 kb. the test showed that the mission controller can send approximately 400 messages per second. however, a 4 kb mission specification will cover 31 hosts. to cover 200 hosts, the corresponding mission specification will be 20 k, and the total output would need to be 4 mb. in the test, the first 200 messages were sent by server 2 and received on server 3 after 0.6 s. four megabytes was transmitted and received after 1.6 s (output rate 2.5 mb/s). how fast a real system would be able to distribute the mission specifications depends on many parameters, which makes it difficult to predict a realistic estimate for a deployed system. however, our results show that it is realistic to support 200 ehr host per post office server. the theoretical bandwidth of a server connected to a 100 mbit/s network is an output of 12.5 mb/s, which corresponds to a mission with approximately 350 participating hosts. with utilisation of about 70% of the network adaptor, the server can support 295 hosts (8.6 mb mission controller output) in 1 s. for the scalability of the input, it would be beneficial to distribute the activation of the sub mission agents over time, because this will distribute the load on the network and the receiving server over time. the input is also dependent on how many cases of infected persons are identified in a single query. in a situation with many infected persons, the size of the mission result will increase towards the maximum of all postcodes covered by the gp clinic. the realistic capacity requirement of the system can therefore be estimated by counting the number of postcodes covered by the gp clinic and calculating the maximum throughput required for the system. the goal of syndromic and disease surveillance systems is to detect disease outbreaks as early as possible. however, syndromic surveillance systems has not yet proved its capacity to discover outbreaks earlier than traditional surveillance methods [24, 29, 38] . also, the reliance on the ability of "astute clinicians" to detect disease outbreaks in time has been questioned [32] . in the future, syndromic surveillance may be based on data provided by human sensors, as described bẙ arsand et al. [67] . the use of human sensors may enable earlier detection of disease outbreaks than what is possible today. a study of healthcare seeking behaviour in new york showed that in that healthcare context, only 8.8% of patients with ili symptoms attend an emergency department. about 50% of the patients consulted a physician (gp) [6] . because most patients in norway see a gp as their first point of contact with the healthcare system, we should provide the general practitioners with the necessary tools to discover local disease outbreaks. a syndromic surveillance system should therefore include data from the ehr system used by doctors to increase the coverage of the system. in norway this implies using ehr data from gps for surveillance purpose. this is feasible, because norway has close to 100% coverage of ehr systems among gps. most gp surgeries are now becoming connected to the national health network. the current disease surveillance system in norway (msis) [40] is based on reporting both from laboratories and from gps. the gps report the cases on pre-printed paper-based forms provided by the laboratory that performed the diseaseconfirming lab test. in addition to the mandatory reporting undertaken by all gps, approximately 200 gp clinics report weekly, during the winter, about the occurrence of a selection of diseases (including ili). achieving good coverage and high accuracy has been a problem in norway as in other countries [39, [68] [69] [70] [71] and is also limited by the question of cost. previous attempts to report the number of cases of selected diseases on pre-printed postcards proved too expensive to be sustainable. the norwegian institute of public health produces a weekly report (the msis report) containing the number of occurrences at a national and county level. the reporting frequency has been too slow and the geographic level too coarse-grained to be of any practical use in triggering general practitioners' awareness of disease outbreaks. however, the norwegian institute of public health is changing its reporting system to a web-based interface that supports a higher level of geographical granularity. the inclusion of data from a large number of ehr systems used by gps raises the need to rethink what technology might be appropriate for this purpose. current middleware approaches seem to be focused on the idea of a centralised, highly available and powerful server (or pool of servers), which utilises the request reply call semantics for distributing the computations between clients and servers or tiers. in contrast, our need is to have distributed autonomous computations on many ehr system servers, with potentially periodic connectivity and narrow bandwidth that cannot support the request reply semantics for security reasons. from our point of view, a peer-to-peer system among gps within the local communities seems to be the appropriate model on which to base a tool for discovering local disease outbreaks. such a system could also avoid the privacy issues associated with the use of ehr data. however, building a distributed and decentralised system for syndromic or disease surveillance raise new challenges not present in centralised systems such as rods [12] . distributed and decentralised surveillance systems need to address scalability, system maintenance and the need for two-way communication, as outlined in the requirements section above. we have described the requirements for such a system, and found that an ideal surveillance system is one that: (1) enables epidemiologists to define the disease to be monitored, including codes, symptom lists and constraints on biometrical measurements, (2) enables distribution of a disease definition to all systems participating in disease surveillance, (3) enables clinicians to start reporting cases in their normal working tool, the ehr system, immediately after receiving the disease definition, (4) immediately enables collection of statistical data after dissemination of a disease definition, and, (5) supports revision of the disease definition as more knowledge becomes available about the disease. this procedure needs to be achievable in a minimum of time. the human factor should be the main time consumer. the surveillance system needs to support flexible connection methods, enabling automatic, independent and dynamic reporting mechanisms. it must support asynchronous (and autonomous) computations and must provide adequate security mechanisms. syndromic and disease surveillance systems traditionally report events. an event is typically a positive laboratory test that confirms a diagnosis. in our system, it seems more natural to report the number of cases affected by the disease or syndrome classification used. this is natural because the ehr system is constantly updated with new electronic lab reports. it also provides the opportunity to review the development of cases retrospectively, because historical data is also available in the ehr system. the real-time outbreak and disease surveillance system (rods) [12] uses a centralised approach to perform syndromic surveillance. an interesting feature of rods is the use of a health system resident component (hsrc) that runs within the health institutions' firewall. this component has much more data available because it can operate on patient-identifiable clinical data. hsrc is able to link and use laboratory and radiology data in its inference and is able to achieve much higher specificity of patient categorization through access to more information. these benefits also apply to our system. the sars outbreak showed that a disease outbreak can easily cross borders. this represents a problem when an outbreak occurs on the border between two disease surveillance systems. the use of a common disease surveillance system can help to overcome such problems. the decentralised approach used in the snow agent system has the potential to solve the problem of cross-border disease surveillance because it removes the need to export patient related information to a centralised storage for processing. from our survey of available literature, we have not discovered any systems that are able to update the coding system or symptom list as a feature of the disease surveillance system. however, the centralised inference and detection system could easily be updated to use new diagnosis codes or syndrome definitions. in distributed and decentralised surveillance systems, this issue can be solved by enabling two-way communication, passing disease or syndrome definitions in one direction and data about the distribution of cases in the other direction. while rods is a monitoring system that feeds data from many sources into a centralised inference system, the snow agent system is more a "peer-to-peer" approach aimed at helping clinicians to discover ongoing outbreaks in the local community and aiding the diagnosis process on a daily basis. we believe such use of the system will have a positive effect on data quality in primary care. the snow agent system approach has the advantage that it is capable of accessing high quality data that precedes disease outbreak detection in a laboratory by many hours. the time advantage may sometimes be as much as a day, depending on geography and sample delivery routines. if the system counts the occurrences of prodromes or even just the rate of lab test requests, it can raise the awareness of local gps about an outbreak long before the laboratory sees enough cases to issue an alarm. the snow agent system also has the advantage that it can count cases that are diagnosed using epidemiological techniques, for instance in situations where a whole family is affected. such cases may never be confirmed by sending a sample to a laboratory. these two features are unique advantages using the peer-to-peer approach between gps. our performance test shows that the system will scale to national level in norway. the test shows that a normal desktop computer running the linux operating system can distribute a computing request to all gp clinics within any norwegian county within a reasonable time. the results from the epidemiology processes running on all ehr system installations can be merged, and may be depicted on an interactive map by an end user anywhere, using a normal web browser with a svg viewer component installed. if caching of computing results is performed, the response time may be reduced dramatically. by scheduling epidemiological queries at regular intervals, it is possible to achieve an acceptable level of responsiveness by the system. the asynchronous nature of the system, together with the autonomy of the mobile agents, makes it possible to compensate for the use of old servers with narrow bandwidth and periodic connectivity. we believe that the system can scale further, by adding more levels to the hierarchical organisation implemented by the system. the mission controller component is the critical component regarding the scalability of the system. the jabber server supports several server configurations that enable the server to handle many server connections. this feature makes it possible to scale up the processing capacity of a snow agent server dramatically. the snow agent system's epidemiology service is not yet sufficiently developed to qualify as an ideal surveillance system. however, it has some benefits over existing approaches. most important, it avoids the privacy problems associated with using ehr data. the two-way communication supported by the snow agent system's epidemiology service makes it possible to update the computer systems that feed the disease surveillance system with new kinds of data. the system is designed for everyday use by general practitioners who need epidemiological information in their daily work, which is beneficial for surveillance systems [3] . the system may also be used as a feeder system to a centralised outbreak detection system such as rods. in this paper we have presented the requirements that we propose distributed surveillance and epidemiology systems must meet when facing a local, national or a pandemic outbreak of a communicable disease. the surveillance systems should make use of the ehr systems used by gp clinics, because most patients consult their gp as the first point of contact with the health service when they get ill. from our point of view, such systems should be distributed and decentralised to provide doctors with an appropriate tool to discover local disease outbreaks, and to avoid the associated privacy issues. such a system must support two-way communication, to enable the participating ehr systems to incorporate new disease or syndrome definitions. the snow agent system makes it possible to collect epidemiological data directly from the ehr system in a geographical area based on a task specification. the core principle utilised to achieve this is propagation of program control, which lets computing processes or programs migrate between hosts in the same manner as a mobile-code-based system. in our approach, we do not move code, because of the security and authorisation problems inherent in mobile-code-based systems. instead, we propagate program control by moving a task specification and data between the hosts. we have tested our approach in realistic settings to learn about the efficiency and scalability of a system based on the propagation of program control principle. in our test, we placed servers around the globe. the results of our test show little or no performance penalty associated with covering a larger geographical area in an epidemiology query. the response time of the slowest server limits the response time of the epidemiological service. by applying a cache, we reduced the response time of the system from about 25-1.5 s, to produce an interactive map containing the epidemiological data. our scalability tests indicate that the system can easily scale to national level in norway. we would like to thank gudleif arnosen at the microbiology laboratory at the university hospital of north norway, for providing lab data for our system; ivar ness for donating time, bandwidth and processing cycles on his server in spain to the project; andreas neumann for letting us use the javascript source code from the vienna svg map example; and dstc, in particular dr. hoylen sue and the titanium project members, for all their help and technological insight which they have so willingly shared. we are also very grateful for the help from hilde pettersen and eva henriksen with proofreading of this manuscript. what was known before the study: the utility of syndromic surveillance systems, compared to traditional surveillance systems, has so far not been demonstrated [24, 29, 38] . syndromic surveillance systems utilise many different information sources and data from emergency departments is an important one. a telephone survey from new york showed that about 50% of patients with influenza like illness consult a physician, about 9% visit an emergency department and about 4% called a nurse or a health hotline [6] . however, very few of these systems utilise data from general practitioners, gpsurv from auckland, new zealand is one exception [72] . syndromic and disease surveillance systems should complement, rather than replace the "astute clinician" [26, 30] . protection of patient privacy is an important issue to address for disease surveillance systems [7] [8] [9] [10] [11] . what this study contributes: the propagation of program control principle provides a platform for mobile autonomous agent solutions that can be applied in the healthcare environment. this platform have been used to build an epidemiology service that may be expanded to include a distributed syndromic and/or disease surveillance system. distributed syndromic and/or disease surveillance systems can be implemented without being dependent of transferring patient identifiable data which eliminates the privacy issue. a surveillance system built using the propagation of program control principle can easily scale to national level in norway and is able to provide up to date data from the participating source systems within 25 s, even if server technology from 1997 is used. distributed syndromic and/or disease surveillance systems should provide two-way communication to enable distribution and update of syndrome and disease definitions for use by the ehr systems contributing data. investigating public health emergency response information system initiatives in china setting standards for improved syndromic surveillance linking better surveillance to better outcomes value of icd-9 coded chief complaints for detection of epidemics roundtable on bioterrorism detection: information system-based surveillance how many illnesses does one emergency department visit represent? using a population-based telephone survey to estimate the syndromic multiplier review of the 2003 national syndromic surveillance conference-lessons learned and questions to be answered health information privacy and syndromic surveillance systems information system architectures for syndromic surveillance biosense-a national initiative for early detection and quantification of public health emergencies removing a barrier to computer-based outbreak and disease surveillance-the rods open source project technical description of rods: a real-time public health surveillance system sars surveillance project-internet-enabled multiregion surveillance for rapidly emerging disease building dual-or multiple-use infrastructures is the task at hand for state and local health departments results from the fielding of the bio-surveillance analysis, feedback, evaluation and response (b-safer) system in albuquerque an international collaboratory based on virtual patient records on site: past and future emergency response information systems prerequisites for developing and deploying a system for ubiquitous access to patient information what is syndromic surveillance? national bioterrorism syndromic surveillance demonstration program if syndromic surveillance is the answer, what is the question? knowledge-based bioterrorism surveillance on the threshold information technology in the age of emergency public health response. the framework for an integrated disease surveillance system for rapid detection, tracking, and managing of public health threats the economic impact of a bioterrorist attack: are prevention and postattack intervention programs justifiable? emerg syndromic surveillance and bioterrorism-related epidemics disease outbreak detection system using syndromic data in the greater washington dc area using automated medical records for rapid identification of illness syndromes (syndromic surveillance): the example of lower respiratory infection syndromic surveillance syndromic surveillance systems hospital admissions syndromic surveillance-connecticut syndromic surveillance at hospital emergency departments-southeastern virginia the informatics response in disaster, terrorism, and war three years of emergency department gastrointestinal syndromic surveillance in new york city: what have we found? a system for electronic disease reporting and management. determining the extent/spread of problems and minimizing consequences through rapid reporting and dissemination of critical information the norwegian institute of public health. surveillance of communicable diseases in norway the virtual library secretary-a user model based software agent the norwegian institute of public health scalable vector graphics (svg) specification w3c recommendation bogomips mini-howto importing clinical data into electronic health records: lessons learnt from the first australian gehr trials seda: sentrale data fra allmennlegetjenesten -teknisk dokumentasjon sentrale data fra allmennlegetjenesten. sluttrapport fra pilotprosjektet an introduction to the tacoma distributed system version 1.0 mobile agents implementing remote procedure calls security in the ajanta mobile agent system mole concepts of a mobile agent system mobile agents: are they a good idea? jabber software foundation. jabber technology overview exodus -escape from proprietary im systen the data inspectorate. veiledning i informasjonssikkerhet for kommuner og fylker ehr design principles archetypes: the proper way archetype definition language carto:net using blood glucose data as an indicator for epidemic disease outbreaks timeliness of case reporting in the swedish statutory surveillance of communicable diseases notification of infectious diseases by general practitioners: a quantitative and qualitative study mandatory notification of communicable diseases: what physicians think reasons for under-reporting of notifiable conditions evaluation of an electronic general-practitioner-based syndromic surveillance system-auckland this work has been funded by the norwegian research council, grant no. 142162/320 and by the norwegian centre for telemedicine. many individuals have contributed to its realisation, and their help is very much appreciated. in particular, key: cord-018947-d4im0p9e authors: helbing, dirk title: challenges in economics date: 2012-02-10 journal: social self-organization doi: 10.1007/978-3-642-24004-1_16 sha: doc_id: 18947 cord_uid: d4im0p9e in the same way as the hilbert program was a response to the foundational crisis of mathematics [1], this article tries to formulate a research program for the socio-economic sciences. the aim of this contribution is to stimulate research in order to close serious knowledge gaps in mainstream economics that the recent financial and economic crisis has revealed. by identifying weak points of conventional approaches in economics, we identify the scientific problems which need to be addressed. we expect that solving these questions will bring scientists in a position to give better decision support and policy advice. we also indicate, what kinds of insights can be contributed by scientists from other research fields such as physics, biology, computer and social science. in order to make a quick progress and gain a systemic understanding of the whole interconnected socio-economic-environmental system, using the data, information and computer systems available today and in the near future, we suggest a multi-disciplinary collaboration as most promising research approach. static where the world was dynamic, it assumed competitive markets where few existed, it assumed rationality when we knew full well that economic agents were not rational . . . economics had no way of dealing with changing tastes and technology . . . econometrics was equally plagued with intractable problems: economic observations are never randomly drawn and seldom independent, the number of excluded variables is always unmanageably large, the degrees of freedom unacceptably small, the stability of significance tests seldom unequivocably established, the errors in measurement too large to yield meaningful results . . . " [5] . in the following, we will try to identify the scientific challenges that must be addressed to come up with better theories in the near future. this comprises practical challenges, i.e. the real-life problems that must be faced (see sect. 16 .2), and fundamental challenges, i.e. the methodological advances that are required to solve these problems (see sect. 16.3) . after this, we will discuss, which contribution can be made by related scientific disciplines such as econophysics and the social sciences. the intention of this contribution is constructive. it tries to stimulate a fruitful scientific exchange, in order to find the best way out of the crisis. according to our perception, the economic challenges we are currently facing can only be mastered by large-scale, multi-disciplinary efforts and by innovative approaches [6] . we fully recognize the large variety of non-mainstream approaches that has been developed by "heterodox economists". however, the research traditions in economics seem to be so powerful that these are not paid much attention to. besides, there is no agreement on which of the alternative modeling approaches would be the most promising ones, i.e. the heterogeneity of alternatives is one of the problems, which slows down their success. this situation clearly implies institutional challenges as well, but these go beyond the scope of this contribution and will therefore be addressed in the future. since decades, if not since hundreds of years, the world is facing a number of recurrent socio-economic problems, which are obviously hard to solve. before addressing related fundamental scientific challenges in economics, we will therefore point out practical challenges one needs to pay attention to. this basically requires to classify the multitude of problems into packages of interrelated problems. probably, such classification attempts are subjective to a certain extent. at least, the list presented below differs from the one elaborated by lomborg et al. [7] , who identified the following top ten problems: air pollution, security/conflict, disease control, education, climate change, hunger/malnutrition, water sanitation, barriers to migration and trade, transnational terrorism and, finally, women and development. the following (non-ranked) list, in contrast, is more focused on socio-economic factors rather than resource and engineering issues, and it is more oriented at the roots of problems rather than their symptoms: 1. demographic change of the population structure (change of birth rate, migration, integration. . . ) 2. financial and economic (in)stability (government debts, taxation, and inflation/ deflation; sustainability of social benefit systems; consumption and investment behavior. . . ) 3. social, economic and political participation and inclusion (of people of different gender, age, health, education, income, religion, culture, language, preferences; reduction of unemployment. . . ) 4. balance of power in a multi-polar world (between different countries and economic centers; also between individual and collective rights, political and company power; avoidance of monopolies; formation of coalitions; protection of pluralism, individual freedoms, minorities. . . ) 5. collective social behavior and opinion dynamics (abrupt changes in consumer behavior; social contagion, extremism, hooliganism, changing values; breakdown of cooperation, trust, compliance, solidarity. . . ) 6. security and peace (organized crime, terrorism, social unrest, independence movements, conflict, war. . . ) 7. institutional design (intellectual property rights; over-regulation; corruption; balance between global and local, central and decentral control. . . ) 8. sustainable use of resources and environment (consumption habits, travel behavior, sustainable and efficient use of energy and other resources, participation in recycling efforts, environmental protection. . . ) 9. information management (cyber risks, misuse of sensitive data, espionage, violation of privacy; data deluge, spam; education and inheritance of culture. . . ) 10 . public health (food safety; spreading of epidemics [flu, sars, h1n1, hiv], obesity, smoking, or unhealthy diets. . . ) some of these challenges are interdependent. in the following, we will try to identify the fundamental theoretical challenges that need to be addressed in order to understand the above practical problems and to draw conclusions regarding possible solutions. the most difficult part of scientific research is often not to find the right answer. the problem is to ask the right questions. in this context it can be a problem that people are trained to think in certain ways. it is not easy to leave these ways and see the problem from a new angle, thereby revealing a previously unnoticed solution. three factors contribute to this: 1. we may overlook the relevant facts because we have not learned to see them, i.e. we do not pay attention to them. the issue is known from internalized norms, which prevent people from considering possible alternatives. 2. we know the stylized facts, but may not have the right tools at hand to interpret them. it is often difficult to make sense of patterns detected in data. turning data into knowledge is quite challenging. 3. we know the stylized facts and can interpret them, but may not take them seriously enough, as we underestimate their implications. this may result from misjudgements or from herding effects, i.e. from a tendency to follow traditions and majority opinions. in fact, most of the issues discussed below have been pointed out before, but it seems that this did not have an effect on mainstream economics so far or on what decision-makers know about economics. this is probably because mainstream theory has become a norm [8] , and alternative approaches are sanctioned as norm-deviant behavior [9, 10] . as we will try to explain, the following fundamental issues are not just a matter of approximations (which often lead to the right understanding, but wrong numbers). rather they concern fundamental errors in the sense that certain conclusions following from them are seriously misleading. as the recent financial crisis has demonstrated, such errors can be very costly. however, it is not trivial to see what dramatic consequences factors such as dynamics, spatial interactions, randomness, non-linearity, network effects, differentiation and heterogeneity, irreversibility or irrationality can have. despite of criticisms by several nobel prize winners such as reinhard selten (1994), joseph stiglitz and george akerlof (2001) , or daniel kahneman (2002) , the paradigm of the homo economicus, i.e. of the "perfect egoist", is still the dominating approach in economics. it assumes that people would have quasi-infinite memory and processing capacities and would determine the best one among all possible alternative behaviors by strategic thinking (systematic utility optimization), and would implement it into practice without mistakes. the nobel prize winner of 1976, milton friedman, supported the hypothesis of homo economicus by the following argument: "irrational agents will lose money and will be driven out of the market by rational agents" [11] . more recently, robert e. lucas jr., the nobel prize winner of 1995, used the rationality hypothesis to narrow down the class of empirically relevant equilibria [12] . the rational agent hypothesis is very charming, as its implications are clear and it is possible to derive beautiful and powerful economic theorems and theories from it. the best way to illustrate homo economicus is maybe a company that is run by using optimization methods from operation research, applying supercomputers. another example are professional chess players, who are trying to anticipate the possible future moves of their opponents. obviously, in both examples, the future course of actions can not be fully predicted, even if there are no random effects and mistakes. it is, therefore, no wonder that people have repeatedly expressed doubts regarding the realism of the rational agent approach [13, 14] . bertrand russell, for example, claimed: "most people would rather die than think". while this seems to be a rather extreme opinion, the following scientific arguments must be taken seriously: 1. human cognitive capacities are bounded [16, 17] . already phone calls or conversations can reduce people's attention to events in the environment a lot. also, the abilities to memorize facts and to perform complicated logical analyses are clearly limited. 2. in case of np-hard optimization problems, even supercomputers are facing limits, i.e. optimization jobs cannot be performed in real-time anymore. therefore, approximations or simplifications such as the application of heuristics may be necessary. in fact, psychologists have identified a number of heuristics, which people use when making decisions [18] . 3. people perform strategic thinking mainly in important new situations. in normal, everyday situation, however, they seem to pursue a satisficing rather than optimizing strategy [17] . meeting a certain aspiration level rather than finding the optimal strategy can save time and energy spent on problem solving. in many situation, people even seem to perform routine choices [14] , for example, when evading other pedestrians in counterflows. 4. there is a long list of cognitive biases which question rational behavior [19] . for example, individuals are favorable of taking small risks (which are preceived as "chances", as the participation in lotteries shows), but they avoid large risks [20] . furthermore, non-exponential temporal discounting may lead to paradoxical behaviors [21] and requires one to rethink, how future expectations must be modeled. 5. most individuals have a tendency towards other-regarding behavior and fairness [22, 23] . for example, the dictator game [24] and other experiments [25] show that people tend to share, even if there is no reason for this. leaving a tip for the waiter in a restaurant that people visit only once is a typical example (particularly in countries where tipping is not common) [26] . such behavior has often been interpreted as sign of social norms. while social norms can certainly change the payoff structure, it has been found that the overall payoffs resulting from them do not need to create a user or system optimum [27] [28] [29] . this suggests that behavioral choices may be irrational in the sense of non-optimal. a typical example is the existence of unfavorable norms, which are supported by people although nobody likes them [30] . 6. certain optimization problems can have an infinite number of local optima or nash equilibria, which makes it impossible to decide what is the best strategy [31] . 7. convergence towards the optimal solution may require such a huge amount of time that the folk theorem becomes useless. this can make it practically impossible to play the best response strategy [32] . 8. the optimal strategy may be deterministically chaotic, i.e. sensitive to arbitrarily small details of the initial condition, which makes the dynamic solution unpredictable on the long run ("butterfly effect") [33, 34] . this fundamental limit of predictability also implies a limit of control-two circumstances that are even more true for non-deterministic systems with a certain degree of randomness. in conclusion, although the rational agent paradigm (the paradigm of homo economicus) is theoretically powerful and appealing, there are a number of empirical and theoretical facts, which suggest deficiencies. in fact, most methods used in financial trading (such as technical analysis) are not well compatible with the rational agent approach. even if an optimal solution exists, it may be undecidable for practical reasons or for theoretical ones [35, 36] . this is also relevant for the following challenges, as boundedly rational agents may react inefficently and with delays, which questions the efficient market hypothesis, the equilibrium paradigm, and other fundamental concepts, calling for the consideration of spatial, network, and time-dependencies, heterogeneity and correlations etc. it will be shown that these points can have dramatic implications regarding the predictions of economic models. the efficient market hypothesis (emh) was first developed by eugene fama [37] in his ph.d. thesis and rapidly spread among leading economists, who used it as an argument to promote laissez-faire policies. the emh states that current prices reflect all publicly available information and (in its stronger formulation) that prices instantly change to reflect new public information. the idea of self-regulating markets goes back to adam smith [38] , who believed that "the free market, while appearing chaotic and unrestrained, is actually guided to produce the right amount and variety of goods by a so-called "invisible hand"." furthermore, "by pursuing his own interest, [the individual] frequently promotes that of the society more effectually than when he intends to promote it" [39] . for this reason, adam smith is often considered to be the father of free market economics. curiously enough, however, he also wrote a book on "the theory of moral sentiments" [40] . "his goal in writing the work was to explain the source of mankind's ability to form moral judgements, in spite of man's natural inclinations towards self-interest. smith proposes a theory of sympathy, in which the act of observing others makes people aware of themselves and the morality of their own behavior . . . [and] seek the approval of the "impartial spectator" as a result of a natural desire to have outside observers sympathize with them" [38] . such a reputation-based concept would be considered today as indirect reciprocity [41] . of course, there are criticisms of the efficient market hypothesis [42] , and the nobel prize winner of 2001, joseph stiglitz, even believes that "there is not invisible hand" [43] . the following list gives a number of empirical and theoretical arguments questioning the efficient market hypothesis: 1. examples of market failures are well-known and can result, for example, in cases of monopolies or oligopolies, if there is not enough liquidity or if information symmetry is not given. 2. while the concept of the "invisible hand" assumes something like an optimal self-organization [44] , it is well-known that this requires certain conditions, such as symmetrical interactions. in general, however, self-organization does not necessarily imply system-optimal solutions. stop-and-go traffic [45] or crowd disasters [46] are two obvious examples for systems, in which individuals competitively try to reach individually optimal outcomes, but where the optimal solution is dynamically unstable. 3. the limited processing capacity of boundedly rational individuals implies potential delays in their responses to sensorial inputs, which can cause such instabilities [47] . for example, a delayed adaptation in production systems may contribute to the occurrence of business cycles [48] . the same applies to the labor market of specially skilled people, which cannot adjust on short time scales. even without delayed reactions, however, the competitive optimization of individuals can lead to suboptimal individual results, as the "tragedy of the commons" in public goods dilemmas demonstrates [49, 50] . 4. bubbles and crashes, or more generally, extreme events in financial markets should not occur, if the efficient market hypothesis was correct (see next subsection). 5. collective social behavior such as "herding effects" as well as deviations of human behavior from what is expected from rational agents can lead to such bubbles and crashes [51] , or can further increase their size through feedback effects [52] . cyclical feedbacks leading to oscillations are also known from the beer game [53] or from business cycles [48] . the efficient market paradigm implies the equilibrium paradigm. this becomes clear, if we split it up into its underlying hypotheses: 1. the market can be in equilibrium, i.e. there exists an equilibrium. 2. there is one and only one equilibrium. 3. the equilibrium is stable, i.e. any deviations from the equilibrium due to "fluctuations" or "perturbations" tend to disappear eventually. 4. the relaxation to the equilibrium occurs at an infinite rate. note that, in order to act like an "invisible hand", the stable equilibrium (nash equilibrium) furthermore needs to be a system optimum, i.e. to maximize the average utility. this is true for coordination games, when interactions are well-mixed and exploration behavior as well as transaction costs can be neglected [54] . however, it is not fulfilled by so-called social dilemmas [49] . let us discuss the evidence for the validity of the above hypotheses one by one: 1. a market is a system of extremely many dynamically coupled variables. theoretically, it is not obvious that such a system would have a stationary solution. for example, the system could behave periodic, quasi-periodic, chaotic, or turbulent [81-83, 85-87, 94] . in all these cases, there would be no convergence to a stationary solution. 2. if a stationary solution exists, it is not clear that there are no further stationary solutions. if many variables are non-linearly coupled, the phenomenon of multistability can easily occur [55] . that is, the solution to which the system converges may not only depend on the model parameters, but also on the initial condition, history, or perturbation size. such facts are known as path-dependencies or hysteresis effects and are usually visualized by so-called phase diagrams [56] . 3. in systems of non-linearly interacting variables, the existence of a stationary solution does not necessarily imply that it is stable, i.e. that the system will converge to this solution. for example, the stationary solution could be a focal point with orbiting solutions (as for the classical lotka-volterra equations [57] ), or it could be unstable and give rise to a limit cycle [58] or a chaotic solution [33] , for example (see also item 1). in fact, experimental results suggest that volatility clusters in financial markets may be a result of over-reactions to deviations from the fundamental value [59] . 4. an infinite relaxation rate is rather unusual, as most decisions and related implemenations take time [15, 60] . the points listed in the beginning of this subsection are also questioned by empirical evidence. in this connection, one may mention the existence of business cycles [48] or unstable orders and deliveries observed in the experimental beer game [53] . moreover, bubbles and crashes have been found in financial market games [61] . today, there seems to be more evidence against than for the equilibrium paradigm. in the past, however, most economists assumed that bubbles and crashes would not exist (and many of them still do). the following quotes are quite typical for this kind of thinking (from [62] ): in 2004, the federal reserve chairman of the u.s., alan greenspan, stated that the rise in house values was "not enough in our judgment to raise major concerns". in july 2005 when asked about the possibility of a housing bubble and the potential for this to lead to a recession in the future, the present u.s. federal reserve chairman ben bernanke (then chairman of the council of economic advisors) said: "it's a pretty unlikely possibility. we've never had a decline in housing prices on a nationwide basis. so, what i think is more likely is that house prices will slow, maybe stabilize, might slow consumption spending a bit. i don't think it's going to drive the economy too far from it's full path though." as late as may 2007 bernanke stated that the federal reserve "do not expect significant spillovers from the subprime market to the rest of the economy". according to the classical interpretation, sudden changes in stock prices result from new information, e.g. from innovations ("technological shocks"). the dynamics in such systems has, for example, been described by the method of comparative statics (i.e. a series of snapshots). here, the system is assumed to be in equilibrium in each moment, but the equilibrium changes adiabatically (i.e. without delay), as the system parameters change (e.g. through new facts). such a treatment of system dynamics, however, has certain deficiencies: 1. the approach cannot explain changes in or of the system, such as phase transitions ("systemic shifts"), when the system is at a critical point ("tipping point"). 2. it does not allow one to understand innovations and other changes as results of an endogeneous system dynamics. 3. it cannot describe effects of delays or instabilities, such as overshooting, self-organization, emergence, systemic breakdowns or extreme events (see sect. 16.3.4). 4. it does not allow one to study effects of different time scales. for example, when there are fast autocatalytic (self-reinfocing) effects and slow inhibitory effects, this may lead to pattern formation phenomena in space and time [63, 64] . the formation of settlements, where people agglomerate in space, may serve as an example [65, 66] . 5. it ignores long-term correlations such as memory effects. 6. it neglects frictional effects, which are often proportional to change ("speed") and occur in most complex systems. without friction, however, it is difficult to understand entropy and other path-dependent effects, in particular irreversibility (i.e. the fact that the system may not be able to get back to the previous state) [67] . for example, the unemployment rate has the property that it does not go back to the previous level in most countries after a business cycle [68] . comparative statics is, of course, not the only method used in economics to describe the dynamics of the system under consideration. as in physics and other fields, one may use a linear approximation around a stationary solution to study the response of the system to fluctuations or perturbations [69] . such a linear stability analysis allows one to study, whether the system will return to the stationary solution (which is the case for a stable [nash] equilibrium) or not (which implies that the system will eventually be driven into a new state or regime). in fact, the great majority of statistical analyses use linear models to fit empirical data (also when they do not involve time-dependencies). it is know, however, that linear models have special features, which are not representative for the rich variety of possible functional dependencies, dynamics, and outcomes. therefore, the neglection of non-linearity has serious consequences: 1. as it was mentioned before, phenomena like multiple equilibria, chaos or turbulence cannot be understood by linear models. the same is true for selforganization phenomena or emergence. additionally, in non-linearly coupled systems, usually "more is different", i.e. the system may change its behavior fundamentally as it grows beyond a certain size. furthermore, the system is often hard to predict and difficult to control (see sect. 16.3.8). 2. linear modeling tends to overlook that a strong coupling of variables, which would show a normally distributed behavior in separation, often leads to fat tail distributions (such as "power laws") [70, 71] . this implies that extreme events are much more frequent than expected according to a gaussian distribution. for example, when additive noise is replaced by multiplicative noise, a number of surprising phenomena may result, including noise-induced transitions [72] or directed random walks ("ratchet effects") [73] . 3. phenomena such as catastrophes [74] or phase transition ("system shifts") [75] cannot be well understood within a linear modeling framework. the same applies to the phenomenon of "self-organized criticality" [79] (where the system drives itself to a critical state, typically with power-law characteristics) or cascading effects, which can result from network interactions (overcritically challenged network nodes or links) [77, 78] . it should be added that the relevance of network effects resulting from the on-going globalization is often underestimated. for example, "the stock market crash of 1987, began with a small drop in prices which triggered an avalanche of sell orders in computerized trading programs, causing a further price decline that triggered more automatic sales." [80] therefore, while linear models have the advantage of being analytically solvable, they are often unrealistic. studying non-linear behavior, in contrast, often requires numerical computational approaches. it is likely that most of today's unsolved economic puzzles cannot be well understood through linear models, no matter how complicated they may be (in terms of the number of variables and parameters) [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] . the following list mentions some areas, where the importance of non-linear interdependencies is most likely underestimated: â�¢ collective opinions, such as trends, fashions, or herding effects. â�¢ the success of new (and old) technologies, products, etc. â�¢ cultural or opinion shifts, e.g. regarding nuclear power, genetically manipulated food, etc. â�¢ the "fitness" or competitiveness of a product, value, quality perceptions, etc. â�¢ the respect for copyrights. â�¢ social capital (trust, cooperation, compliance, solidarity, . . . ). â�¢ booms and recessions, bubbles and crashes. â�¢ bank panics. â�¢ community, cluster, or group formation. â�¢ relationships between different countries, including war (or trade war) and peace. another common simplification in economic modeling is the representative agent approach, which is known in physics as mean field approximation. within this framework, time-dependencies and non-linear dependencies are often considered, but it is assumed that the interaction with other agents (e.g. of one company with all the other companies) can be treated as if this agent would interact with an average agent, the "representative agent". let us illustrate this with the example of the public goods dilemma. here, everyone can decide whether to make an individual contribution to the public good or not. the sum of all contributions is multiplied by a synergy factor, reflecting the benefit of cooperation, and the resulting value is equally shared among all people. the prediction of the representative agent approach is that, due to the selfishness of agents, a "tragedy of the commons" would result [49] . according to this, everybody should free-ride, i.e. nobody should make a contribution to the public good and nobody would gain anything. however, if everybody would contribute, everybody could multiply his or her contribution by the synergy factor. this example is particularly relevant as society is facing a lot of public goods problems and would not work without cooperation. everything from the creation of public infrastructures (streets, theaters, universities, libraries, schools, the world wide web, wikipedia etc.) over the use of environmental resources (water, forests, air, etc.) or of social benefit systems (such as a public health insurance), maybe even the creation and maintainance of a commonly shared language and culture are public goods problems (although the last examples are often viewed as coordination problems). even the process of creating public goods is a public good [95] . while it is a well-known problem that people tend to make unfair contributions to public goods or try to get a bigger share of them, individuals cooperate much more than one would expect according to the representative agent approach. if they would not, society could simply not exist. in economics, one tries to solve the problem by introducing taxes (i.e. another incentive structure) or a "shadow of the future" (i.e. a strategic optimization over infinite time horizons in accordance with the rational agent approach) [96, 97] . both comes down to changing the payoff structure in a way that transforms the public good problem into another one that does not constitute a social dilemma [98] . however, there are other solutions of the problem. when the realm of the mean-field approximation underlying the representative agent approach is left and spatial or network interactions or the heterogeneity among agents are considered, a miracle occurs: cooperation can survive or even thrive through correlations and co-evolutionary effects [99] [100] [101] . a similar result is found for the public goods game with costly punishment. here, the representative agent model predicts that individuals avoid to invest into punishment, so that punishment efforts eventually disappear (and, as a consequence, cooperation as well). however, this "second-order free-rider problem" is naturally resolved and cooperation can spread, if one discards the mean-field approximation and considers the fact that interactions take place in space or social networks [56] . societies can overcome the tragedy of the commons even without transforming the incentive structure through taxes. for example, social norms as well as group dynamical and reputation effects can do so [102] . the representative agent approach implies just the opposite conclusion and cannot well explain the mechanisms on which society is built. it is worth pointing out that the relevance of public goods dilemmas is probably underestimated in economics. partially related to adam smith's belief in an "invisible hand", one often assumes underlying coordination games and that they would automatically create a harmony between an individually and system optimal state in the course of time [54] . however, running a stable financial system and economy is most likely a public goods problem. considering unemployment, recessions always go along with a breakdown of solidarity and cooperation. efficient production clearly requires mutual cooperation (as the counter-example of countries with many strikes illustrates). the failure of the interbank market when banks stop lending to each other, is a good example for the breakdown of both, trust and cooperation. we must be aware that there are many other systems that would not work anymore, if people would lose their trust: electronic banking, e-mail and internet use, facebook, ebusiness and egovernance, for example. money itself would not work without trust, as bank panics and hyperinflation scenarios show. similarly, cheating customers by selling low-quality products or selling products at overrated prices, or by manipulating their choices by advertisements rather than informing them objectively and when they want, may create profits on the short run, but it affects the trust of customers (and their willingness to invest). the failure of the immunization campaign during the swine flu pandemics may serve as an example. furthermore, people would probably spend more money, if the products of competing companies were better compatible with each other. therefore, on the long run, more cooperation among companies and with the customers would pay off and create additional value. besides providing a misleading picture of how cooperation comes about, there are a number of other deficiencies of the representative agent approach, which are listed below: 1. correlations between variables are neglected, which is acceptable only for "well-mixing" systems. according to what is known from critical phenomena in physics, this approximation is valid only, when the interactions take place in high-dimensional spaces or if the system elements are well connected. (however, as the example of the public goods dilemma showed, this case does not necessarily have beneficial consequences. well-mixed interactions could rather cause a breakdown of social or economic institutions, and it is conceivable that this played a role in the recent financial crisis.) 2. percolation phenomena, describing how far an idea, innovation, technology, or (computer) virus spreads through a social or business network, are not well reproduced, as they depend on details of the network structure, not just on the average node degree [103] . 3. the heterogeneity of agents is ignored. for this reason, factors underlying economic exchange, perturbations, or systemic robustness [104] cannot be well described. moreover, as socio-economic differentiation and specialization imply heterogeneity, they cannot be understood as emergent phenomena within a representative agent approach. finally, it is not possible to grasp innovation without the consideration of variability. in fact, according to evolutionary theory, the innovation rate would be zero, if the variability was zero [105] . furthermore, in order to explain innovation in modern societies, schumpeter introduced the concept of the "political entrepreneur" [106] , an extra-ordinarily gifted person capable of creating disruptive change and innovation. such an extraordinary individual can, by definition, not be modeled by a "representative agent". one of the most important drawbacks of the representative agent approach is that it cannot explain the fundamental fact of economic exchange, since it requires one to assume a heterogeneity in resources or production costs, or to consider a variation in the value of goods among individuals. ken arrow, nobel prize winner in 1972, formulated this point as follows [107] : "one of the things that microeconomics teaches you is that individuals are not alike. there is heterogeneity, and probably the most important heterogeneity here is heterogeneity of expectations. if we didn't have heterogeneity, there would be no trade." we close this section by mentioning that economic approaches, which go beyond the representative agent approach, can be found in refs. [108, 109] . another deficiency of economic theory that needs to be mentioned is the lack of a link between micro-and macroeconomics. neoclassical economics implicitly assumes that individuals make their decisions in isolation, using only the information received from static market signals. within this oversimplified framework, macro-aggregates are just projections of some representative agent behavior, instead of the outcome of complex interactions with asymmetric information among a myriad of heterogeneous agents. in principle, it should be understandable how the macroeconomic dynamics results from the microscopic decisions and interactions on the level of producers and consumers [81, 110] (as it was possible in the past to derive micro-macro links for other systems with a complex dynamical behavior such as interactive vehicle traffic [111] ). it should also be comprehensible how the macroscopic level (the aggregate econonomic situation) feeds back on the microscopic level (the behavior of consumers and producers), and to understand the economy as a complex, adaptive, self-organizing system [112, 113] . concepts from evolutionary theory [114] and ecology [115] appear to be particularly promising [116] . this, however, requires a recognition of the importance of heterogeneity for the system (see the the previous subsection). the lack of ecological thinking implies not only that the sensitive network interdependencies between the various agents in an economic system (as well as minority solutions) are not properly valued. it also causes deficiencies in the development and implementation of a sustainable economic approach based on recycling and renewable resources. today, forestry science is probably the best developed scientific discipline concerning sustainability concepts [117] . economic growth to maintain social welfare is a serious misconception. from other scientific disciplines, it is well known that stable pattern formation is also possible for a constant (and potentially sustainable) inflow of energy [69, 118] . one of the great achievements of economics is that it has developed a multitude of methods to use scarce resources efficiently. a conventional approach to this is optimization. in principle, there is nothing wrong about this approach. nevertheless, there are a number of problems with the way it is usually applied: 1. one can only optimize for one goal at a time, while usually, one needs to meet several objectives. this is mostly addressed by weighting the different goals (objectives), by executing a hierarchy of optimization steps (through ranking and prioritization), or by applying a satisficing strategy (requiring a minimum performance for each goal) [119, 120] . however, when different optimization goals are in conflict with each other (such as maximizing the throughput and minimizing the queue length in a production system), a sophisticated timedependent strategy may be needed [121] . high profit? best customer satisfaction? large throughput? competitive advantage? resilience? [122] in fact, the choice of the optimization function is arbitrary to a certain extent and, therefore, the result of optimization may vary largely. goal selection requires strategic decisions, which may involve normative or moral factors (as in politics). in fact, one can often observe that, in the course of time, different goal functions are chosen. moreover, note that the maximization of certain objectives such as resilience or "fitness" depends not only on factors that are under the control of a company. resilience and "fitness" are functions of the whole system, in particularly, they also depend on the competitors and the strategies chosen by them. 3. the best solution may be the combination of two bad solutions and may, therefore, be overlooked. in other words, there are "evolutionary dead ends", so that gradual optimization may not work. (this problem can be partially overcome by the application of evolutionary mechanisms [120] ). 4. in certain systems (such as many transport, logistic, or production systems), optimization tends to drive the system towards instability, since the point of maximum efficiency is often in the neighborhood or even identical with the point of breakdown of performance. such breakdowns in capacity or performance can result from inefficiencies due to dynamic interaction effects. for example, when traffic flow reaches its maximum capacity, sooner or later it breaks down. as a consequence, the road capacity tends to drop during the time period where it is most urgently needed, namely during the rush hour [45, 123] . 5 . optimization often eliminates reduncancies in the system and, thereby, increases the vulnerability to perturbations, i.e. it decreases robustness and resilience. 6. optimization tends to eliminate heterogeneity in the system [80] , while heterogeneity frequently supports adaptability and resilience. 7. optimization is often performed with centralized concepts (e.g. by using supercomputers that process information collected all over the system). such centralized systems are vulnerable to disturbances or failures of the central control unit. they are also sensitive to information overload, wrong selection of control parameters, and delays in adaptive feedback control. in contrast, decentralized control (with a certain degree of autonomy of local control units) may perform better, when the system is complex and composed of many heterogeneous elements, when the optimization problem is np hard, the degree of fluctuations is large, and predictability is restricted to short time periods [77, 124] . under such conditions, decentralized control strategies can perform well by adaptation to the actual local conditions, while being robust to perturbations. urban traffic light control is a good example for this [121, 125] . 8. further on, today's concept of quality control appears to be awkward. it leads to a never-ending contest, requiring people and organizations to fulfil permanently increasing standards. this leads to over-emphasizing measured performance criteria, while non-measured success factors are neglected. the engagement into non-rewarded activities is discouraged, and innovation may be suppressed (e.g. when evaluating scientists by means of their h-index, which requires them to focus on a big research field that generates many citations in a short time). while so-called "beauty contests" are considered to produce the best results, they will eventually absorb more and more resources for this contest, while less and less time remains for the work that is actually to be performed, when the contest is won. besides, a large number of competitors have to waste considerable resources for these contests which, of course, have to be paid by someone. in this way, private and public sectors (from physicians over hospitals, administrations, up to schools and universities) are aching under the evaluationrelated administrative load, while little time remains to perform the work that the corresponding experts have been trained for. it seems naã¯ve to believe that this would not waste resources. rather than making use of individual strengths, which are highly heterogeneous, today's way of evaluating performance enforces a large degree of conformity. there are also some problems with parameter fitting, a method based on optimization as well. in this case, the goal function is typically an error function or a likelihood function. not only are calibration methods often "blindly" applied in practice (by people who are not experts in statistics), which can lead to overfitting (the fitting of meaningless "noise"), to the neglection of collinearities (implying largely variable parameter values), or to inaccurate and problematic parameter determinations (when the data set is insufficient in size, for example, when large portfolios are to be optimized [126] ). as estimates for past data are not necessarily indicative for the future, making predictions with interpolation approaches can be quite problematic (see also sect. 16.3.3 for the challenge of time dependence). moreover, classical calibration methods do not reveal inappropriate model specifications (e.g. linear ones, when non-linear models would be needed, or unsuitable choices of model variables). finally, they do not identify unknown unknowns (i.e. relevant explanatory variables, which have been overlooked in the modeling process). managing economic systems is a particular challenge, not only for the reasons discussed in the previous section. as large economic systems belong to the class of complex systems, they are hard or even impossible to manage with classical control approaches [76, 77] . complex systems are characterized by a large number of system elements (e.g. individuals, companies, countries, . . . ), which have non-linear or network interactions causing mutual dependencies and responses. such systems tend to behave dynamic rather than static and probabilistic rather than deterministic. they usually show a rich, hardly predictable, and sometimes paradoxical system behavior. therefore, they challenge our way of thinking [127] , and their controllability is often overestimated (which is sometimes paraphrased as "illusion of control") [80, 128, 129] . in particular, causes and effects are typically not proportional to each other, which makes it difficult to predict the impact of a control attempt. a complex system may be unresponsive to a control attempt, or the latter may lead to unexpected, large changes in the system behavior (so-called "phase transitions", "regime shifts", or "catastrophes") [75] . the unresponsiveness is known as principle of le chatelier or goodhart's law [130] , according to which a complex system tends to counteract external control attempts. however, regime shifts can occur, when the system gets close to so-called "critical points" (also known as "tipping points"). examples are sudden changes in public opinion (e.g. from pro to anti-war mood, from a smoking tolerance to a public smoking ban, or from buying energy-hungry sport utilities vehicles (suvs) to buying environmentally-friendly cars). particularly in case of network interactions, big changes may have small, no, or unexpected effects. feedback loops, unwanted side effects, and circuli vitiosi are quite typical. delays may cause unstable system behavior (such as bull-whip effects) [53] , and over-critical perturbations can create cascading failures [78] . systemic breakdowns (such as large-scale blackouts, bankruptcy cascades, etc.) are often a result of such domino or avalanche effects [77] , and their probability of occurrence as well as their resulting damage are usually underestimated. further examples are epidemic spreading phenomena or disasters with an impact on the socio-economic system. a more detailed discussion is given in refs. [76, 77] . other factors contributing to the difficulty to manage economic systems are the large heterogeneity of system elements and the considerable level of randomness as well as the possibility of a chaotic or turbulent dynamics (see sect. 16.3.4) . furthermore, the agents in economic systems are responsive to information, which can create self-fulfilling or self-destroying prophecy effects. inflation may be viewed as example of such an effect. interestingly, in some cases one even does not know in advance, which of these effects will occur. it is also not obvious that the control mechanisms are well designed from a cybernetic perspective, i.e. that we have sufficient information about the system and suitable control variables to make control feasible. for example, central banks do not have terribly many options to influence the economic system. among them are performing open-market operations (to control money supply), adjustments in fractional-reserve banking (keeping only a limited deposit, while lending a large part of the assets to others), or adaptations in the discount rate (the interest rate charged to banks for borrowing short-term funds directly from a central bank). nevertheless, the central banks are asked to meet multiple goals such as: â�¢ to guarantee well-functioning and robust financial markets. â�¢ to support economic growth. â�¢ to balance between inflation and unemployment. â�¢ to keep exchange rates within reasonable limits. furthermore, the one-dimensional variable of "money" is also used to influence individual behavior via taxes (by changing behavioral incentives). it is questionable, whether money can optimally meet all these goals at the same time (see sect. 16.3.7) . we believe that a computer, good food, friendship, social status, love, fairness, and knowledge can only to a certain extent be replaced by and traded against each other. probably for this reason, social exchange comprises more than just material exchange [131] [132] [133] . it is conceivable that financial markets as well are trying to meet too many goals at the same time. this includes: â�¢ to match supply and demand. â�¢ to discover a fair price. â�¢ to raise the foreign direct investment (fdi). â�¢ to couple local economies with the international system. â�¢ to facilitate large-scale investments. â�¢ to boost development. â�¢ to share risk. â�¢ to support a robust economy, and â�¢ to create opportunities (to gamble, to become rich, etc.). therefore, it would be worth stuyding the system from a cybernetic control perspective. maybe, it would work better to separate some of these functions from each other rather than mixing them. another aspect that tends to be overlooked in mainstream economics is the relevance of psychological and social factors such as emotions, creativity, social norms, herding effects, etc. it would probably be wrong to interpret these effects just as a result of perception biases (see sect. 16.3.1) . most likely, these human factors serve certain functions such as supporting the creation of public goods [102] or collective intelligence [134, 135] . as bruno frey has pointed out, economics should be seen from a social science perspective [136] . in particular, research on happiness has revealed that there are more incentives than just financial ones that motivate people to work hard [133] . interestingly, there are quite a number of factors which promote volunteering [132] . it would also be misleading to judge emotions from the perspective of irrational behavior. they are a quite universal and a relatively energy-consuming way of signalling. therefore, they are probably more reliable than non-emotional signals. moreover, they create empathy and, consequently, stimulate mutual support and a readiness for compromises. it is quite likely that this creates a higher degree of cooperativeness in social dilemma situations and, thereby, a higher payoff on average as compared to emotionless decisions, which often have drawbacks later on. finally, there is no good theory that would allow one to assess the relevance of information in economic systems. most economic models do not consider information as an explanatory variable, although information is actually a stronger driving force of urban growth and social dynamics than energy [137] . while we have an information theory to determine the number of bits required to encode a message, we are lacking a theory, which would allow us to assess what kind of information is relevant or important, or what kind of information will change the social or economic world, or history. this may actually be largely dependent on the perception of pieces of information, and on normative or moral issues filtering or weighting information. moreover, we lack theories describing what will happen in cases of coincidence or contradiction of several pieces of information. when pieces of information interact, this can change their interpretation and, thereby, the decisions and behaviors resulting from them. that is one of the reasons why socio-economic systems are so hard to predict: "unknown unknowns", structural instabilities, and innovations cause emergent results and create a dynamics of surprise [138] . the problems discussed in the previous two sections pose interesting practical and fundamental challenges for economists, but also other disciplines interested in understanding economic systems. econophysics, for example, pursues a physical approach to economic systems, applying methods from statistical physics [81] , network theory [139, 140] , and the theory of complex systems [85, 87] . a contribution of physics appears quite natural, in fact, not only because of its tradition in detecting and modeling regularities in large data sets [141] . physics also has a lot of experience how to theoretically deal with problems such as time-dependence, fluctuations, friction, entropy, non-linearity, strong interactions, correlations, heterogeneity, and many-particle simulations (which can be easily extended towards multi-agent simulations). in fact, physics has influenced economic modeling already in the past. macroeconomic models, for example, were inspired by thermodynamics. more recent examples of relevant contributions by physicists concern models of self-organizing conventions [54] , of geographic agglomeration [65] , of innovation spreading [142] , or of financial markets [143] , to mention just a few examples. one can probably say that physicists have been among the pioneers calling for new approaches in economics [81, 87, [143] [144] [145] [146] [147] . a particularly visionary book beside wolfgang weidlich's work was the "introduction to quantitative aspects of social phenomena" by elliott w. montroll and wade w. badger, which addressed by mathematical and empirical analysis subjects as diverse as population dynamics, the arms race, speculation patterns in stock markets, congestion in vehicular traffic as well as the problems of atmospheric pollution, city growth and developing countries already in 1974 [148] . unfortunately, it is impossible in our paper to reflect the numerous contributions of the field of econophysics in any adequate way. the richness of scientific contributions is probably reflected best by the econophysics forum run by yi-cheng zhang [149] . many econophysics solutions are interesting, but so far they are not broad and mighty enough to replace the rational agent paradigm with its large body of implications and applications. nevertheless, considering the relatively small number of econophysicists, there have been many promising results. the probably largest fraction of publications in econophysics in the last years had a data-driven or computer modeling approach to financial markets [143] . but econophyics has more to offer than the analysis of financial data (such as fluctuations in stock and foreign currency exchange markets), the creation of interaction models for stock markets, or the development of risk management strategies. other scientists have focused on statistical laws underlying income and wealth distributions, nonlinear market dynamics, macroeconomic production functions and conditions for economic growth or agglomeration, sustainable economic systems, business cycles, microeconomic interaction models, network models, the growth of companies, supply and production systems, logistic and transport networks, or innovation dynamics and diffusion. an overview of subjects is given, for example, by ref. [152] and the contributions to annual spring workshop of the physics of socio-economic systems division of the dpg [153] . to the dissatisfaction of many econophysicists, the transfer of knowledge often did not work very well or, if so, has not been well recognized [150] . besides scepticism on the side of many economists with regard to novel approaches introduced by "outsiders", the limited resonance and level of interdisciplinary exchange in the past was also caused in part by econophysicists. in many cases, questions have been answered, which no economist asked, rather than addressing puzzles economists are interested in. apart from this, the econophysics work was not always presented in a way that linked it to the traditions of economics and pointed out deficiencies of existing models, highlighting the relevance of the new approach well. typical responses are: why has this model been proposed and not another one? why has this simplification been used (e.g. an ising model of interacting spins rather than a rational agent model)? why are existing models not good enough to describe the same facts? what is the relevance of the work compared to previous publications? what practical implications does the finding have? what kind of paradigm shift does the approach imply? can existing models be modified or extended in a way that solves the problem without requiring a paradigm shift? correspondingly, there have been criticisms not only by mainstream economists, but also by colleagues, who are open to new approaches [151] . therefore, we would like to suggest to study the various economic subjects from the perspective of the above-mentioned fundamental challenges, and to contrast econophysics models with traditional economic models, showing that the latter leave out important features. it is important to demonstrate what properties of economic systems cannot be understood for fundamental reasons within the mainstream framework (i.e. cannot be dealt with by additional terms within the modeling class that is conventionally used). in other words, one needs to show why a paradigm shift is unavoidable, and this requires careful argumentation. we are not claiming that this has not been done in the past, but it certainly takes an additional effort to explain the essence of the econophysics approach in the language of economics, particularly as mainstream economics may not always provide suitable terms and frameworks to do this. this is particularly important, as the number of econophysicists is small compared to the number of economists, i.e. a minority wants to convince an established majority. to be taken seriously, one must also demonstrate a solid knowledge of related previous work of economists, to prevent the stereotypical reaction that the subject of the paper has been studied already long time ago (tacitly implying that it does not require another paper or model to address what has already been looked at before). a reasonable and promising strategy to address the above fundamental and practical challenges is to set up multi-disciplinary collaborations in order to combine the best of all relevant scientific methods and knowledge. it seems plausible that this will generate better models and higher impact than working in separation, and it will stimulate scientific innovation. physicists can contribute with their experience in handling large data sets, in creating and simulating mathematical models, in developing useful approximations, in setting up laboratory experiments and measurement concepts. current research activities in economics do not seem to put enough focus on: â�¢ modeling approaches for complex systems [154] . â�¢ computational modeling of what is not analytically tractable anymore, e.g. by agent-based models [155] [156] [157] . â�¢ testable predictions and their empirical or experimental validation [164] . â�¢ managing complexity and systems engineering approaches to identify alternative ways of organizing financial markets and economic systems [91, 93, 165] , and â�¢ an advance testing of the effectiveness, efficiency, safety, and systemic impact (side effects) of innovations, before they are implemented in economic systems. this is in sharp contrast to mechanical, electrical, nuclear, chemical and medical drug engineering, for example. expanding the scope of economic thinking and paying more attention to these natural, computer and engineering science aspects will certainly help to address the theoretical and practical challenges posed by economic systems. besides physics, we anticipate that also evolutionary biology, ecology, psychology, neuroscience, and artificial intelligence will be able to make significant contributions to the understanding of the roots of economic problems and how to solve them. in conclusion, there are interesting scientific times ahead. it is a good question, whether answering the above list of fundamental challenges will sooner or later solve the practical problems as well. we think, this is a precondition, but it takes more, namely the consideration of social factors. in particular, the following questions need to be answered: 8. how do costly punishment, antisocial punishment, and discrimination come about? 9. how can the formation of social norms and conventions, social roles and socialization, conformity and integration be understood? 10. how do language and culture evolve? 11. how to comprehend the formation of group identity and group dynamics? what are the laws of coalition formation, crowd behavior, and social movements? 12. how to understand social networks, social structure, stratification, organizations and institutions? 13. how do social differentiation, specialization, inequality and segregation come about? 14. how to model deviance and crime, conflicts, violence, and wars? 15. how to understand social exchange, trading, and market dynamics? we think that, despite the large amount of research performed on these subjects, they are still not fully understood. the ultimate goal would be to formulate mathematical models, which would allow one to understand these issues as emergent phenomena based on first principles, e.g. as a result of (co-)evolutionary processes. such first principles would be the basic facts of human capabilities and the kinds of interactions resulting from them, namely: 1. birth, death, and reproduction. 2. the need of and competition for resources (such as food and water). 3. the ability to observe their environment (with different senses). 4. the capability to memorize, learn, and imitate. 5. empathy and emotions. 6. signaling and communication abilities. 7. constructive (e.g. tool-making) and destructive (e.g. fighting) abilities. 8. mobility and (limited) carrying capacity. 9. the possibility of social and economic exchange. such features can, in principle, be implemented in agent-based models [158] [159] [160] [161] [162] [163] . computer simulations of many interacting agents would allow one to study the phenomena emerging in the resulting (artificial or) model societies, and to compare them with stylized facts [163, 168, 169] . the main challenge, however, is not to program a seemingly realistic computer game. we are looking for scientific models, i.e. the underlying assumptions need to be validated, and this requires to link computer simulations with empirical and experimental research [170] , and with massive (but privacy-respecting) mining of social interaction data [141] . in the ideal case, there would also be an analytical understanding in the end, as it has been recently gained for interactive driver behavior [111] . as well as for inspiring discussions during a visioneer workshop in zurich from january 13 how to understand human decision-making? how to explain deviations from rational choice theory and the decision-theoretical paradoxes? why are people risk averse? 2. how does consciousness and self-consciousness come about? how to understand creativity and innovation? 4. how to explain homophily, i.e. the fact that individuals tend to agglomerate, interact with and imitate similar others? how to explain social influence, collective decision making, opinion dynamics and voting behavior? why do individuals often cooperate in social dilemma situations? how do indirect reciprocity, trust and reputation evolve? references 1 how did economists get it so wrong? the financial crisis and the systemic failure of academic economics (dahlem report address of the governor of the central bank of barbados on the futurict knowledge accelerator: unleashing the power of information for a sustainable future global solutions: costs and benefits economics as religion: from samuelson to chicago and beyond the counter revolution of science the case of flexible exchange rates adaptive behavior and economic theory bounded rationality. the adaptive toolbox the bounds of reason: game theory and the unification of the behavioral sciences on information efficiency and financial stability simon models of man simple heuristics that make us smart judgment under uncertainty: heuristics and biases physics of risk and uncertainty in quantum decision making a theory of fairness, competition, and cooperation foundations of human sociality: economic experiments and ethnographic evidence from fifteen small-scale societies social distance and other-regarding behavior in dictator games measuring social value orientation economics and restaurant gratuities: determining tip rates the cement of society: a study of social order the rewards of punishment. a relational theory of norm enforcement the emergence of homogeneous norms in heterogeneous populations the emperor's dilemma: a computational model of selfenforcing norms non-explanatory equilibria: an extremely simple game with (mostly) unattainable fixed points the myth of the folktheorem deterministic chaos for example, three-body planetary motion has deterministic chaotic solutions, although it is a problem in classical mechanics, where the equations of motion optimize a lagrangian functional on formally undecidable propositions of principia mathematica and related systems on computable numbers, with an application to the entscheidungsproblem the behavior of stock market prices an inquiry into the nature and causes of the wealth of nations the theory of moral sentiments (1759) evolution of indirect reciprocity revisiting market efficiency: the stock market as a complex adaptive system there is no invisible hand. the guardian optimal self-organization traffic and related self-driven many-particle systems from crowd dynamics to crowd safety: a video-based analysis stability and stabilization of time-delay systems network-induced oscillatory behavior in material flow networks and irregular business cycles the tragedy of the commons altruistic punishment in humans switching phenomena in a system with no switches animal spirits: how human psychology drives the economy, and why it matters for global capitalism testing behavioral simulation models by direct experiment a mathematical model for behavioral changes by pair interactions multistability in a dynamic cournot game with three oligopolists evolutionary establishment of moral and double moral standards through spatial interactions abzweigungen einer periodischen lã¶sung von einer stationã¤ren lã¶sung eines differentialgleichungssystems dynamic decision behavior and optimal guidance through information services: models and experiments the dynamics of general equilibrium modeling the stylized facts in finance through simple nonlinear adaptive systems predicting moments of crisis in physics and finance" during the workshop "windows to complexity the chemical basis of morphogenesis lectures on nonlinear differential equation-models in biology settlement formation ii: numerical simulation self-organization in space and induced by fluctuations stokes integral of economic growth: calculus and the solow model hysteresis and the european unemployment problem pattern formation and dynamics in nonequilibrium systems critical phenomena in natural sciences: chaos, fractals, selforganization and disorder the science of disasters noise-induced transitions: theory and applications in physics, chemistry, and biology brownian motors: noisy transport far from equilibrium catastrophe theory introduction to phase transitions and critical phenomena managing complexity: an introduction. pages 1-16 in systemic risks in society and economics systemic risk in a unifying framework for cascading processes on networks how nature works: the science of self-organized criticality why economists failed to predict the financial crisis physics and social science-the approach of synergetics nonlinear economic dynamics the self-organizing economy self-organization of complex structures: from individual to collective dynamics complex economic dynamics sociodynamics: a systematic approach to mathematical modelling in the social sciences business dynamics: systems thinking and modeling for a complex world growth theory, non-linear dynamics and economic modelling complexity hints for economic policy coping with the complexity of economics handbook of research on complexity the rise and decline of nations: economic growth, stagflation, and social rigidities the evolution of cooperation a general theory of equilibrium selection phase transitions to cooperation in the prisoner's dilemma evolutionary games and spatial chaos social diversity promotes the emergence of cooperation in public goods games the outbreak of cooperation among success-driven individuals under noisy conditions governing the commons. the evolution of institutions for collective action topological traps control flow on real networks: the case of coordination failures modelling supply networks and business cycles as unstable transport phenomena the selforganization of matter and the evolution of biological macromolecules the theory of economic development the changing face of economics. conversations with cutting edge economists beyond the representative agent economics with heterogeneous interacting agents modeling aggregate behavior and fluctuations in economics collection of papers on an analytical theory of traffic flows in the economy as an evolving complex system ii the economy as an evolving complex system iii evolutionary economics ecology for bankers origin of wealth: evolution, complexity, and the radical remaking of economics introduction to forest ecosystem science and management models of biological pattern formation multi-objective management in freight logistics: increasing capacity, service level and safety with optimization algorithms handbook of research on nature-inspired computing for economics and management self-control of traffic lights and vehicle flows in urban road networks biologistics and the struggle for efficiency: concepts and perspectives analytical calculation of critical perturbation amplitudes and critical densities by non-linear stability analysis of a simple traffic flow model complexity cube for the characterization of complex production systems verfahren zur koordination konkurrierender prozesse oder zur steuerung des transports von mobilen einheiten innerhalb eines netzwerkes (method for coordination of concurrent processes for noise sensitivity of portfolio selection under various risk measures the logic of failure: recognizing and avoiding error in complex situations complexity and the enterprise: the illusion of control nurturing breakthroughs: lessons from complexity theory monetary relationships: a view from threadneedle street. (papers in monetary economics, reserve bank of australia, 1975); for applications of le chatelier's principle to economics see also p. a. samuelson, foundations of economic analysis structures of social life: the four elementary forms of human relations understanding and assessing the motivations of volunteers: a functional approach happiness: a revolution in economics the wisdom of crowds: why the many are smarter than the few and how collective wisdom shapes business swarm intelligence. introduction and applications economics as a science of human behaviour: towards a new social science paradigm growth, innovation, scaling and the pace of life in cities uncertainty and surprise in complex systems scale-free networks: a decade and beyond economic networks: the new challenges hyperselection and innovation described by a stochastic model of technological evolution introduction to econophysics: correlations and complexity in finance minority games: interacting agents in financial markets economics needs a scientific revolution the economy needs agent-based modelling meltdown modelling introduction to quantitative aspects of social phenomena econophysics forum fifteen years of econophysics: worries, hopes and prospects worrying trends in econophysics econophysics and sociophysics: trends and perspectives aims and scope of the physics of socio-economic systems division of the german physical society pluralistic modeling of complex systems handbook of computational economics simulation modeling in organizational and management research developing theory through simulation methods understanding complex social dynamics: a plea for cellular automata based modelling platforms and methods for agent-based modeling from factors to actors: computational sociology and agent-based modeling artificial societies. multiagent systems and the micro-macro link in sociological theory generative social science: studies in agent-based computational modeling the handbook of experimental economics managing complexity: concepts, insights, applications engineering economy statistical physics of social dynamics the future of social experimenting the authors are grateful for partial financial support by the eth competence center "coping with crises in complex socio-economic systems" (ccss) through eth research grant ch1-01 08-2 and by the future and emerging technologies programme fp7-cosi-ict of the european commission through the project visioneer (grant no.: 248438). they would like to thank for feedbacks on the manuscript by kenett dror, tobias preis and gabriele tedeschi key: cord-293148-t2dk2syq authors: nadini, matthieu; zino, lorenzo; rizzo, alessandro; porfiri, maurizio title: a multi-agent model to study epidemic spreading and vaccination strategies in an urban-like environment date: 2020-09-22 journal: appl netw sci doi: 10.1007/s41109-020-00299-7 sha: doc_id: 293148 cord_uid: t2dk2syq worldwide urbanization calls for a deeper understanding of epidemic spreading within urban environments. here, we tackle this problem through an agent-based model, in which agents move in a two-dimensional physical space and interact according to proximity criteria. the planar space comprises several locations, which represent bounded regions of the urban space. based on empirical evidence, we consider locations of different density and place them in a core-periphery structure, with higher density in the central areas and lower density in the peripheral ones. each agent is assigned to a base location, which represents where their home is. through analytical tools and numerical techniques, we study the formation mechanism of the network of contacts, which is characterized by the emergence of heterogeneous interaction patterns. we put forward an extensive simulation campaign to analyze the onset and evolution of contagious diseases spreading in the urban environment. interestingly, we find that, in the presence of a core-periphery structure, the diffusion of the disease is not affected by the time agents spend inside their base location before leaving it, but it is influenced by their motion outside their base location: a strong tendency to return to the base location favors the spreading of the disease. a simplified one-dimensional version of the model is examined to gain analytical insight into the spreading process and support our numerical findings. finally, we investigate the effectiveness of vaccination campaigns, supporting the intuition that vaccination in central and dense areas should be prioritized. the number of people living in urban areas has already exceeded 4 billions and it is estimated to reach 7 billions by 2050 (ritchie and roser 2020) . global urbanization poses new challenges in different sectors, ranging from transportation to energy supply, environmental degradation, and healthcare (cohen 2006) . among these challenges, understanding how urban environments shape the evolution of epidemic outbreaks and designing effective containment strategies have recently drawn considerable attention from researchers and media. paradigmatic are the examples of recent outbreaks, such as the 2003 sars (smith 2006 mers (de groot et al. 2013 , and the ongoing covid-19 (chang et al. 2020; chinazzi et al. 2020; ferguson et al. 2020) . analyzing how diseases spread within urban environments has been the topic of various experimental and theoretical studies (eubank et al. 2004; satterthwaite 2007; alirol et al. 2011; neiderud 2015; telle et al. 2016; massaro et al. 2019) . experimental studies have offered a detailed analysis of urban environments (satterthwaite 2007; neiderud 2015) , suggesting specific preventive measures for both urban residents and travelers (alirol et al. 2011) . theoretical studies have provided insights on how to contain outbreaks (eubank et al. 2004) , as well on possible key drivers of contagion, such as the role of human mobility patterns (massaro et al. 2019 ) and socio-economical risk factors (telle et al. 2016) . despite the importance of urban environments in the global diffusion of diseases (brockmann and helbing 2013) , how epidemic outbreaks unfold therein is yet to be fully elucidated. some attempts to mathematically describe the diffusion of diseases within and among cities can be found in metapopulation models (colizza and vespignani 2007; colizza and vespignani 2008; liu et al. 2013) . in these models, a fixed network of spatial localities is used to model the mobility patterns between cities, where homogeneously-mixed populations are affected by the epidemic. while metapopulation models can be, at least partially, tackled through analytical methods (colizza and vespignani 2007; colizza and vespignani 2008; liu et al. 2013) , considerable experimental evidence challenges the assumption of homogeneously-mixed populations, which could yield misleading estimates of the extent of epidemic outbreaks (pastor-satorras et al. 2015) . on the other side of the spectrum of epidemic models, agent-based models (eubank et al. 2004; degli atti et al. 2008; gilbert 2008 ) constitute a valuable framework to offer a realistic description of how diseases diffuse within urban environments. currently, this class of models is being leveraged to predict the diffusion of the covid-19 (chang et al. 2020; ferguson et al. 2020) , informing the design and implementation of timely containment measures. however, those advantageous features are accompanied by some drawbacks, including the need of mobility data and models, the use of massive computational resources when the system size scales up, and the lack of analytical techniques for model characterizations. a viable approach to agent-based modeling is based on two-dimensional representations, where agents move and interact according to proximity criteria (frasca et al. 2006; frasca et al. 2008; zhou and liu 2009; buscarino et al. 2010; yang et al. 2012; buscarino et al. 2014; huang et al. 2016; peng et al. 2019) . as a first approximation, the motion of the agents can be described according to a random walk with sporadic long range jumps (frasca et al. 2006) . building on this approximation, it is possible to include realistic features such as nonhomogeneous infection rates (buscarino et al. 2014 ) and heterogeneous radii of interaction (huang et al. 2016; peng et al. 2019) . much work is needed, however, to fully capture and describe realistic patterns of human mobility, which are shaped by the complex structure of urban environments (alessandretti et al. 2017) . here, we contribute to the field of agent-based modeling by presenting a twodimensional model that is capable of reproducing a spatially inhomogeneous urban-like environment, in which a heterogeneous population follows realistic rules of mobility. inspired by previous theoretical studies (huang et al. 2016; peng et al. 2019) , we assume that agents have a heterogeneous radius of interaction, which accounts for variations among individuals in their involvement in social behavior and activities. we consider a urban-like environment composed of multiple locations, each of them representing a well-defined region of the urban space (that is, a neighborhood of a city). through this spatial organization, our model is able to encapsulate two key features of urban environments. first, it can reproduce typical core-periphery structures, where central regions are more densely populated than peripheral ones (makse et al. 1998; de nadai et al. 2016) . second, it allows to mimic the inhomogeneity in movement patterns of humans, where people tend to spend most of their time in a few neighborhoods -for example, experimental studies suggest that individuals spend most time either at home or at work, while only sporadically visiting other neighborhoods (brasche and bischof 2005; schweizer et al. 2007; matz et al. 2015) . to reproduce realistic conditions for agents' movement patterns, we posit two different mobility schemes, applied within and outside the agents' base location (that is, where their home is). while the homogeneous mixing assumption seems reasonable within the agents' base location, we assume that agents tend to move outside of their base location following a gravity model and a biased random walk. hence, agents are more likely to explore regions close to their base location rather than remotely-located regions (brasche and bischof 2005; schweizer et al. 2007; matz et al. 2015) . from this mobility pattern, we construct a network of contacts, whose topology is examined in this study. through some mathematical derivations and numerical simulations, we seek to identify analogies between the proposed agent-based model and existing temporal network approaches, where spatial mobility is lumped into nodal parameters (perra et al. 2012; zino et al. 2016; zino et al. 2018; nadini et al. 2018; nadini et al. 2020) . we adopt the proposed framework to study how urban-like environments shape the diffusion of infectious diseases, using the illustrative epidemic models with the possibility of reinfection (susceptible-infected-susceptible, or sis) or permanent removal (susceptible-infected-removed, or sir) (keeling and rohani 2011) . our results confirm the intuition that agents' density plays a critical role on the diffusion of both sis and sir processes. in the limit case where the entire urban area consists of one location, agents that move outside the location only seldom interact with other agents, thereby hindering the contagion process. in the more realistic scenario of a core-periphery structure with multiple locations, we unexpectedly find that the time spent by agents in their base location does not influence the endemic prevalence in the sis model and the epidemic size in the sir model, which are measures of the overall fraction of population that is affected by the disease. a possible explanation for this counterintuitive phenomenon may be found in the agents' mobility rules. in fact, commuting patterns that bring agents from central areas to peripheral ones may yield a reduction in the diffusion in the central areas. contrarily, commuting patterns from peripheral to central areas lead to the opposite effects. to detail the working principles of this unexpected result, we present a minimalistic one-dimensional version of the model, which is amenable to a complete analytical treatment, thereby offering some preliminary analytical insight into the role of model parameters on epidemic spreading. we also explore the interplay between the agents' radius of interaction and their positioning in the core-periphery structure. we find that when agents' with larger radii are assigned to the less dense and peripheral locations, then the endemic prevalence (in the sis model) and the epidemic size (in the sir model) strongly decrease with respect to a random assignment. moreover, when agents' with larger radii are assigned to denser (and central) locations, the fraction of population affected by the disease is not sensibly increased. hence, our results support the intuition that more central areas are the crossroads of individuals commuting in a city and are critical for the spread of diseases. finally, we numerically analyze the effect of targeted vaccination strategies, which consist of immunizing a portion of the population in a specific location, prior to the disease onset. consistent with the intuition that central locations play a key role on the spread of epidemic diseases, we find that the best strategy is to prioritize the vaccination of agents belonging to central urban areas. the rest of the manuscript is organized as follows. in table 1 , we summarize the notation and the nomenclature used throughout the paper. in "model", we introduce the model of agents' mobility. in "temporal network of contacts", we describe and analyze the temporal network formation mechanism. in "epidemic processes", we analytically and numerically study the spread of epidemic processes and compare several vaccination strategies. in "discussion and conclusion" sections, we discuss our main findings and propose further research directions. we consider n ≥ 1 agents, labeled by positive integers v := {1, . . . , n} . agents move in a square planar space with side length d > 0 and with periodic boundary conditions (frasca et al. 2006) , that is, when an agent exits through one side of the square planar space, it re-appears on the opposite side. the position of agent i ∈ v at the discrete time t ∈ z ≥0 in a cartesian reference frame is denoted by ( we deploy the n agents over l locations, each of them representing a bounded portion of the square space. the set of all locations is l = {1, . . . , l} and each location ∈ l occupies a convex region of the planar space we assume that all the locations are mutually disjoint and we order them in ascending order according to their area, that is, a 1 ≤ · · · ≤ a l . we hypothesize that a l d 2 , that is, each location is much smaller than the whole square space. each agent is assigned a specific base location (that is, their home) according to a map: β : v −→ l; we assume that each base location is associated with the same number of agents, n = n/l. 1 as a result, the density of agents assigned to location , varies with the location. also, locations are sorted in descending order of density, that is ρ 1 ≥ · · · ≥ ρ l . for simplicity, in the numerical simulations implemented throughout this paper, the convex regions are taken as circles with nondecreasing radii 1 ≤ · · · ≤ l . inspired by empirical and theoretical studies (witten jr and sander 1981; vicsek 1992; makse et al. 1998; de nadai et al. 2016) , radii of the locations are extracted from a power law distribution with probability density function p( ) ∝ −γ , with bilateral cutoffs such that ∈ [ min , max ], for any ∈ l. the upper bound guarantees that all locations fit in the exponents of the power law distribution of locations' radii min , max lower and higher cut-off of the distribution of locations' radii g(σ ) probability density function of agents' radii of interaction ω exponents of the power law distribution of radii of interaction σ min , σ max lower and higher cut-off of the distribution of radii of interaction k i degree of agent i λ infection probability per contact μ recovery probability per unit time fraction of susceptible, infected, and recovered agents in the system fraction of susceptible, infected, and recovered agents in fraction of susceptible, infected, and recovered agents at distance d from the closest location (t) contagion probability in out,d (t) contagion probability at a distance d from the closest location · statistical average of the quantity "·" expected value of the quantity "·" probability of an event "·" square, and the lower bound sets a maximum to the locations' density. since the radii are power law distributed with exponent −γ , the areas of the locations are also power law distributed with exponent −2γ and cutoffs such that a ∈ π 2 min , π 2 max . empirical studies on urban environments suggest that cities are constructed according to a core-periphery structure, whereby locations with smaller areas and denser population are located in their center, while locations with larger areas and sparser population pertain to peripheral areas (makse et al. 1998; de nadai et al. 2016) , as shown in fig. 1a . we implement a heuristic algorithm to generate a locations' layout according to a coreperiphery structure and qualitatively reproduce empirical results. figure 1b shows the output generated by our algorithm, whose structure is qualitatively consistent with the empirical observations reported in fig. 1a . details of the algorithm used to create such a core-periphery structure are presented in appendix a. at the present time, the technical literature has yet to empirically study the relationship between the agents' radius of interaction and the density of their base location. in this paper, we explore different scenarios aiming at offering a first theoretical understanding of the impact of this potential relationship on the evolution of disease processes. unless otherwise specified, we consider that the n members of each location are randomly chosen, independently of their radius of interaction. we also examine the cases in which there is a correlation (positive or negative) between the agents' radius of interaction and the density in their base location: a positive correlation means that agents with larger radius are assigned to denser (central) locations, while a negative correlation identifies the case in which agents with larger radius are placed in the less dense (peripheral) locations. agents' positions evolve according to a discrete-time dynamics. hence, their positions are updated at each discrete time-step t ∈ z ≥0 . the law of motion of the generic agent i depends on whether it is outside or inside its base location β(i) ∈ l. if agent i ∈ v is outside its base location, that is, (x i (t), y i (t)) / ∈ β(i) , it performs a biased random walk fig. 1 qualitative comparison between real datasets from an experimental study (de nadai et al. 2016) , and the output of our algorithm. a experimental results about human digital activity density in the cities of milan and rome, italy. the highest density is registered in central areas, while lower densities are observed in peripheral ones. b using our algorithm, we generate l = 1, 000 circular locations distributed in rings of decreasing densities. the first few rings contain the denser locations (darker central regions) and may parallel the city center of a urban environment, while the outer rings are less dense and represent peripheral areas (light gray regions). source of a: (de nadai et al. 2016) . parameters used to generate b: d = 1, 000, min = 3, max = 30, and γ = 2.1. details of the generative algorithm used are available in appendix a toward its base location; 2 on the contrary, if it is inside its base location, it can move to a random position (within its base location), or exit according to a probabilistic mechanism. specifically, if the agent is not in its base location, then here, v > 0 is the (constant) speed and θ i (t) is an angle, determined as the sum of two terms: ( 3 ) the first term, i (t), represents the angle of the direction of the shortest path from (x i (t), y i (t)) to the region β(i) , that is, to the agent base location. this quantity is formally defined by introducing so that the second term, α θ it , is modulated by θ it , which is a random variable that takes values uniformly in [ −π, π] and is extracted independently at every time t and for every agent i, and by α ∈[ 0, 1], which is a randomness parameter that regulates how much the agents tend to deviate from the shortest path to return to their base location, when they are outside it. when α = 1, the agent moves completely at random, while, when α = 0, it moves along the shortest path toward its location. when the agent is in its base location, (x i (t), y i (t)) ∈ β(i) , the law of motion is defined as follows. given a parameter p ∈[ 0, 1] (constant in time and equal for all agents), with probability 1 − p, the agent moves to a position chosen uniformly at random within its base location, so that its position is completely uncorrelated with the previous one. otherwise, with probability p, the agent jumps outside its base location, ending in a position of the remaining space according to a distance decay law. in particular, we assume that the distance from the border of the base location at which an agent jumps is the realization of a random variable exponentially distributed with exponent c > 0. the corresponding probability density function p jump (d) is equal to for d ≥ 0. hence, the expected distance at which an agent jumps is equal to 1/c. a sensible choice of the exponent in the law in eq. (6) yields a typical behavior observed in many empirical studies (levinson and el-geneidy 2009; boussauw et al. 2011) , whereby agents tend to gravitate within and around their base location, while sporadically initiating journeys toward further locations (liu et al. 2014) . two salient snapshots of agents' motion are illustrated in fig. 2a and c. . direction and modulus of the agents' velocity is drawn with solid red arrows. the position where agent 1 will jump is indicated with a dotted red arrow. the four arrows around an agent indicate that it will move in a random position inside its own location. in b and d, we show the temporal network formation mechanism. agents' radii of interaction are represented with solid circles, and undirected links are represented with solid blue lines upon the mobility model, we construct the network of contacts, which is the means through which the disease is transmitted. in this vein, agents create undirected temporal links based on proximity with other agents. specifically, agent i ∈ v contacts all other agents located within a circle of radius σ i centered in its current position (x i (t), y i (t)). we assume that agents have heterogeneously distributed radii, extracted from a power law distribution with probability density function an undirected temporal link between two agents i and j is created when the euclidean distance at time t between the position of agent i, (x i (t), y i (t)), and the position of agent j, x j (t), y j (t) , is less than or equal to the maximum of the two radii σ i and σ j , that is, figure 2b and d show two consecutive instances of the network formation process. toward modeling of epidemics in urban environments, our model allows agents inside a location to interact with agents outside the location, see, for example, agents 2 and 3 in fig. 2a and b. the intricacy of the motion patterns and the nonsmooth process for generating the network of contacts hinder the analytical tractability of the model in its general formulation. however, for some cases it is possible to establish analytical insight on some model features. in appendix b, we analyze the system in the two specific cases of: i) a free space without any location (l = 0), and ii) when the law of motion of the agents outside their base locations is deterministic (α = 0) and the locations are uniformly distributed in the plane. in these two cases it is possible to apply a meanfield approach in the limit of large systems (n → ∞) to analytically study the number of connections generated by the agents, which represent potential paths of infection throughout the population. therein, numerical simulations for large systems are provided to validate theoretical findings. the general case of a core-periphery structure and stochasticity in the motion out of the location is treated only through numerical simulations, in which we record all the interactions and use their time-evolution over sufficiently long time-windows (t 1, where t is the duration of the observation) to study key topological features (average degree and clustering coefficient). in fig. 3a , we consider the case without locations. our numerical results are consistent with analytical predictions in appendix b, which are exact in the thermodynamic limit of large systems n → ∞. specifically, the expected degree of agent i is equal to so that agents with a larger radius of interaction have a greater average degree. note that when the agent radius is close to the minimum, that is, σ i ≈ σ min , eq. (8) is dominated by the second summand, while when the radius is close to the maximum, that is, σ i ≈ σ max , the right-hand side of eq. (8) scales with σ 2 i . equation 8 highlights a nontrivial relationship between the expected degree of an agent and its radius of interaction, which is due to the links passively received by the agent when it is located within the radii of interaction of other agents. this relationship is different from the case of directed interactions analyzed in huang et al. (2016) ; peng et al. (2019) , where e[ k i ] is proportional to σ 2 i . in fig. 3b , we examine the case of multiple locations uniformly distributed in the plane. based on the theoretical derivations in appendix b, we obtain the following expression for the expected number of interactions that agent i establishes in its own location in the thermodynamic limit of large systems n → ∞: we observe that e k in,i is inversely proportional to the square of the radius of location β(i) , that is, 2 β(i) . in fig. 3b , we multiply the numerical estimation of each agents' average degree by the square of the radius of the corresponding location, to allow a graphical representation of the comparison between numerical estimations and analytical predictions. numerical results in finite-size systems are in close agreement with analytical predictions of eq. (9), which are exact in the limit of large systems. in order to offer insight into the influence that a core-periphery structure has on the agents' average degree, we analyze three different scenarios. first, we study the case in (9) provides the expected degree, which is numerically estimated by tracking the corresponding agent in time. numerical results are presented using different colors and markers, corresponding to each of the locations (numerical findings share a common trend, which is well captured by the theory). in the simulations, we use the following parameters: l = 10, d = 10 9 , min = 1, 000, max = 10, 000, σ min = 10, σ max = 100, p = 0.3, and α = 0. agents are initially inside their base location and interactions are recorded after 100 steps to allow agents to reach a steady-state configuration. other parameter values are n = 10, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, and t = 5, 000 which agents are strongly tied to their base location, such that they have low probability of jumping outside their base location (small p) and low probability of deviating from the shortest path to return to the base location, when they are outside (small randomness α), in fig. 4a . second, we examine the case in which the probability of jumping outside their base location and the agents' randomness in the random walk are intermediate, in fig. 4b . finally, we investigate the case in which agents tend to spend most of their time outside their base location (large p and α), in fig. 4c . as expected from the formulation of the model, we determine that agents with larger radii of interaction tend to have larger average degrees. also, agents with larger radii of interaction are more likely to contact agents outside of their base location, thereby leading to lower clustering coefficients c (saramäki et al. 2007 ), which is a measure of the agents' tendency to form clusters. 3 the results of our simulations are reported in figs. 4d-f. during the evolution of an epidemic process, agents with large radii might act as "superspreaders (lloydsmith et al. 2005) , " which are known to have a key role on the disease spreading, by creating many connections and infecting agents from different locations. less expected are the relationships between agents' radii of interaction and their base location, and between agents' clustering coefficients and their radii of interaction. among the agents with a small radius of interaction, the agents that are assigned to central locations have a larger average degree than those that are assigned to peripheral locations. this result is independent of the time spent outside their base location (that is, independent of p and α). interestingly, the same argument does not apply when agents have a fig. 4 influence of the location radius on the agents' average degree a-c and clustering coefficient d-f, for three different parameter settings. average degree and cluster coefficient are numerically estimated by tracking every agent in the system. darker circles represent agents assigned to more peripheral locations, while brighter ones indicates agents belonging to more central locations. we set: a-d p = 0.1 and α = 0, b-e p = 0.4 and α = 0.2, and c-f p = 0.8 and α = 0.4. agents are initially inside their base location and contacts are recorded after 100 steps to allow the agents to reach a steady-state configuration. other parameter values are l = 100, n = 10, 000, d = 10 9 , min = 100, max = 10, 000, σ min = 1, σ max = 1, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, and t = 5, 000 large radius of interaction. in this situation, agents assigned to peripheral locations may have a larger degree than agents assigned to central locations, because their high radius of interaction allows a multitude of interactions, independent of the position of their base location. in addition, agents assigned to central locations have a lower clustering coefficient than agents assigned to peripheral locations. this is because the former group interacts with more agents and creates less tight clusters than the latter group that is assigned to peripheral locations. further, we comment that time spent outside the base location (regulated by p and α) is inversely proportional to the dispersion of the agents' degree. in fact, the largest dispersion in the agents' degree is registered when the probability of jumping outside the base location and the agent's randomness are small, in fig. 4a . dispersion in the agents' degree decreases as the probability of jumping outside the base location and the agent's randomness increase, in fig. 4b and in fig. 4c . a possible explanation for this phenomenon can be based on the following argument. the more the agents spend time inside their base location, the more they remain isolated from other agents in the system. on the contrary, agents' isolation is reduced when they spend more time outside their base location: they are able to interact with all the agents in the system, and, as a consequence, the dispersion in their degree decreases. here, we investigate the spreading of epidemics over spatially-distributed populations that behave according to the presented agent-based model. even though the complexity of the mobility mechanism and the presence of a geographical structure hinders the general mathematical treatment of the epidemics, some mathematical insight can be obtained by studying a simplified, one-dimensional version of the model. we start by presenting the one-dimensional simplification, discussing our main analytical results and validating them against numerical simulations. specifically, we focus on the impact of three salient model characteristics on epidemic processes. namely, i) the random exploration of the space governed by the parameter α, ii) the probability of jumping outside the base location p, and iii) the presence of multiple locations. interestingly, when multiple locations are present, the time spent inside the base location does not play an important role in the evolution of the contagion process. then, we consider the two-dimensional agent-based model and explore the effect of the same three salient model characteristics. we determine that results are qualitatively equivalent to those obtained in the one-dimensional case. we continue our numerical campaign on the two-dimensional agent-based model by studying whether the disease spreading is influenced by the presence of agents with larger radii in specific regions of the core-periphery structure. to this end, we study the presence of agents with greater radius of interaction in either the more central or more peripheral locations, thereby discovering that central locations are important for sustaining the overall diffusion. finally, we analyze the outcome of vaccination strategies, finding that the highest beneficial effect for the entire population is registered when the vaccination of agents in central locations is prioritized. we consider an infectious disease with the possibility of re-infection (sis model) or immunization (sir model), after the contraction of the infection. in the sis model, agents can be either susceptible to the disease or infected (keeling and rohani 2011) . two mechanisms characterize the epidemic dynamics: infection propagation and recovery process. the former occurs when an infected agent contacts a susceptible one, who may become infected with a probability λ, independently of the others. the latter consists of the spontaneous transition from the infected state to the susceptible one and occurs with probability μ at each unit time, independently of the others. in the sir model, instead, individuals who recover cannot be infected again and transition from the infected state to a removed state with probability μ per unit time (keeling and rohani 2011) . in the sis model, we examine the endemic prevalence (that is, the number of active cases in the long-term), which has typically two possible outcomes: either it quickly dies out and tends to zero, or it fluctuates around a quantity greater than zero for a nonnegligible amount of time, denoted by i * . for the sir model, instead, the fraction of infected individuals in the system always goes to zero in the long-run. however, the total fraction of individuals who have been infected may vary, depending on the model parameters. the sir epidemic size, denoted as r ∞ , is defined as the fraction of recovered individuals at the end of the epidemic process. here, we propose a one-dimensional model that provides some analytical intuitions on the influence that the randomness α, the probability of jumping outside the base location p, and the presence of a core-periphery structure have in the evolution of sis and sir epidemic processes. this model simplifies the two-dimensional case study by constraining agents to move in a discrete, infinitely long, one-dimensional lattice with periodic boundary conditions (that is, a ring). the l locations occupy consecutive positions on the lattice (labeled from 1 to l), and a fixed number of n = n/l agents belong to each one, as their base location. to generate a contact, agents should occupy the same position along the lattice. agents belong to a unique base location in the lattice, which they may leave with probability p. we use a geometric distribution (chung and zhong 2001) to describe the agents' law of motion, that is, the probability of jumping at a distance d from the base location is equal to where c ∈ (0, 1) is a constant parameter that governs the decay rate, similar to eq. (6). once outside their base location, agents move toward their base location by making one step toward it, similar to the two-dimensional model with α = 0. a schematic representation of the one-dimensional model is provided in fig. 5 . we remark that this one-dimensional model maintains some key features of the original two-dimensional agent-based model, that is: i) the presence of closely-spaced base locations, ii) a stochastic mechanism that governs the probability of jumping outside the base location, and iii) a gravity law that biases the agents to jump close to the base location according to an exponential distribution. a key feature that is not captured by this simplified model is the heterogeneity in the locations' density and agents' radii of interaction, which are numerically investigated in the two-dimensional model. we start our analysis by reporting the probability that a generic agent i ∈ v is inside location , which is explicitly derived in appendix c, similarly, the probability that a generic agent is in a position that is not occupied by any location and at a distance d from the closest location is computed in appendix c as where we assume that the closest location is = 1. by a simple change of variables, we can write an equivalent expression when the closest location is = l. in the sir and sis processes, the disease propagates from infected agents to susceptible ones occupying the same position of the one-dimensional lattice. we define as s(t), i(t), and (for the sir model only) r(t) the fractions of susceptible, infected, and recovered agents at time t, respectively. for large-scale systems, we can compute the fraction of susceptible, infected, and recovered agents along the lattice by using the law of large numbers (chung and zhong 2001) . in the thermodynamic limit of large systems n → ∞, the fraction of susceptible, infected, and recovered agents inside location is s (t) = q s(t), i (t) = q i(t), and r (t) = q r(t), respectively. similarly, the fraction of susceptible, infected, and recovered agents at a distance d from the closest location is s out,d (t) = q out,d s(t), i out,d (t) = q out,d i(t), and r out,d (t) = q out,d r(t), respectively. in the thermodynamic limit of large systems n → ∞, the evolution of the fraction of infected agents at time t + 1 is determined by the following equation: where (t) is the contagion probability of an agent inside its base location at time t, that is and out,d (t) is the contagion probability of an agent at distance d from the closest location at time t, that is, the derivation of these expressions is reported in appendix c. the evolution of the fraction of infected agents in eq. (13) depends on four terms: i) the fraction of infected at time t, ii) the fraction of newly recovered, iii) the fraction of newly infected in any location, and iv) the fraction of newly infected outside all the locations. the evolution of the sis model is fully determined by eq. (13), since s(t) = 1 − i(t). for the sir model, instead, eq. (13) should be coupled with the following equation, which describes the evolution of the fraction of recovered agents, and with the conservation constraint s(t) = 1 − r(t) − i(t). the evolution of the fraction of recovered agents only depends on the fraction of recovered at time t and the fraction of newly recovered. in order to gain qualitative insight into the behavior of the sis and sir epidemic processes described by eqs. (13) and (16), we compute the epidemic threshold of both processes by studying the stability of the disease-free equilibrium in eq. (13). we linearize eq. (13) and expand the expressions for the contagion probabilities in eqs. (14) and (15) about the disease-free equilibrium i * = 0, obtaining the epidemic threshold is computed by imposing i(t + 1) ≤ i(t) in eq. (17), obtaining in the case of one location, l = 1, the threshold in eq. (18) reduces to where the last equality is obtained by substituting the explicit terms for q 1 and q out,d from eqs. (11) and (12), respectively, and computing the sum of the obtained series. from inspection of eq. (19), we observe that increasing the probability of jumping outside the location, p, contributes to increasing the epidemic threshold and thus lowers the endemic prevalence and epidemic size. when many locations are present, that is, l → ∞, the second term at the denominator yields a marginal contribution to the epidemic threshold in eq. (19), so that, we observe that the epidemic threshold is now independent from any choice of the probability of jumping outside the location, p. we conclude the analysis of the one-dimensional model by numerically studying the effect of the agents' randomness α and of the probability of jumping outside the location p on the sis endemic prevalence and the sir epidemic size. these numerical simulations extend our analytical predictions, which are limited to the case α = 0. we consider two scenarios, one with l = 1 locations, presented in fig. 6 , and the other with l = 100 locations, illustrated in fig. 7 . our simulations suggest that increasing the agents' randomness α hinders the diffusion of both sis and sir epidemic processes. when only one location is present, increasing the probability of jumping outside the location (that is, shortening the time spent inside the base location) hinders both sis and sir epidemic processes. interestingly, when multiple locations are present, increasing p does not impact the evolution of the epidemic processes. our numerics for α = 0 in figs. 6b,d and 7b ,d indicate the potential use of the analytical expressions in eqs. (13) and (16) for systems of finite size, with n = 10, 000 agents. we consider the two-dimensional agent-based model and numerically study the influence of the randomness α, the probability of jumping outside the base location p, and the presence of a core-periphery structure on the evolution of sis and sir epidemic processes. we start our analysis by exploring the case of a space containing one location, that is, l = 1, fig. 6 influence of the agents' randomness, α, and probability of jumping outside the location, p, on the sis endemic prevalence, a-b, and sir epidemic size, c-d. theoretical values of the sis endemic prevalence, b, and sir epidemic size, d, are computed by evaluating the steady state in eqs. (13) and (16), respectively. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected agents is 0.01. other parameter values are l = 1, n = 10, 000, d = 100, 000, r = 0.0004, c = 0.3333, λ = 0.1, and μ = 0.1 which is the base for all the agents. agents can be either inside or outside their base location. their motion is constrained by the boundary of the location when they are inside it, while it is governed by the parameters α or p when they are outside their base location. our results reveal that increasing either α or p reduces the impact of the epidemic disease, both in the case of possible reinfection (sis), as shown in fig. 8a , and in the case of immunization after recovery (sir), as illustrated in fig. 8b . specifically, in the sis process, the endemic prevalence, i * , is high when α and p are low because agents spend more time inside the location, which is the densest region of the entire space, thus favoring interactions between agents. on the contrary, when agents spend more time outside the location (by increasing either α or p 4 ), they explore a less dense region of the space and interactions become more sporadic. as a result, the likelihood that the disease spreads is lower. from our numerical simulations, we observe that there is a threshold for α (for α close toᾱ = 0.5), beyond which the disease spreading is halted. simulations with different values of the parameters show a similar behavior, with varying values of the threshold α. hence, in the sis dynamics, the disease is not able to spread and the endemic prevalence tends to zero, as shown in fig. 8a ; a similar behavior is observed for the sir process. similar results are obtained for the one-dimensional lattice, as illustrated in fig. 6. fig. 7 influence of the agents' randomness, α, and probability of jumping outside the location, p, on the sis endemic prevalence, a-b, and sir epidemic size, c-d. theoretical values of the sis endemic prevalence, b, and sir epidemic size, d, are computed by evaluating the steady state in eqs. (13) and (16), respectively. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected is 0.01. other parameter values are l = 100, n = 10, 000, d = 100, 000, r = 0.01, c = 0.3333, λ = 0.05, and μ = 0.03 next, we consider the case in which multiple locations are present, forming a coreperiphery structure, as described in appendix a and illustrated in fig. 1b . agents that exit their base location are likely to jump inside another location and interact with other agents occupying a different portion of the urban environment. we investigate a scenario with l = 100 locations, as illustrated in fig. 9 . our numerical results suggest that increasing the agent's randomness, α, still reduces the endemic prevalence (in the sis model) and the epidemic size (in the sir model), i * and r ∞ , similar to the case of a single location. numerical results in fig. 9a and c, however, seem to display a nonmonotonic behavior of the fraction of population affected by the disease, whereby small values of α may favor the epidemic outbreak instead of hindering its inception. we record the existence of a threshold for α (in our simulations, this is close to 0.5) at which a sharp transitions takes place for both the endemic prevalence (in the sis model) and the epidemic size (in the sir model). according to eq. (3), by increasing α, agents' randomness is increased and, as a consequence, agents tend to explore a larger portion of the urban environment and to occupy peripheral locations with a lower density of agents. hence, they become are less likely to interact with each other and support disease spreading. fig. 8 influence of the agents' randomness, α, and the probability of jumping outside the location, p, on the endemic prevalence of the sis model, a-b, and the epidemic size of the sir model, c-d. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected is 0.01. other parameter values are l = 1, n = 10, 000, d = 10 9 , min = 100, max = 10, 000, σ min = 1, σ max = 1, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, λ = 0.15, and μ = 0.1 surprisingly, we observe that the probability of jumping outside the base location, p, seems to have a negligible effect on the outcome of the sis and sir disease processes, similar to predictions from the one-dimensional simplified version of the model in eq. (20) and fig. 7 . a reason for this phenomenon may be found in the following intuition. the core-periphery structure analyzed in our work, illustrated in fig. 1 , allows two contrasting effect to simultaneously occurs. on the one hand, agents moving outside the central areas are likely to end in peripheral ones, decreasing the agents' density in the central regions and increasing the density in the peripheral ones. on the other hand, agents moving outside the peripheral areas are likely to end in the central ones, thereby increasing the density in the central regions and decreasing the density in the peripheral ones. overall, these two opposite effects tend to balance each other. here, we study the impact of the correlation between the radius of interaction of agent i, σ i , and the density of its base location, ρ β(i) . we compare the uncorrelated case (analyzed earlier in fig. 9a and c) , where agents are randomly assigned to locations, with the cases of either positive or negative correlation. in the case of positive correlation, agents with fig. 9 influence of the agents' randomness, α, and the probability of jumping outside the location, p, on the endemic prevalence (sis model), a-b, and the epidemic size (sir model), c-d. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected is 0.01. other parameter values are l = 100, n = 10, 000, d = 10 9 , min = 100, max = 10, 000, σ min = 1, σ max = 1, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, λ = 0.15, and μ = 0.1 larger radius are assigned to denser (and central) locations. in the case of negative correlation, agents with larger radius belong to the less dense (and peripheral) locations. we consider a scenario with l = 100 locations, whose results are illustrated in fig. 10 . both the endemic prevalence, i * , and the epidemic size, r ∞ , are marginally affected by a positive correlation, while they strongly diminish if the radii and density of locations are negatively correlated, as shown in fig. 10a and b, respectively. in both the positivecorrelated and uncorrelated cases, agents with larger radii occupy the central locations, thereby sustaining the diffusion of the disease. on the other hand, if agents with large radii are relegated to peripheral and sparser areas, it would be more difficult for them to create connections and fuel the diffusion process. finally, we examine the effect of different vaccination strategies applied to our population. specifically, we consider a purely randomized strategy and two targeted vaccination policies. in the three strategies, we assume that a fraction of the population is vaccinated and is thus immune to the disease. in the "random" vaccination mechanism, we vaccinate a fraction of the population, sampled uniformly at random. in the "center" targeted mechanism, we select such a fraction starting from the agents assigned to the most central locations. in the "peripheral" targeted mechanism, we choose such a fraction starting from the agents assigned to the most peripheral locations. fig. 10 impact of different ways of assigning agents to their locations on the endemic prevalence (sis model), a, and the epidemic size (sir model), b. the "uncorrelated" case represents a random assignment. in the "pos. correlated" case, agents with larger radii are assigned to the denser (central) locations, while, in the "neg. correlated" case, agents with larger radii belong to the less dense (peripheral) locations. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected is 0.01. other parameter values are l = 100, n = 10, 000, d = 10 9 , min = 100, max = 10, 000, σ min = 1, σ max = 1, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, λ = 0.15, and μ = 0.1 from fig. 11 , we observe that prioritizing the vaccination of agents assigned to the most central locations has the most beneficial effect for the prevention of the diffusion of the epidemic disease, while the worst strategy targets vaccination to peripheral areas. as detailed in fig. 4 , agents assigned to more central base locations tend to have a larger expected degree than agents assigned to more peripheral locations, thereby potentially acting as "superspreaders" (lloyd-smith et al. 2005) . also, agents whose base locations are in the center can easily reach all portions of the environment, thereby contacting the majority of the agents. by focusing the vaccination on central areas, the contacts generated by these agents do not contribute to the spread, thereby significantly reducing the contagion. in this paper, we studied a class of agent-based models (frasca et al. 2006) , in which agents move in a two-dimensional space and interact according to proximity criteria. we extended such class of models by encapsulating a core-periphery structure, typical of urban environments (makse et al. 1998; de nadai et al. 2016) , where central areas are more densely populated than peripheral ones. our urban-like environment is partitioned in several closely spaced locations, each of them representing a restricted portion of the space. when agents are inside their base location, they take a random position within the base location at every time-step. when outside, they tend to move back to their base location by following a biased random walk. the contribution of the study is fourfold. first, we analytically and numerically studied the temporal network formation mechanism, demonstrating that heterogeneously distributed radii of interaction in the population generate heterogeneity in the degree distribution of the temporal network of contacts, similar to what is observed in many realworld systems albert et al. 1999; barrat et al. 2004 ). the role fig. 11 effect of different vaccination strategies on the endemic prevalence (sis model), a-b-c, and epidemic size (sir model), d-e-f. the vaccination coverage represents the fraction of immune agents prior to the disease onset. in "random", we select the fraction of agents to vaccinate at random; in "center", we vaccinate first the agents that are assigned to central base locations, while in "peripheral", we prioritize vaccination for agents that belongs to the peripheral agents. we set: a-d p = 0.1 and α = 0.0, b-e p = 0.4 and α = 0.2, and c-f p = 0.8 and α = 0.4. curves represent the median of 100 independent simulations; 95% confidence bands are displayed in gray. agents are initially inside their base location and the infection starts after 100 steps to allow the agents to reach a steady-state configuration. the fraction of randomly chosen initial infected is 0.01. other parameter values are l = 100, n = 10, 000, d = 10 9 , min = 100, max = 10, 000, σ min = 1, σ max = 1, 000, v = 500, c = 4 · 10 −4 , ω = 2.4, γ = 2.1, λ = 0.15, and μ = 0.1 of the interaction radius is also evident in the study of the clustering coefficient, whereby we found that agents' with larger degree have a lower clustering coefficient. second, we investigated the role of the urban-like environment on the spread of epidemic outbreaks. specifically, we considered epidemic prevalence in the susceptibleinfected-susceptible (sis) model and epidemic size in the susceptible-infectedrecovered (sir) model. we found that both these quantities, which measure the fraction of the system that is affected by the disease, are lowered by increasing the randomness of the agents' law of motion. in fact, increasing agents' randomness improves the chance that agents randomly explore peripheral urban areas, where less agents are present and less contacts are thus created. a lower number of interactions hinders the contagion process. interestingly, we discovered that the endemic prevalence and epidemic size have nontrivial relationships with the probability of jumping outside the base location. when the entire urban environment is modeled as a unique location, larger probabilities of jumping outside hinder the epidemic diffusion. in fact, inside the location the density of agents is higher than outside it. as a consequence, interactions between agents are rare, slowing down the disease spread. instead, when multiple locations are arranged in a core-periphery structure, our numerical results suggest that epidemic prevalence and size are independent of the probability of jumping outside the base location. a possible explanation for this phenomenon might be that, when agents in central locations jump outside them, they are likely to end in peripheral locations, diminishing the fraction of agents in central areas. this event is compensated by agents from peripheral locations that jump in central ones. our numerical results are in agreement with the theoretical findings in the simplified, one-dimensional, version of our agent-based model. third, we found that central locations play a key role on the diffusion of epidemic diseases. in particular, we studied the influence of the correlation between agents' radius and locations' density. when these quantities are negatively correlated, agents with larger radius belong to less dense (peripheral) locations, while when positively correlated, agents with larger radius belong to denser (central) locations. the endemic prevalence (in the sis model) and the epidemic size (in the sir model) are only marginally favored by the presence of many agents with large radius in the more central locations (positive correlation), while the diffusion of the epidemic is hindered if central locations are mostly assigned to agents with small radius of interaction (negative correlation). finally, we studied the effect of targeted vaccination strategies. we found that the vaccination of agents that belong to central locations is the most beneficial approach for the entire population, leading to the smallest epidemic prevalence. our analysis corroborates our previous observation that central (and more dense) locations are crucial in the diffusion of disease processes. we emphasize that the proposed vaccination strategy can be implemented with information about the system at the mesoscopic level of locations, that is, without any information on the specific properties of single individuals (for instance, their radius of interaction). with information at the individual level, the proposed policy may be improved by combining knowledge about locations and radii of interaction prioritizing vaccination of central agents with large radius of interaction, which acts as "superspreaders. " a main limitation of our work resides in the assumption that each agent belongs to a unique location, while the remaining urban area, occupied by other locations, is only seldom explored. a more realistic approach could consider agents that may be assigned to multiple locations. our theoretical study of the one-dimensional case provides insight into some aspects of epidemic processes in urban environments. however, a general mathematical theory is missing. we believe that our preliminary results constitute a starting point for performing a more general theoretical analysis of the two-dimensional model. furthermore, variations of the proposed model can be easily generated. for instance, the gravity law in our model could be replaced by other laws, such as, the well-established radiation law (simini et al. 2012) or the one recently proposed in (schläpfer et al. 2020) . overall, our work determines that central urban areas are critical in the diffusion of epidemic diseases within a city, being the crossroad of most of the urban population, and thus should be carefully included into mathematical models of epidemic outbreaks. by vaccinating individuals in central urban areas, we can halt the overall contagion better than randomly distributing limited vaccination supplies. our proposed vaccination strategy may offer practitioners and epidemiologists general guidelines for emergency situations, complementing other strategies (braha and bar-yam 2006; génois et al. 2015) toward effective containment measures and herd immunity (fine 1993 ) within urban environments. when the system is in its steady state, the expected number of other agents within a distance σ i from agent i is proportional to the ratio between the area of a circle of radius σ i and the whole planar space. hence, the expected number of interactions created by agent i is equal to further, agent i can form undirected interactions with other agents if it is located within their radii of interaction. to avoid double counting and exclude connections that are already counted in eq. (21), the radius of agent j should be greater than the one of agent i, which should be at a distance greater than σ i but smaller than σ j . when the system reaches the steady state, the probability of such an event is π σ 2 j − σ 2 i /d 2 . let us introduce the set c i of agents with radius of interaction greater than σ i and let us define σ 2 i = |c i | −1 j∈c i σ 2 j as their average square radius. the expected number of connections formed by agent i with other agents beyond those included in eq. (21) is by summing eqs. (21) and (22), we conclude that the average number of agents that an agent interacts with in a unit time, termed its average degree k i , is equal to the computation of the quantities |c i | and σ 2 i , can be performed in the limit of large systems n → ∞, by means of the strong law of large numbers (chung and zhong 2001) . we start by explicitly writing the probability density function g(σ ) of the power law distribution of the radii of interaction with cutoffs σ ∈ [σ min , σ max ], as where ω is the exponent. from the expression of g(σ ), we compute |c i | using the strong law of large numbers (chung and zhong 2001) , which ensures that almost surely we define the conditional probability density function where the first equality holds due to scale invariance of the power law distribution, and then explicit computation is performed using the expression of g(σ ). using again the strong law of large numbers (chung and zhong 2001) and eq. (26), we compute σ 2 i as almost surely. finally, by substituting eqs. (25) and (27) in eq. (23), the expected degree of agent i in the limit of large systems, n → ∞, almost surely reads neglecting the terms on smaller order in n. now, we consider the limit case in which agents move straight toward their base location, that is, α = 0, and we assume that locations are uniformly distributed in the planar space. we consider the generic agent i that belongs to location β(i) . since a β(i) d 2 , we use the approximation d → ∞. the probability for this agent to be in its base location, q in , can be computed by introducing the following partition of the planar space, for any h ∈ z ≥0 . note that c (i) h is the region of the plane from which agent i reaches its base location β(i) in exactly h time-steps. consequently, when h = 0 agents are inside their base location, that is, c using the mapping z (i) , for each agent i ∈ v, we define the stochastic process z i (t) : z ≥0 −→ z ≥0 as z i (t) := z (i) (x i (t), y i (t)). since α = 0, when an agent is outside its base location, then its law of motion is purely deterministic and it moves in the direction of the location. therefore, if z i (t) = h = 0, then, z i (t + 1) = h − 1. if z i (t) = 0, the agent is inside its base location, from which it exits only through a jump, which is statistically characterized by eq. (6). hence, with probability 1 − p the process z i (t) remains in state 0 at the following time-step. else, if a jump occurs, the process z i evolves to state h with probability equal to the transition probabilities of z i (t) depend only on the state h in which the process is and on the model parameters. the process z i (t) is a markov chain, whose structure is illustrated in fig. 12 and whose transition matrix is we observe that, if p > 0, then the markov chain is ergodic and it converges to its unique stationary distribution π, which can be computed as the left eigenvector of m associated with the eigenvalue 1 (levin et al. 2006 ). when the system has reached its steady state, the probability for each agent to be inside its base location, q in = π 0 , that is derived from the left eigenvalue equation for m in eq. (32) (with unitary eigenvalue), that is, from eq. (33), the expression of q h in eq. (31), and using that ∞ h=0 π h = 1, we derive q in = π 0 = e cv − 1 (1 + p)e cv − 1 . (34) when the system reaches its stationary state, the number of agents in location is equal to the sum of two contributions. the first one consists of agents whose base location is and are in that location, that is, on average, nq in . the second one is due to agents whose base location is not , but are in . the second contribution is relatively small since locations are placed randomly in the entire space d × d, and we discard it when the system is large. the steady-state density in location can be approximated by considering only the agents assigned to it. hence, the expected number of connections of agent i within its base location is approximated by where c i, and σ 2 i, are the set of agents with radius greater than σ i in location and their average square radius, respectively. assuming the distribution of the radii of interaction to be independent of the agents' base locations, then, c i, = n−1 n c i and σ 2 i, = σ 2 i . under this assumption, in the limit of large systems, n → ∞, combining eqs. (25) and (27) into eq. (35), we obtain e k in,i ≈ q 2 in 2 (n − 1) when a core-periphery structure is present, as in fig. 1 , locations are not uniformly distributed in space and often are close to each other. for instance, a central location is surrounded by other locations and interactions generated by agents whose base location is not cannot be neglected. this case is discussed in the main text by means of numerical simulations. where the first term refers to the probability that the agent is in its base location and its base location is , while the second and third terms correspond to the probability that the agent belongs to another base location and it occupies location . similarly, we compute the probability that agents are in a position not occupied by any location and at a distance d from the closest location as where we assume that the closest location is = 1. through a simple change of variables, we can write an equivalent expression when the closest location is = l. substituting expressions in eqs. (38) and (39) in eqs. (40) and (41) yields the two expressions reported in the main text, that is, eqs. (11) and (12). we now compute the probability that an agent becomes infected at time t. we first consider the probability of not being infected. in location , such a probability is equal to 1 − λi (t) for each contact. on average, an agent contacts q n other agents, the probability of not being infected in location is equal to (t) = (1 − λi (t)) q n . similarly, the probability of not being infected at a distance d from the closest location is equal to out,d = 1 − λi out,d (t) q out,d n . thus, the contagion probability of an agent inside its base location is the complement of (t), that is, internet: diameter of the world-wide web multi-scale spatio-temporal analysis of human mobility urbanisation and infectious diseases in a globalised world emergence of scaling in random networks the architecture of complex weighted networks minimum commuting distance as a spatial characteristic in a non-monocentric urban system: the case of flanders from centrality to temporary fame: dynamic centrality in complex networks daily time spent indoors in german homes-baseline data for the assessment of indoor exposure of german occupants the hidden geometry of complex, network-driven contagion phenomena effects of motion on epidemic spreading local and global epidemic outbreaks in populations moving in inhomogeneous environments solving circle packing problems by global optimization: numerical results and industrial applications modelling transmission and control of the covid-19 pandemic in australia tthe effect of travel restrictions on the spread of the 2019 novel coronavirus (covid-19) outbreak a course in probability theory urbanization in developing countries: current trends, future projections, and key challenges for sustainability epidemic modeling in metapopulation systems with heterogeneous coupling pattern: theory and simulations commentary: middle east respiratory syndrome coronavirus (mers-cov): announcement of the coronavirus study group the death and life of great italian cities: a mobile phone data perspective mitigation measures for pandemic influenza in italy: an individual based model considering different scenarios circle packing for origami design is hard modeling disease outbreaks in realistic urban social networks impact of non-pharmaceutical interventions (npis) to reduce covid19 mortality and healthcare demand herd immunity: history, theory, practice a packing problem with applications to lettering of maps dynamical network model of infective mobile agents synchronization of moving chaotic agents data on face-to-face contacts in an office building suggest a low-cost vaccination strategy based on community linkers agent-based models epidemic spreading in random walkers with heterogeneous interaction radius modeling infectious diseases in humans and animals the minimum circuity frontier and the journey to work contagion dynamics in time-varying metapopulation networks uncovering patterns of inter-urban trip and spatial interaction from social media check-in data superspreading and the effect of individual variation on disease emergence modeling urban growth patterns with correlated percolation exact solution of the two-dimensional finite bin packing problem assessing the interplay between human mobility and mosquito borne diseases in urban environments urban-rural differences in daily time-activity patterns, occupational activity and housing characteristics urban-like environment epidemic spreading in temporal and adaptive networks with static backbone epidemic spreading in modular time-varying networks how urbanization affects the epidemiology of emerging infectious diseases epidemic processes in complex networks an sis epidemic model with vaccination in a dynamical contact network of mobile individuals with heterogeneous spatial constraints activity driven modeling of time varying networks urbanization. our world in data generalizations of the clustering coefficient to weighted complex networks the transition to a predominantly urban world and its underpinnings indoor time-microenvironment-activity patterns in seven regions of europe a universal model for mobility and migration patterns responding to global infectious disease outbreaks: lessons from sars on the role of risk perception, communication and management new approaches to circle packing in a square: with program codes the spread of dengue in an endemic urban milieu -the case of delhi diffusion-limited aggregation, a kinetic critical phenomenon traffic-driven epidemic spreading on networks of mobile agents epidemic spreading in communities with mobile agents continuous-time discrete-distribution theory for activity-driven networks modeling memory effects in activity-driven networks publisher's note springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations received: 14 april 2020 accepted: 3 august 2020 appendix a: algorithm to generate a core-periphery structure from a practical point of view, packing all convex regions l (locations) in the square space d×d is a nondeterministic polynomial-time hardness (np-hard) problem (martello and vigo 1998; demaine et al. 2010) , often requiring to find approximate methods (formann and wagner 1991; szabó et al. 2007; castillo et al. 2008) . in our study, we aim at reproducing the core-periphery structure present in real urban areas, as shown in fig. 1a , while minimizing the space between locations. agents that exit from their base location occupy nearby locations, thereby interacting with agents that are assigned to different regions of the urban area. we developed a heuristic algorithm that unfolds according to the following steps.1 place the center of the denser location, x c 1 , y c 1 , in the center of the square space, x c 1 , y c 1 = (d/2, d/2). 2 initialize ← 1, σ in ← 0, and σ out ← σ = l =1 σ /l. 3 create a circular crown centered in (d/2, d/2) with internal radius σ in and external radius σ out . 4 randomly place the center of location + 1 in the crown and check for overlaps. i) if location + 1 does not overlap with other locations, then the location is placed. increase the index by 1, that is, ← + 1. if = l, then terminate the algorithm. otherwise, resume it to step 4. ii) if an overlap occurs, then repeat the current assignment in 4. after a number of consecutive failed attempts (we set this limit to 100), stop the current iteration and move to step 5.5 setσ in ← σ out and σ out ← σ out + σ , and resume the algorithm to step 3. here, we detail the analytical derivations of eqs. (8) and (9). to this end, we analyze the formation of the temporal network of interactions in the two specific cases of a free space, without any location (l = 0); and when the law of motion of the agents outside their base locations is deterministic (α = 0) and the locations are uniformly distributed in the plane. we begin our analysis by considering the case of a free space, that is, l = 0, where agents perform simple random walks with constant velocity equal to v in the plane. in this scenario, eq.(3) should be intended without the component associated with the location i (t) and with α = 1. according to eq. (7), at time t, agent i creates undirected interactions with other agents if their euclidean distance is less than or equal to the maximum of their radii of interaction. in practice, the expected number of interactions of agent i, e [k i ], is equal to the sum of two contributions: the expected number of interactions that are generated by agent i with agents that are in its radius of interaction, denoted as e k + i , and the expected number of interactions that are generated by other agents with i, denoted as e k − i . here, we compute the contagion probability of an agent inside its base location at time t, (t), and the contagion probability of an agent at distance d from the closest location at time t, out,d (t). we start our analysis by computing the probability that agents are in their base location, denoted by ψ in , or in a position at a distance d from it, ψ d , when the system is at steady state. for p > 0, the system is ergodic and we can compute ψ in and ψ d at steady state (levin et al. 2006 ). similar to appendix b, from the steady-state equation we derive the following recursive system of equations:where the factor 2 is because there are two positions at a distance d from any location ∈ l, as in fig. 5 . from eq. (37), the expression of p jump (d) in eq. (10), and using that ψ in + 2 ∞ d=1 ψ d = 1, we deriveandgiven that each agent is randomly assigned to one of the l locations, the probability that a generic agent i ∈ v is inside location is equal to likewise, the contagion probability when the agent is at a distance d from the closest location is the complement of out,d (t) , that is, a sample of the algorithms generated is available at (nadini 2020) . the entire set of algorithms is available upon request. the authors declare that they have no competing interests. key: cord-342636-mmlnm3mz authors: situngkir, h. title: the pandemics in artificial society: agent-based model to reflect strategies on covid-19 date: 2020-07-29 journal: nan doi: 10.1101/2020.07.27.20162511 sha: doc_id: 342636 cord_uid: mmlnm3mz various social policies and strategies have been deliberated and used within many countries to handle the covid-19 pandemic. some of those basic ideas are strongly related to the understanding of human social interactions and the nature of disease transmission and spread. in this paper, we present an agentbased approach to model epidemiological phenomena as well as the interventions upon it. we elaborate on micro-social structures such as social-psychological factors and distributed ruling behaviors to grow an artificial society where the interactions among agents may exhibit the spreading of the virus. capturing policies and strategies during the pandemic, four types of intervention are also applied in society. emerged macro-properties of epidemics are delivered from sets of simulations, lead to comparisons between each effectivity of the policy/strategy. when it comes to coronavirus disease pandemic, the world witnesses heterogeneous ways of managing. since the beginning of the outbreak in many places, different ways of managing the virus spreadings reflect the various healthcare situations faced by each government in the world, due to the testing and hospitalizations facilities. the situations are getting more complex for the unique characteristics of the covid-19 itself. as illustrated in figure 1 , the fatality rate and the contagiousness of the covid-19 are in between many other known diseases. the fact that the infected persons have some degree of possibilities to be asymptomatic, as well as the recognized pre-symptomatic cases, made a direct impact on the social life on which the epidemic occurs. most infected people have mild symptoms and still be able to deliver their daily social activities in which they can easily spreading the virus rapidly in the population. thus, policies of those limiting social interactions for the whole population are the most suggested solutions epidemiologically, as well as closing borders between places (countries, provinces, districts, etc.) to avoid imported/exported cases between regions. the consequences of the policies would apparently give a direct impact on many social aspects since interactions are fundamental elements in social life (smelser, 1997) . modern human life is formed in the patterns of social interactions, in which all social aspects are embedded, including religions, customs, economy, politics, and all. thus, the epidemiological suggestions for the policies' pro-social limitations and closing administrative borders are being resisted by the need of economical aspects of the social life. people are needed to be supported economically while on the other hand should be stay put with distant social interactions for the sake of damping the easily doubling trends of the infections. one popular policy is the "lock-down" within areas, where people are encouraged (and restricted) to "stay at home" (who, 2020) . in contrast, some other countries enforce other kinds of the policy of not applying the "lock-down" at all, like in sweden (dahlber, et. al., 2020) with very mild restrictions due to social activities within the country. many aspects happened to be put into account when it comes to social restrictions, from the reasons related to human rights to the economic aspects due to processes within the economy, macroeconomic reserve and workforces, and a lot more. furthermore, the considerations backing the varsities of enforced policies and strategies facing the covid-19 may also be coming from the heterogeneous natural aspects of countries around the world. for instance, the interactive effects based on meteorological influences in the covid-19 transmission and spread, due to the interactive effect between daily temperature and relative humidity on covid-19 incidence (qi, et. al., 2020) . some countries in tropical areas may have a milder effect within their infection cases relative to those in sub-tropical regions. the urge for maximum social restrictions in warmer tropical areas may not as great as those in other colder places. the exhibited variations are there for different existing situations faced countries in battling the pandemics. the paper is preparing the toy model reflecting the social structure that in return presenting a proposal for enriching observations toward various strategies within many places due to the pandemic of covid-19. while other previous works may present the computational models and simulations to approach the dynamics of an epidemic outbreak (cf. , the coverage of the paper is focused on the preparations of artificial societies in which some policies may be grown (epstein & axtell, 1996) and explained. the model used tried to capture the micro-social structure on which the infected people came along. the people are interacting with one another with particular individual motives and move around the artificial world as lattices and grids (cf. rhodes & anderson, 1996) . as an infected agent come along, the interactions exhibit the spreading of the virus in a sort of social network based on their bounded situations (cf. newman, 2002) . the model runs as an agent-based model (gilbert & terna, 2000) living the landscape that we can use to monitor the macro-properties of the epidemics (cf. situngkir, 2003) . the observations are delivered in the emerged aspects of epidemics, i.e.: the number of people infected and how it spreads over the landscape as the complex adaptive system (miller, 2007) . if the behavior of the agents is related to the micro-level of description, the government policies can be seen as in the mesoone (medium level, between micro and macro). all of those aspects are then depicting the relations of "factors to actors" relations (macy & willer, 2002) . some possible conjectures in the ways to verify the results are also discussed. eventually, the paper is demonstrated how interdisciplinary works can enhance social policies handling pandemics (angulo, 1987) . we can see the micro-social used in the simulations as three parts, i.e.: the internal state of the agents, the mobility in our artificial world, and the spreading of the disease based on the first two properties. agent has internal state ruled by the social-psych-wellness index, denoted as ! ( ) ∈ ℜ, ! ( ) = [0.1]. this index represents the wellness of agents regarding many complex aspects in social life and it is composed of three factors, two of them are related directly to its corresponding spatial situations. the first one is denoted as ! ( , , ), how agent fulfill her necessities when she nears a static point, we can say this as attraction points located in 2-dimensional , . one can imagine these points as public surroundings where people meet and have their needs from others, e.g.: marketplace, recreational spaces, offices. index ! ( , , ) is determined by the euclidean distance to the nearest attraction point, adjusted by parameter "##$"%#!&' . however, the relation is inversely proportional so that the nearer the agent to her chosen point, the higher ! ( , , ). its value is ! ( , , ) ∈ ℜ = [0,1] by a negative exponential function, decaying with term − 7 () , () '*"$*+# !"#$% 9. the second aspect is the agent neighborhood ! ( , ) which is simply a fraction of the number of her neighbors divided by a global constant maximum neighbors m. the factor corresponds to understanding that human tends to reach out others for their individual needs, be it economic, cultural, and social well-being. the conceptualizations of the neighborhood are other agents as perceived within radii '*!,-.&$+ , compared to her distance to others. this ! ( , ) ∈ ℜ = [0.1], weighted by a parameter as a feedback of the growing index of social-psycho-wellness, there is a boredom factor with proportions to agent 's, − ! ( ). the higher the value of her gladness, the more this factor reducing ! ( ). the three factors which determine the agent's internal state can formally be stated by the following system of equations: our artificial world is 2-dimensional lattice and grids of which form torus in 3 dimensionalities since the edge of the left is glued with the right one, and the top with the bottom edge as well. each agent spatial states are represented as a vector of position () ! and her velocity () ! . initially, agents are randomly located uniformly with zero speed of movement. following the terms of classical mechanics, in every iteration the spatial states of agents are updated by applying an amount of normalized steering force to agent : whereby default, each value of 1 mmm⃗ components is taken from an unbiased uniform distribution and the mass is assumed to be unity. furthermore, some conditional force is computationally adapted based upon the agent internal state, ! ( ). thus, the agent's mobility is ruled by her micro-motives: maximizing the wellness index. in short, the dynamics adopt boids model (reynold, 1987) which follows some simple rules distributed among the agents, the separation and cohesion-like rule. only if ! ( ) is below the threshold value 2&3 , we replace || $"'/&0 1 mmmmmmmmmmmmmmmm⃗ || to || +**41', 1 mmmmmmmmmmmmmmm⃗ ||. this force will make agents move towards their respective neighbors' average position and their corresponding nearest attraction points. however, dynamically, the neighboring factor is set to be zero when her number of neighbors m exceeds maximum limit forbidding the overcrowded localities. this also happened when an agent has no neighboring agents. in case no attraction point that they can perceive, or no neighboring force, they will perform a random walk by default. where denotes the value of weighting factor times normalization factor for each force. another rule is the separation rule which is necessary for later when we do the simulations. this rule makes agent keep their distance from each other by applying force: the equation corresponds to +*8"$"#!&' value which determines how an agent takes into account their neighbors perceived as too close, 7 () ! , () ! #&&_%2&+* 9 < +*8"$"#!&' . the denominator gives proportionality to the applied force so that the nearer her neighbor, the stronger she will stay away. to summarize, the change of agent velocity can be restated as: from here on, the simulation can be demonstrated and ready for some further epidemiological features. figure 2. the screen of our artificial societies with the control panel for adjustments of the experiments we can deliver. the dots represent moving agents, while the red dots are infected, the green one represents the recovered agents, the white one is susceptible agents. the dark blue is the center of attraction in which agents fulfill their utmost social and psychological wellness. when it comes to the simulated epidemics, agents can have an epidemiological state, be it susceptible, infectious, recovered, or dead, in a mutually exclusive manner. the state will be updated on each iteration according to some rules. as discussed previously, the coronavirus disease (covid-19) is highly contagious from human to human. via the droplets due to respiratory process, when people are close enough to talk or touch one another, or one human gets from droplets staying on surfaces as touched by hands as one-touch eyes, nose, or mouth. physical distancing, sanitizing hands, and preventing touching faces are the campaign for people to resist the contagions. these aspects are the core thing to be simulated in our artificial societies. firstly we have a population with all susceptible agents, except the patient zero. the only susceptible agent can be infected due to her spatial interaction with other infectious agents, occurs with probability !'9*%#*/ and only examined if the distance between the two agents is below a certain constant of !'9*%#!&' . after the $*+&2:#!&' -th iteration (distributed normally for all agents), infectious agent can be recovered with probability $*%&;*$) , otherwise, she is dead and removed from the artificial world. in return, recovered agents are still susceptible again with a particular probability +:+%*8#!.2* after +:+%*8#!.2* -th iteration (also distributed normally). countries around the world manage the covid-19 pandemic in various ways deliberating all aspects and social dimensions. the basic idea of strategy against covid-19 in the absence of any pharmaceutical intervention is reducing the possibility of susceptible and infectious mix one another. this is brought by early ascertainment of cases or reduction of contact. "lockdown" is one of the most heard solutions since it was the practice in china during the first outbreak epidemic (lau et.al., 2020) . however, the basic idea for reducing the mixing of susceptible people with the infected one can be interpreted as applied to various enforced policies. the campaign of social distancing is one of them: people are given the understanding to be in the distance when around other people (lewnward, et. al., 2020) . this is including the order of not letting people be in gatherings in not more than a few. in most impacted regions and countries, the public spaces closures, e.g.: restaurant, recreational park, entertainment venues, market, school are forbidden to operate regularly. a late addition by the world health organization's advice as a suggested intervention for the pandemic is universal face mask use. there are some rationales of the use of face masking as effective personal protective measures in the era of pandemic (sunjaya & jenkins, 2020) . social distancing in a social distancing scenario, all agents are expected to keep their distance from each other to reduce the disease transmissibility. this can be applied by adjusting the dseparation value then see if the agents perform separation rule. if the separating distance among agents is greater than infection distance, no infection will occur in the probability of (1-pseparate), the fseparation is set to zero vector. the mask-wearing scenario is simulated by multiplying the probability pinfected with mask reducing factor k = [0.1, 1]. in the basic model, one can say that k value is equal to 1 so that it will not affect the occurrence of infections. public spaces closures this scenario is implemented by removing all the attraction points that previously exist in the basic model. the attraction force for each agent then automatically turns to zero vector. it's worth noting that this also affects the wellness index of agents where the distance between the agent and her nearest attraction point becomes infinity. for the lockdown scenario, agents are redistributed to the initial vector in their "homes" then within the "home" uniform grids sized as lgrid x wgrid where each has centroid rgrid. once in her "home" position, the agent is attached to one of those grids in which the distance d(r, rgrid) is minimum. whenever the agent wants to leave her grid, she will be pushed back by the wall so that she can't interact with other agents in another grid. in this scenario, all public spaces are omitted from the landscape. growing the epidemic in silico would be valuable in observing how some of those intervention strategies enforced in many countries. four categorical types of interventions that can be observed in the artificial societies are social distancing encouragement, mask-wearing campaign, public spaces closures, and lockdown as presented in table 1. in its actuality, the four categorical types can be used as interventions in many places, countries, around the world in the existing combination as well. in our artificial society, without the intervention at all, the dynamics of the epidemic rely only on the herd immunity formations within the population. however the aspects of the contagiousness of the covid-19 would likely in demand of quite a long time to reach it since most fraction of the population should be infected to get there (clemente-suárez, et. al., 2020) . there is also the risk of an unacceptable number of death tolls without any intervention. within the simulation, we check the popular "lockdown" intervention. the effective lockdown is impressively suppressed the number of infection rate. there are some social (and psychological) effects, nonetheless to the population in the period of locking down the population, the agent's social and psychological index. this index can somehow represent the situation of social and economic aspects of the intervention since within the interval time people cannot do the social and activities at all effectively. at the end of the day, locking-down the population needs accurate momentum along with social and economic aspects of individual and community life, especially when it comes to the timing of the reopening. the too-early reopening comes to the risk of the second wave. the world witnesses this situation in some countries of which reopening too soon of the measure. in order to suppress the second wave, another lock-down intervention should be enforced. in some places around the world, there are some cities and countries that need to do such measures. when it comes to closing down the public spaces (in the simulation we omit the social attraction points) and encouraging the effective physical distancing measures to the population, the number infection rate is suppressed a little. but since the agents are still able to get along wandering within our landscape, the infections are still there. people are still getting infected and in the long run, the active cases are merely slowing down in a big period of time until the definitely possible herd-immunity comes along. this is shown in figure 5 (left). as discussed earlier, the who later gives recommendations on the use (cloth) mask as personal protection in the era of a pandemic. as we simulated the usage of masks in our agent-based model, the slowing rate of infection does give effect even though it needs time to suppress the number of active cases. it is worth noting that the usage of masks in our simulation in figure 5 (right) is delivered exclusively without any other intervention. the effectiveness of wearing masks with a combination of physical distancing encouragement and closure of public spaces for a period of time is simulated with the interesting results as shown in figure 6 . the effectiveness of this combination, relative to the effective lockdown is the interval time for the cases eventually significantly decreased. however, since agents are still given the opportunity with their social and economic life, the things average social and psychological index is not as drop like the one with full effective lockdown intervention observed in the previous experiment. this may explain with the cases in south korea, taiwan, and some other eastern asia, where people are encouraged strongly to wear masks even after the lockdown phase and the cases have been decreased (lee & you, 2020) . thus from our sets of experiments in the agent-based simulation, some tweaks of interventions due to the pandemic at the micro-level, the emerged macro-level is observed, including some emerged social aspects. there is no single solution of intervention when it comes to complex social systems, including due to the policy harnessing the pandemic. the computational simulation, whatsoever, is open for modifications and other changes due to many other aspects to be included in the intervention. our simulation explains the varsities of governance and social and economic policies applied in different countries, regions, and areas. the agent-based model of the complex social system to observe some aspects of intervention due to the covid-19 pandemic can give insights on the cause and emergence of intervention on trying to handle the pandemics in the absence of a vaccine. some interventions potentially hurt the social and economy of the people while "flattening" the exponential rise of infections. on the other hand, no intervention can bring people with the risk of unacceptable death while naturally, the eco-social system adjusts itself for the collective immunity. on the other hand, the characteristics of the disease and virus are not the same either for different regions, areas, and countries. many aspects, not necessarily the social and economic one per se, should be put into account when it comes to policies. computational simulation, by growing deliberately the structure of the social system into computation, provides the artificial societies in which many different observed aspects and characteristics of the disease can be applied (by tweaking variables in either micro and larger (macro) levels of description). our simulation may explain and computationally demonstrated the various pathways of intervention delivered by the governments in different countries and regions. initial number of agents n 300 initial agents positions rx, ry u(0, 700), u(0, 500) initial g(0) n(0.5, 0.125) !""#!$"%&' interdisciplinary approaches in epidemic studies -ii: four geographic models of the flow of contagious disease dynamics of population immunity due to the herd effect in the covid-19 pandemic:. vaccines effects of the covid-19 pandemic on population mobility under mild policies: causal evidence from sweden growing artificial societies: social sciences from the bottom up how to build and use agent-based models in social science the positive impact of lockdown in wuhan on containing the covid-19 outbreak in china psychological and behavioral responses in south korea during the early stages of coronavirus disease 2019 (covid-19 scientific and ethical basis for social-distancing interventions against covid-19. the lancet. infectious diseases from factors to actors: computational sociology and agent based modeling complex adaptive systems: an introduction to computational models of social life the spread of epidemic disease on networks dynamics in a lattice epidemic model flocks, herds and schools: a distributed behavioral model. siggraph '87: proceedings of the 14th annual conference on computer graphics and interactive techniques moneyscape: a generic agent-based model of corruption how far can we go through social system epidemiology through cellular automata: case of study avian influenza in indonesia problematics of sociology: the georg simmel lectures rationale for universal face masks in public against covid-19 covid-19 transmission in mainland china is associated with temperature and humidity: a time-series analysis coronavirus disease (covid-19) advice for the public we thank colleagues in bandung fe institute for discussions in the early draft of the manuscript. key: cord-353297-jizitnfl authors: meyer, r.f.; morse, s.a. title: viruses and bioterrorism date: 2008-07-30 journal: encyclopedia of virology doi: 10.1016/b978-012374410-4.00549-5 sha: doc_id: 353297 cord_uid: jizitnfl the use of viral agents for biological warfare has a long history, which predates their recognition and isolation by culture. advances in viral culture and virus stabilization made during the second half of the twentieth century raised the level of concern by facilitating the large-scale production of viral agents for aerosol dissemination. furthermore, the nucleic acid of many viruses, including some that are currently not threats, can be manipulated in the laboratory. thus, the potential for genetic engineering and misuse of biotechnology is a serious threat. an effective defense against viral agents requires a comprehensive approach including restricting access to viral stocks, detecting deliberately induced disease outbreaks, rapid laboratory identification of viral agents in clinical specimens, preventing person-to-person transmission, using reliable decontamination procedures, and developing effective vaccines and antiviral drugs. between virus names and species names would be to change the current names of virus species into nonlatinized binomial names. such a system, which has been advocated by plant virologists for many years, consists in replacing the word 'virus' appearing at the end of the existing species name by the genus name which also ends in '-virus'. measles virus then becomes measles morbillivirus, hepatitis a virus becomes hepatitis a hepatovirus, and tobacco mosaic virus becomes tobacco mosaic tobamovirus. the advantage of such a system, which could be implemented without problems for about 98% of all virus species names, is that inclusion of the genus name in the species name provides additional information about the properties of the virus. a changeover to binomial species names would not affect the common names of viruses in english or other languages since names such as measles virus or 'virus de la rougeole' would remain the same. the ictv is currently debating the possibility of introducing binomial species names and some decision is likely to be made in the near future. given that the common names of viruses are used repeatedly in scientific texts there is a need for abbreviating them and the ictv has published several lists of recommended acronyms for virus names. since the names of virus species are used only very seldom in publications, there is no need to abbreviate them. if binomial names of virus species were introduced in the future, the abbreviations of common names of viruses will of course not be affected. man has known that biological organisms and toxins were useful as weapons of war long before the germ theory of disease was understood. however, as the twentieth century came to a close, the perceived difficulties in production, weaponization, and deployment of these biological weapons as well as a belief that moral restraints would preclude the use of these weapons gave many a false sense of security. recently, a number of events have served to focus attention on the threat of terrorism and the potential for the use of biological, chemical, or nuclear weapons against the military, civilian populations, or agriculture for the purpose of causing illness, death, or economic loss. this possibility became a reality in october 2001 when someone sent spores of bacillus anthracis to media companies in new york city and boca raton, florida, resulting in five deaths, considerable panic throughout the united states and other countries, and raised the awareness of our vulnerability. there are more than 1400 species of infectious organisms that are known to be pathogenic for humans; many additional organisms are capable of causing disease in animals or plants. realistically, only a few of these infectious agents pose serious problems or are capable of affecting human, animal, or plant health on a large scale. even fewer of these agents are viruses. viruses that could be used as weapons against humans, animals, or plants generally possess traits including ease of production and dissemination, transmissibility, environmental stability, and high morbidity and mortality rates. the use of biological agents is often characterized by the manner in which they are used. for the purposes of this article 'biological warfare' is defined as a special type of warfare conducted by a government against a target; 'bioterrorism' is defined as the threat or use of a biological agent (or toxin) against humans, animals, or plants by individuals or groups motivated by political, religious, ecological, or other ideological objectives. furthermore, terrorists can be distinguished from other types of criminals by their motivation and objective; however, criminals may also be driven by psychological pathologies and may use biological agents. when criminals use biological agents for murder, extortion, or revenge, it is called a 'biocrime'. the use of viral agents for biological warfare has a long history, which predates their recognition and isolation by culture. their early use is consistent with what, at the time, was known about infectious diseases, particularly smallpox. in the sixteenth century, the spanish explorer, francisco pizarro, presented the indigenous peoples of south america with variola-contaminated clothing, which resulted in widespread epidemics of smallpox. during the french and indian war (1745-67), sir jeffrey amherst, commander of the british forces in north america, suggested the deliberate use of smallpox to 'reduce' native american tribes hostile to the british. captain ecuyer (one of amherst's subordinates), fearing an attack on ft. pitt from native americans, acquired two variola-contaminated blankets and a handkerchief from a smallpox hospital and, in a gesture of good will, distributed them to the native americans. as a result, several outbreaks of smallpox occurred in various tribes in the ohio river valley. in 1775, during the revolutionary war, the british attempted to spread smallpox among the continental forces by inoculating (variolation) civilians fleeing boston. in the south, there is evidence that the british were going to distribute slaves who had escaped during hostilities, and were sick with smallpox, back to the rebel plantations in order to spread the disease. the use of viruses other than variola major is a more recent phenomenon and reflects an increased knowledge of how to grow and stabilize viruses for delivery purposes. allegations have been made by the government of cuba that the cia was responsible for the massive outbreaks of swine fever in 1971 and dengue fever in 1980 that ravaged the country. however, subsequent investigations have failed to find substantive proof of cia involvement in these outbreaks. the aum shinrikyo, a religious cult responsible for the 1995 release of sarin gas in the tokyo subway system, was also involved in biological warfare activity and sent a team of 40 people to zaire to acquire ebola virus. fortunately, they were unsuccessful in this endeavor. in 1997, unknown farmers in new zealand deliberately and illegally introduced rabbit hemorrhagic disease virus (a calicivirus) onto the south island as an animal control tool to kill feral rabbits. over the past two decades, the human immunodeficiency virus (hiv) has been involved in a number of biocrimes. this most likely reflects the availability of hivcontaminated blood as a source of this virus. for example, in 1990, graham farlow, an asymptomatic hiv-positive inmate at a prison in new south wales, australia, injected a guard with hiv-contaminated blood. the guard became infected with hiv; farlow subsequently died of aids. in 1992, brian t. stewart, a phlebotomist at a st. louis, mo hospital, injected his 11-month-old son with hivcontaminated blood during a fight over payment of child support. in 1993, iwan e. injected his former girlfriend with 2.5 ml of hiv-contaminated blood after she broke up with him. in 1994, dr. richard j. schmidt, a married louisiana gastroenterologist, injected a former lover with hivcontaminated blood. molecular typing of the hiv strains demonstrated that she contracted the same strain of hiv as found in one of dr. schmidt's patients. in perhaps the most famous case, dr. david acer, a florida dentist infected with hiv, transmitted the disease to six of his patients between 1987 and 1990. the intentional infection of these patients is a possibility although there is no direct evidence. in spite of these incidents, hiv has not been included on lists of threat agents for public health bioterrorism preparedness. however, some contend that hiv has great weapon potential if the goal is to destabilize a society. viruses have also been involved in suspected incidents or hoaxes. in 1999, an article appeared suggesting that the cia was investigating whether iraq was responsible for causing the outbreak of west nile fever in the new york city area. the story relied heavily on a previous story written by an iraqi defector, claiming that saddam hussein planned to use west nile virus strain sv 1417 to mount an attack. the investigation indicated that there was no known evidence of bioterrorism involved in the spread of west nile virus. a fictional 'virus' was also involved in one of the largest bioterrorism hoaxes in 2000. according to e-mail messages widely circulated on the internet, an organization known as the klingerman foundation was mailing blue envelopes containing sponges contaminated with a fictional pathogen called the 'klingerman virus'. according to the e-mail alert, 23 people had been infected with the virus, including 7 who died. advances in viral culture and virus stabilization made during the second half of the twentieth century facilitated large-scale production of viral agents for aerosol dissemination. a report for the united nations on chemical and biological weapons and the effects of their possible use gave estimates on the numbers of casualties produced by a hypothetical biological attack (table 1) . three viruses (rift valley fever virus, tick-borne encephalitis virus, and venezuelan equine encephalomyelitis (vee) virus) were evaluated in a scenario in which 50 kg of the agent was released by aircraft along a 2 km line upwind of a population center of 500 000. the viral agents produced fewer casualties and impacted a smaller area when compared with the bacterial agents used in this hypothetical model. of note, smallpox was apparently not evaluated because it had not yet been eradicated and level of vaccine-induced immunity in the population was high. viral agents were part of the biological weapons arsenal of both the soviet union and the united states ( table 2) . vee virus was stockpiled by both countries as an incapacitating agent; variola major and marburg viruses were stockpiled as lethal agents by the soviet union. the soviet union reportedly conducted a live field test of variola major virus on vozrozhdeniye island in the aral sea in the 1970s, in which 400 g of the virus was released into the atmosphere by explosion. unfortunately, a laboratory technician who was collecting plankton samples from an oceanographic research vessel 15 km from the island became infected. it was reported that after returning home to aralsk, she transmitted the infection to several people including children. all those infected died. a number of other viruses that infect humans (e.g., ebola virus, lassa fever virus, enterovirus 70) or livestock (e.g., foot and mouth disease virus, rinderpest, newcastle disease virus) have also been studied for their offensive capabilities or for the development of medical and veterinary countermeasures. today, with the increased level of concern, a number of viruses have been cited as possible weapons for use against humans or animals ( table 2 ). the requirements for an ideal biological warfare agent include availability, ease of production, stability after production, a susceptible population, absence of specific treatment, ability to incapacitate or kill the host, appropriate particle size in aerosol so that the virus can be carried long distances by prevailing winds and inhaled deeply into the lungs of unsuspecting victims, ability to be disseminated via food or water, and the availability of a vaccine to protect certain groups. other factors such as the economic and psychological impact of an attack on animal agriculture with a viral agent must also be considered. variola major is considered to be the major viral threat agent for humans. thus, considerable effort has been expended toward preparing the public health and medical communities for the possibility that this agent will be employed by a terrorist. variola major is considered to be an ideal terrorist weapon because it is highly transmissible by the aerosol route from infected to susceptible persons; the civilian populations of most countries contain a high proportion of susceptible persons; the disease is associated with a high morbidity and about 30% mortality; initially, the diagnosis of a disease that has not been seen for almost 30 years would be difficult; and, other than the vaccine, which may be effective in the first few days post infection, there is no proven treatment available. alphaviruses ( table 2 ) are also of concern because they can be produced in large amounts in inexpensive and unsophisticated systems; they are relatively stable and highly infectious for humans as aerosols, and strains are available that produce incapacitating (e.g., vee) or lethal infections (eee case fatality rates range from 50-75%). furthermore, the existence of multiple serotypes of vee and eee viruses, as well as the inherent difficulties of inducing efficient mucosal immunity, make defensive vaccine development difficult. the filoviruses and arenaviruses that cause hemorrhagic fever have also been considered as agents that might be used by terrorists because of their high virulence and capacity for causing fear and anxiety. the filoviruses, ebola and marburg, can also be highly infectious by the airborne route. humans are generally susceptible to infection with these viruses with fatality rates greater than 80%, and infection can be transmitted between humans through direct contact with virus-containing body fluids. there are five species of arenaviruses (lassa fever, junin, machupo, guanarito, and sabia) that can cause viral hemorrhagic fevers with a case fatality rate of about 20%. large quantities of these viruses can be produced by propagation in cell culture. infection occurs via the respiratory pathway suggesting that dissemination via aerosol might be used by a terrorist. human to human transmission has also been reported with aerosol transmission the most likely route for at least some of the secondary cases. the filoviruses and arenaviruses discussed above are bsl-4 bacillus anthracis >20 95 000 125 000 note. these estimates are based on the following scenario: release of 50 kg of agent by aircraft along a 2 km line upwind of a population center of 500 000. agents and diagnostic capacities for infections caused by these viruses are limited. because the nucleic acid of many viruses, including some that are currently not threats, can be manipulated in the laboratory, the potential for genetic engineering remains a serious threat. biotechnology, which has had a tremendous impact on the development of medicines, vaccines, and in the technologies needed to counter the threat of naturally occurring disease, can also be used to modify viruses with unintended consequences or even for the development of novel biological agents. several examples involving viruses are presented below. an australian research group was investigating virally vectored immunocontraceptive vaccines based on ectromelia virus, the causative agent of the disease termed mousepox. they created a recombinant virus, which expressed the mouse cytokine il-4 in order to enhance the antibodymediated response to other recombinant antigens carried on the virus vector. instead, the ectromelia virus vector expressing il-4 altered the host's immune response to this virus resulting in lethal infections in normally genetically classification of viral agents that are considered to be of concern for bioterrorism and biowarfare and those that have been weaponized or studied for offensive or defensive purposes as part of former or current national biological weapons programs resistant mice (e.g., c57bl/6). additionally, this virus also caused lethal infections in mice previously immunized against infection with ectromelia virus. the creation of this 'supermousepox' virus led to speculation that similar genetic engineering could be performed on variola major leading to a biological weapon that would be effective against an immunized population. the influenza pandemic of 1918-19, which followed world war i, was uniquely severe, causing an estimated 20-40 million deaths globally. this pandemic happened before the advent of viral culture and very little was known about the virus until the discovery of the polymerase chain reaction (pcr). recently, the complete coding sequences of all eight viral rna segments has been determined by using reverse transcription-pcr (rt-pcr) to amplify the viral rna sequences from formalin-fixed and frozen tissue samples from individuals who died during this pandemic in an effort to shed light on both the reasons for its extraordinary virulence and evolutionary origin. more recently, researchers reconstructed the 1918 spanish influenza pandemic virus using reverse genetics and observed that this reconstructed virus exhibited exceptional virulence in the model systems examined and that the 1918 hemaglutinin and polymerase genes were essential for optimal virulence. a full-length poliovirus complementary dna (cdna) (c. 7500 bp) has been synthesized in the laboratory by assembling oligonucleotides of plus-and minus-strand polarity. the synthetic poliovirus cdna was transcribed by rna polymerase into viral rna, which translated and replicated in a cytoplasmic extract of uninfected hela s3 cells, resulting in the de novo synthesis of infectious poliovirus. the publication of this research raised concerns that more complicated viruses (e.g., variola major or ebola) could be synthesized from scratch based on publicly available sequences, or that viruses could be created that do not exist in the wild. an effective defense requires a comprehensive approach that includes: prevention of access to viral stocks; improved means of detecting deliberately induced disease outbreaks; rapid medical recognition of specific syndromes (e.g., hemorrhagic fever syndrome); rapid laboratory identification of viruses in patient specimens; prevention of person-person transmission; reliable decontamination procedures; development of effective vaccines; and development of effective antiviral therapy. rapid and accurate detection of biological threat agents is the basis of an effective public health response to bioterrorism. in order to address this issue, cdc in collaboration with other partners established a national network of laboratories called the laboratory response network (lrn), which was provided with the tools to accomplish this mission. rapid assays utilizing advanced molecular and immunological technologies for detection of agents such as variola virus, as well as emerging public health threats such as sars coronavirus and h5n1 influenza virus, were distributed to member laboratories. equipment, training, and proficiency testing are elements of the lrn and contribute to a uniform operational plan. the importance of high-quality standardized testing for detection of these agents is exemplified by the rapid need for medical countermeasures to protect or treat civilian populations. accurate laboratory analysis is a major element in the decision process for deployment of the federal government's strategic national stockpile (sns) of medical countermeasures. as part of the effort to deter biological terrorism and strengthen the law enforcement response to such an act, the us recently established a microbial forensic laboratory known as the national bioforensics analysis center that operates in partnership with the federal bureau of investigation. scientists are already developing methods for the forensic investigation of incidents involving viruses. for the terrorist, the use of a viral agent would pose a challenge due to problems associated with acquisition, cultivation, and dissemination. the target for an attack with a viral agent can range from humans to animals and plants. therefore, agricultural targets are also a major concern. nature has provided many challenges to combating viral diseases. viral agents are much more prone to genetic variation and mutation, and can be manipulated or created in the laboratory to take on desired characteristics. differentiating between natural and intentional viral disease outbreaks can be challenging. unlike bacterial diseases, many of which are treatable, there are fewer medical countermeasures to employ when dealing with viral infections. laboratory diagnostic methods and reagents must continuously be refined to account for genetic changes and variants. thus, the challenge of developing bioterrorism countermeasures is significant. fortunately, this effort contributes to combating natural disease events more effectively, which has global benefits. see also: aids: disease manifestation; aids: global epidemiology; aids: vaccine development. defense against filoviruses used as biological weapons microbial forensics bioterrorism and biocrimes. the illicit use of biological agents since 1900 chemical synthesis of poliovirus cdna: generation of infectious virus in the absence of natural template arena viruses other than lassa virus did bioweapons test cause a deadly smallpox outbreak? smallpox as a biological weapon princes and peasants. smallpox in history expression of mouse interleukin-4 by a recombinant ectromelia virus suppresses cytolytic lymphocyte responses and overcomes genetic resistance to mousepox origin of the west nile virus responsible for an outbreak of encephalitis in the northeastern united states molecular evidence of hiv-1 transmission in a criminal case public health assessment of potential biological terrorism agents viruses of the bunya-and togaviridae families: potential as bioterrorism agents and means of control characterization of the reconstructed 1918 spanish influenza pandemic virus world health organization(1970) health aspects of chemical and biological weapons alkaliphilic having a requirement for an environment with a high ph. burst size the number of infectious virus particles released per cell. carrier state persistent infection of a host cell by a virus, with the surviving host persistently carrying and continually producing virus without entering a lysogenic state. circular permutation a change in the sequence of the linear dna termini that does not alter the relative sequence (e.g., circular permutation of abcdefgh could generate bcdefgha, cdefghab, etc). concatamer two or more dna molecules that are linked together to form a long, linear dna molecule. cured a host cell that was once a lysogen, but no longer carries viral dna in any form. halophilic having a requirement for an environment with a high salt concentration. headful packaging the mechanism of packaging viral dna based on the size of the virus head, rather than the length of the viral genome. hyperthermophile having a requirement for an environment with a high temperature (>80 c).insertion sequences repetitive sequences of dna that can move from one site to another within the viral dna. integrase/recombinase an enzyme which can integrate viral dna into the genome of its host cell. lysogen a host cell that has been infected by a virus that remains dormant, despite the presence of viral dna. lytic virus a virus that is able to infect a host, replicate, and subsequently leave the host cell by rupturing (lysing) the host cell. methanogenic having the ability to produce methane. monovalent a virus that has a host range limited to one species. prophage a virus that is dormant within the host cell. protein-primed replication replication of dna via the interaction of the dna polymerase with specific proteins, rather than dna or rna primers. temperate virus a virus that is able to infect a host, but remain dormant within the host cell. terminal redundancy linear dna with the same sequence at each end. transduction the transfer of host dna from one host cell to another by a virus. transfection the introduction of pure viral genomic dna into a host cell, producing viable virus. key: cord-297287-0i4nc353 authors: braun, benjamin; taraktaş, başak; beckage, brian; molofsky, jane title: simulating phase transitions and control measures for network epidemics caused by infections with presymptomatic, asymptomatic, and symptomatic stages date: 2020-09-10 journal: plos one doi: 10.1371/journal.pone.0238412 sha: doc_id: 297287 cord_uid: 0i4nc353 we investigate phase transitions associated with three control methods for epidemics on small world networks. motivated by the behavior of sars-cov-2, we construct a theoretical sir model of a virus that exhibits presymptomatic, asymptomatic, and symptomatic stages in two possible pathways. using agent-based simulations on small world networks, we observe phase transitions for epidemic spread related to: 1) global social distancing with a fixed probability of adherence. 2) individually initiated social isolation when a threshold number of contacts are infected. 3) viral shedding rate. the primary driver of total number of infections is the viral shedding rate, with probability of social distancing being the next critical factor. individually initiated social isolation was effective when initiated in response to a single infected contact. for each of these control measures, the total number of infections exhibits a sharp phase transition as the strength of the measure is varied. the sars-cov-2 virus that has spread throughout the globe has created societal disruption and had a massive impact on global health [1] . with no known treatment, public policy and human behavior are currently the only tools that are available to mitigate the spread [2] . a fundamental characteristic of sars-cov-2 is that after an individual is exposed, that individual passes through an extended presymptomatic stage followed by either an asymptomatic or symptomatic stage [3] . our goal in this work is to construct a theoretical network disease model with these qualities and investigate phase transitions associated with three types of control measures. while many models related to sars-cov-2 are designed to be forecasting tools, our study is intended as a contribution to the theoretical literature regarding qualitative aspects of control measures for viruses with these pathways of disease progression. in the specific case of sars-cov-2, one control measure that has been used is governmentmandated social distancing. different countries, and different states within the us, have implemented different approaches to this [2, 3] . while most government plans include some social distancing, questions have arisen as to the efficacy of social distancing, how long social distancing should last and to what extent it is needed [4] . a second control method involves individually-determined changes to social behavior, which work in concert with mandated social distancing to mitigate viral transmission [5, 6] . individuals who live with an infected individual are being asked or required to quarantine for 14 days prior to interacting in the larger society [7] . one question is whether these individual responses of behavioral modification are sufficient to moderate epidemic spread and whether there are additive or non-additive effects when implemented with top-down government policy on social distancing [5] . a third type of control measure involves use of personal protective equipment to reduce the rate of viral transmission. for example, mask usage has been found to be effective in this regard for sars-cov-2 [8] [9] [10] . these real-world aspects of sars-cov-2 highlight the need for a more thorough understanding of the general behavior of viruses exhibiting multiple progressions of disease development. with this as motivation, we develop a theoretical model in which we investigate how three types of control measures are associated with sharp phase transitions for the total number of infected individuals. while modeling contacts can be done in mean-field, statistical, and metapopulation sir models [1, [11] [12] [13] [14] , we use an agent-based model (abm) on watts-strogatz small world networks [15] [16] [17] . small world networks have connectedness properties that are found in real-world social networks and have been previously considered in epidemiological contexts [18, 19] . the first control measure in our abm is social distancing imposed on the network at a global scale. our model encodes this global social distancing as complete isolation of an agent from other agents. the likelihood of social distancing is applied uniformly to all agents. the second control measure arises when agents have social connections that are infected and symptomatic [2, 8] . in this case, agents temporarily isolate from their contacts in the network if they are in contact with a sufficient number of symptomatic agents. the third type of control measure is to alter the rate of viral spread, which reflects behavior such as use of personal protective equipment, e.g., masks [8, 10] . we examine how each of these measures alone and in concert with each other influence the viral outbreak. for each of these control measures, we ask the following questions: 1. how does varying the strength of the control measure impact the total number of infections in an epidemic? 2. if a control measure impacts the total number of infections, is there a phase transition associated with changes in strength of that control measure? 3. how do these three control measures interact regarding their impact on total number of infections? we develop an sir, network-based, agent-based model where agents pass through various infection states (fig 1) . agents pass through a presymptomatic infection state followed by either an asymptomatic infected stage or a symptomatic infected stage. in our model, each agent carries an individual pathogen level that changes in response to contact with infected agents. initially, this level is set to 0 pathogen units for susceptible agents. at each time step (conceived as a day), if a susceptible agent has no infected contacts then their pathogen level does not change. for each day that a susceptible agent has one or more infected neighbors, their pathogen level increases by a fixed fraction of the pathogen levels of their infected neighbors. there is a global infection threshold that applies to all agents, which we fix at 25 pathogen units; in other words, the day after the pathogen level for an individual agent exceeds 25 units, that agent enters the presymptomatic infected state. once an agent enters the infected state, their pathogen level stays constant until they have reached the resistant/removed state, at which point it is reset to 0 units. model runs are initiated with a small number of infected agents, whose pathogen levels are initially set at 35 pathogen units, and the remainder of the agents are initially deemed susceptible. these initial and threshold values for the pathogen levels in our model are not based on real-world data, but rather were selected for simplicity to investigate general behavior of phase transitions under a mechanism of viral shedding with individualized pathogen levels. because our model does not use a transmission probability for each contact, but rather a viral shedding rate where each individual agent has varying levels of pathogen load, this model is well-suited to abm simulations and less amenable to ode-based deterministic analysis. once the individual pathogen level for an agent exceeds 25 units, that agent enters a presymptomatic infection stage, followed by either a symptomatic or asymptomatic stage (fig 1) . the length of the presymptomatic stage is the same for all agents, and can be set to last one or more days. the lengths of the two possible main infection stages are set independently from each other, but are the same for all agents. following the main infection stage, the agent is either resistant or removed. in addition to the infection stages, each agent is in one of two daily behavior states: socially distanced or not socially distanced. the behavior state is reset each day. if an agent is not socially distanced on a given day, then that agent can interact with any neighboring agent. if an agent is socially distanced, the agent does not interact with any neighboring agents; in our theoretical model, social distancing is equivalent to self-quarantine. agents socially distance in a given day for one of two reasons. a global social distance probability is set, which determines the chance that an agent will socially distance on a given day. a local social distance threshold is set, and this value dictates individual responses to infected symptomatic neighbors. if the number of infected symptomatic neighbors of an agent equals or exceeds this threshold, the agent will social distance independently of the global parameter. this local threshold is the same for all agents. we implement our agent-based model on watts-strogatz (ws) networks. the ws networks can simultaneously demonstrate both high clustering and short average path length, and thus serve as effective approximations of social networks that are neither completely random nor regular [15] . high clustering and short average path length allow for local interactions and more distant interactions to be incorporated [16, 17, 20] , which are properties often found in real-world networks. the ws small world network in our model is characterized by three parameters: number of nodes n, average node degree k and rewiring probability. the rewiring parameter is used to determine the likelihood of rewiring each edge starting from a regular ring lattice. a rewiring parameter of 0 preserves the original ring lattice; a rewiring parameter of 1 simulates a random network. we fix the number of nodes n = 500 and the average degree k = 20, which allows ln(n) � k � n. we then vary the rewiring probability among the values {0.05, 0.10, 0.25, 0.50}. for each of our four rewiring probabilities we construct 10 networks on which to run simulations. we define our three model parameters as follows: 1. social distance probability: the probability that an agent is socially distancing on any given day. 2. social distance threshold: the minimum number of infected symptomatic contacts required to cause an agent to social distance for that day. 3. viral shedding: the fraction of individual pathogen level that an infected agent passes to each of its contacts. we ran two sets of simulations over different parameter spaces. our primary simulation ran through ten networks for each set of parameters given in table 1 . based on the results of this primary simulation, we ran a secondary set of simulations over the refined parameter space given in table 2 to provide a more detailed analysis of the phase transition behavior observed in the primary simulations. the parameters for the secondary simulation were selected based on our analysis of the primary data using regression trees to identify critical variables and on the observed ranges where phase transitions were observed. we used a regression tree to partition the variation in final number of infected nodes across model parameters and runs in our primary simulation [21] . reductions in viral shedding were associated with the primary partition in the regression tree in fig 2. viral shedding below 15% compared to a value of 25% were associated with a mean number of infections of 46 out of 500 agents. reduced viral shedding with social distancing probability over 25% led to overall infection of approximately 2% of the agents. if the overall viral shedding is reduced dramatically to 5%, even without additional social distancing of any type, less than 1% of the population becomes infected. achieving low levels of infections in populations without reducing viral shedding requires significantly higher levels of global social distancing, where each agent has at least a 75% chance of social distancing each day; this results in an approximately 1% infection rate among agents. if each agent has less than a 75% chance of social distancing each day, the total infection rates for the populations are much higher; these range from a low of 21% (if individuals phase transitions and control measures for network epidemics self-isolate in response to one infected social contact) all the way up to 97% with low levels of any type of social distancing. thus, with a higher level of viral shedding, it becomes important to have agents self-isolate when a contact becomes symptomatic. even if this occurs, the infection rate in the population is an order of magnitude higher (10% vs. 1%) than if the viral shedding is reduced. failure to achieve this strict social distancing in response to an infected social contact results in a widespread outbreak with approximately 62% of the agents infected. because our goal is to understand the behavior of phase transitions regarding total number of infections in our model, we conducted secondary simulations on a refined parameter space based on the results of our regression tree analysis. in these simulations, we observed sharp phase transitions in the total number of infections as a function of all three control methods. these transitions are shown in figs 3, 4 and 5. in these figures, the maximum number of possible infections is 500, as there are 500 nodes in our networks. in fig 3, a phase transition exists between viral shedding of 5% and 10%, across all levels of social distance thresholds and social distance probabilities. in fig 4, a phase transition exists at a social distance threshold of 1, across all levels of social distance probabilities and viral shedding. if the social distance threshold parameter is 2 or more, then it is possible to have epidemics that infect the entire population. in fig 5, a phase transition exists around a social distance probability of 73-74%, across all levels of social distance threshold and viral shedding. if the social distance probability is 74% or more, then our simulations end with a small number of infected agents. phase transitions and control measures for network epidemics given the regression tree analysis of our primary simulations, it is clear that viral shedding and social distance probability play key roles. in our secondary simulations over a refined parameter space, this becomes more clear. in fig 6, we observe additional confirmation of the regression tree findings that the main driver of total number of infections is the viral shedding rate, with social distance probability being the next critical factor. specifically, simulations with large total infections cluster to the upper left of the plot, where viral shedding rates are higher and social distancing is enacted by approximately 60% of agents. there is also a clear interaction between the social distance probability and viral shedding parameters and the resulting number of infected agents and the length of the epidemic. these interactions are shown in figs 7 and 8. in fig 7, there is clustering of long epidemics when the probability is near 60% and the viral shedding rate is high. as the social distance probability increases to 80% and the viral shedding rate decreases, there is a phase transition where simulations result in outbreaks of short duration. in fig 8, most infections result in either a limited outbreak (less than 125 out of 500 agents) or almost all agents infected. as the social distance probability is increased from 60% to 80%, the length of the epidemics increase while remaining limited in total number of infections before sharply transitioning to a high number of infections during a return to short epidemic lengths. mathematical modeling can provide tools to better understand epidemic dynamics and can vary from purely theoretical to more data driven and predictive [22] . while a simple model such as this one should not be used to make policy recommendations, it can provide a framework for empirical investigation and specific hypothesis testing related to social networks of smaller size exhibiting small world characteristics, such as those seen in college settings [23] . here we use our theoretical model to investigate how different control methods impact the total number of infections in an epidemic on a network caused by a virus with a presymptomatic stage and both asymptomatic and symptomatic pathways. we specifically examine three main control measures that can be taken to reduce epidemic spread: 1) global social distancing with a fixed probability of adherence. 2) individually initiated social isolation when a threshold number of contacts are infected. 3) reduction of viral shedding. we observe sharp phase transitions in the total number of infected agents as the strength of each of these control measures are varied. to examine the full potential for global social distancing, we consider a wide range of possible scenarios varying from no social distancing to strong adherence to social distancing (90% of agents). when considering the relationship between our theoretical model and real-world contexts, the two extreme scenarios are easy to envision (zero social distancing is business as usual and 90% is all but non-essential businesses closed), while more moderate social distancing scenarios are harder to translate into direct societal actions. nevertheless, we observe in our small-world models a clear phase transition associated with global social distancing. in general, a global social distancing probability below 65% results in a wide-spread epidemic, while a global social distancing probability above 75% limits the epidemic to a dramatically lower number of total infections. we also found that social distance probabilities that approached the threshold from below resulted in prolonged epidemics while with low overall infection rates. for our secondary simulations over a refined parameter space, in the absence of other control measures we observe that there is a phase transition for total infections that occurs as the percentage of agents socially distancing changes from 73% to 74%. individual behavior taken during a pandemic can greatly affect the dynamics of disease spread. for example, for sars-cov-2, the most commonly recommended guideline after contact with an infected individual is 14 days of self-isolation to avoid exposing other individuals [5, 7, 24] . however, despite these official guidelines, self-isolation following exposure requires that infected individuals inform their contacts and that exposed individuals voluntarily comply. thus, from a theoretical perspective it is important to understand how different self-isolation behaviors following contact with an infected agent impact epidemic spread. in our model, we consider self-imposed social distancing as highly responsive to an agent's short-term perceptions regarding infection risks within their community. thus, self-imposed social distancing/isolation occurs only on the days when the agent has sufficiently many symptomatic contacts in the network. interestingly, for self-isolation to significantly decrease the total number of infections in our model, an extreme level of responsiveness was needed by the agents involved; in our model, it was necessary for self-isolation to occur following exposure to only one infected agent. if self-isolation occurred only after contact with two or more symptomatic agents on the same day, the effect on disease spread was minimal. our findings also support the well-known fact that real-world contact tracing following an individual's positive test is critical for limiting the spread of the infection [4] . an important difference between our theoretical model and viruses such as sars-cov-2 [13, 25] is that, in the real world, individuals who have come in contact with an infected individual are not aware of their exposure. our theoretical reduction of viral shedding is motivated by behaviors such as mask wearing or other use of personal protective equipment [10] . while in real-world contexts individual responsiveness to recommended government actions are highly variable [24] , a decrease in viral shedding rate can be achieved through use of protective equipment [9] . when the viral shedding in our model was set at a high shedding rate of 25%, global social distancing was required to be greater than 80% to control the outbreak, resulting in an approximately 1% infection rate in the population. other less stringent social distancing conditions result in a viral infection rate between 25% and 97.5%. with a moderate rate of viral shedding, the social distance threshold at which someone decides to self-isolate after coming into contact with an infected individual becomes much more important. in our model, if the social distance threshold is set to 1 (agents self-isolate after coming into contact with at least one infected agent), then the final total infection rate in the population is approximately 12%. however, if the behaviorally induced social distancing does not take place or takes place at a higher threshold, then the total number of infections is much larger with the overall infection rate in the population approximately 62%. if the viral shedding rate is very low, then the epidemic does not spread and a low total number of infections is observed. thus, there is a sharp phase transition as the viral shedding rate moves from 5% to 10%. an important observation regarding these phase transitions is the relatively extreme values at which they occur, e.g., a high social distancing probability, a low social distancing threshold, and a low viral shedding rate. these values are very high and low both within the context of our model and of real-world epidemics that motivate our model. it would be of interest to investigate whether or not, given an arbitrary set of values, a specific network and selection of parameters could be found for which phase transitions occur near these values. alternatively, if no such network and choice of parameters exist, it would be of interest if a more rigorous theoretical description could be given of the mechanism preventing this occurrence. we develop an agent-based model of epidemic spread on watts-strogatz small world networks, where infected agents pass through a presymptomatic stage followed by either an asymptomatic or symptomatic stage. we consider the impact of three control measures on the total number of infected agents, with regard to both phase transitions and efficacy. the three control methods we consider all generate sharp phase transitions in the total number of infections as the strength of the method varies. social distancing controls in this model exhibit a phase transition regarding total number of infections, either when imposed globally or when based on individual response to infected contacts. individually-enacted social distancing in the form of temporary self-isolation must be immediately enacted if a social contact is known to be infected in order to halt the spread of an epidemic. reductions in viral shedding lead to significant reductions in the size of the final infected population. forecasting the novel coronavirus covid-19 time-to-death approach in revealing chronicity and severity of covid-19 across the world covid-19: the cidrap viewpoint part 1: the future of the covid-19 pandemic: lessons learned from pandemic influenza feasibility of controlling covid-19 outbreaks by isolation of cases and contacts. the lancet global health using social and behavioural science to support covid-19 pandemic response changes in risk perception and protective behavior during the first week of the covid-19 pandemic in the united states the psychological distress and coping styles in the early stages of the 2019 coronavirus disease (covid-19) epidemic in the general mainland chinese population: a web-based survey behavioural change models for infectious disease transmission: a systematic review & covid-19 systematic urgent review group effort (surge) study authors physical distancing, face masks, and eye protection to prevent person-to-person transmission of sars-cov-2 and covid-19: a systematic review and meta-analysis (2020) journey through an epidemic: some observations of contrasting public health responses to sars analysis and forecast of covid-19 spreading in china a simple sir model with a large set of asymptomatic infectives geo-temporal distribution of 1,688 chinese healthcare workers infected with covid-19 in severe conditions-a secondary data analysis machine learning using intrinsic genomic signatures for rapid classification of novel pathogens: covid-19 case study collective dynamics of 'small-world'networks classes of small-world networks disease transmission in territorial populations: the small-world network of serengeti lions network-based analysis of stochastic sir epidemic models with random and proportionate mixing coupled contagion dynamics of fear and disease: mathematical and computational explorations the watts-strogatz network model developed by including degree distribution: theory and computer simulation rpart: recursive partitioning and regression trees mathematics of epidemics on networks the small world network of college classes: implications for epidemic spread on a university campus perceptions of the adult us population regarding the novel coronavirus outbreak data-based analysis, modelling and forecasting of the covid-19 outbreak this project began during the workshop "understanding and exploring network epidemiology in the time of key: cord-285617-nyocnvvj authors: stramer, s l title: current perspectives in transfusion-transmitted infectious diseases: emerging and re-emerging infections date: 2014-07-28 journal: isbt sci ser doi: 10.1111/voxs.12070 sha: doc_id: 285617 cord_uid: nyocnvvj background: in august 2009, a group from the aabb (stramer et al., transfusion 2009;99:1s–29s, emerging infectious disease agents and their potential threat to transfusion safety; http://www.aabb.org/resources/bct/eid/pages/default.aspx) published a supplement to transfusion that reviewed emerging infectious disease (eid) agents that pose a real or theoretical threat to transfusion safety, but for which an existing effective intervention is lacking. the necessary attributes for transfusion transmission were outlined including: presence of the agent in blood during the donor's asymptomatic phase, the agent's survival/persistence in blood during processing/storage, and lastly that the agent must be recognized as responsible for a clinically apparent outcome in at least a proportion of recipients who become infected. without these attributes, agents are not considered as a transfusion-transmission threat and were excluded. sixty-eight such agents were identified with enough evidence/likelihood of transfusion transmission (e.g., blood phase) and potential for clinical disease to warrant further consideration. in the supplement, fact sheets (fs) were published providing information on: agent classification; disease agent's importance; clinical syndromes/diseases caused; transmission modes (including vectors/reservoirs); likelihood of transfusion transmission, and if proven to be transfusion-transmitted, information on known cases; the feasibility/predicted success of interventions for donor screening (questioning) and tests available for diagnostics/ adapted for donor screening; and finally, the efficacy, if known, of inactivation methods for plasma-derived products. the supplement included a separate section on pathogen reduction using published data. agents were prioritized relative to their scientific/epidemiologic threat and their perceived threat to the community including concerns expressed by the regulators of blood. agents given the highest priority due to a known transfusion-transmission threat and severe/fatal disease in recipients were the vcjd prion, dengue viruses and the obligate red-cell parasite that causes babesiosis (b. microti and related babesia). although the focus of the supplement was towards the united states and canada, many of the agents (and the process) are applicable worldwide. next steps: since the publication of the supplement, six new fss (yellow fever viruses-including vaccine breakthrough infections, miscellaneous arboviruses, xmrv, human parvoviruses/bocaviruses other than b19, and most recently the middle east respiratory syndrome coronavirus, mers-cov) were added and 14 existing fss updated (anaplasma, babesia, bartonella, erhlichia, chronic wasting disease-cwd, human prions other than vcjd, vcjd, coxiella burnetii-the agent of q fever, dengue viruses, hav, hev, japanese encephalitis-je complex, tick-borne encephalitis viruses-tbev, and human parvovirus b19). also, tables were released outlining pathogen reduction clinical trials/results (published) and availability/commercial routine use of such technologies by country. of necessity, the list of eid agents is not, and can never be, complete due to the nature of emergence. we recognized that a system of assessing the risk/threat of eids for their potential impact on blood safety and availability must include processes for monitoring, identifying, evaluating, estimating severity, assessing risk and developing interventions. thus, a ‘toolkit’ containing the necessary ‘tools’ from eid monitoring (horizon scanning) to validation/effectiveness evaluations of interventions is being developed. the goal is, to develop a systematic approach to risk assessment and intervention development for the impact of emerging infectious upon blood safety intended to educate and provide advise about risks/interventions in a timely/accurate fashion. conclusions: the process and final product (toolkit) including methods to monitor eid agent emergence, identification/recognition of a transfusion-transmission threat, methods for quantitative risk assessments, and the appropriate management of such threats should be considered for implementation by all blood systems. disease-cwd, human prions other than vcjd, vcjd, coxiella burnetii-the agent of q fever, dengue viruses, hav, hev, japanese encephalitis-je complex, tick-borne encephalitis viruses-tbev, and human parvovirus b19). also, tables were released outlining pathogen reduction clinical trials/results (published) and availability/commercial routine use of such technologies by country. of necessity, the list of eid agents is not, and can never be, complete due to the nature of emergence. we recognized that a system of assessing the risk/threat of eids for their potential impact on blood safety and availability must include processes for monitoring, identifying, evaluating, estimating severity, assessing risk and developing interventions. thus, a 'toolkit' containing the necessary 'tools' from eid monitoring (horizon scanning) to validation/effectiveness evaluations of interventions is being developed. the goal is, to develop a systematic approach to risk assessment and intervention development for the impact of emerging infectious upon blood safety intended to educate and provide advise about risks/interventions in a timely/accurate fashion. the recognition of newly described infectious disease agents will continue to place demands on the collectors of blood worldwide to ensure safety. approaches are needed to ensure that mechanisms are in place to allow surveillance, threat assessments, triggers for action, and as needed, intervention development, implementation and assessment of efficacy. agents of concern are diverse in nature and emergence is unpredictable. some examples of agents internationally recognized as potential or existing threats to blood safety include the newly described middle east respiratory syndrome coronavirus (mers-cov), dengue viruses, chikungunya virus and hepatitis e virus (hev). in august 2009, a group from the aabb in the united states [1] published a supplement to transfusion that reviewed the definition and background of eid agents that pose a real or theoretical threat to transfusion safety, but for which, an existing effective intervention is lacking. the definition includes recognition of a new agent, a previously undetected agent or that a disease has an infectious origin and lastly, the re-emergence of a known infection after a decline in incidence. one estimate of the rate of emergence from 1940 to 2004 includes 5á3 new viruses discovered every year, of which 60-70% have an animal origin but can infect humans in cases where exposure by one of several routes occurs; the predicted rate of such emergence will continue well beyond the year 2020 [2] [3] [4] . human exposure can occur by a number of overlapping conditions, many of which humans have clearly precipitated in 'forcing' a new agent by mutation, species jump, failure of prior control measures, population growth and movement, transportation, behavioural changes (sanitary, dietary and war) and intensive farming practices to name a few. in addition to most eid agents having a zoonotic origin, they represent a diverse group of agents, infect by multiple transmission routes, may result in acute or chronic infections, derive from human activity, but most importantly, their emergence is unpredictable. the necessary attributes for transfusion transmission include the following: presence of the agent in blood during an asymptomatic phase in the donor, the agent's survival/persistence in blood during processing/storage and lastly that the agent must be recognized as responsible for a clinically apparent outcome in at least a proportion of recipients who become infected [1, 2] . the response with respect to blood safety approaches to eid agents has varied related to the severity of the agent, its incidence and prevalence and rate of emergence. originally, the aabb group identified 68 eid agents of concern to blood safety; each agent had enough evidence/likelihood of transfusion transmission (e.g. blood phase) and potential for clinical disease to warrant further consideration. in the supplement [1] , fact sheets were published providing information on the following: agent classification; background on the disease agent's importance; the clinical syndromes/diseases caused; transmission modes (including vectors/reservoirs); likelihood of transfusion transmission, and if proven to be transfusion transmitted, information on known cases; the feasibility and predicted success of interventions that could be used for donor screening (questioning) and tests available for diagnostics or that could be adapted for donation screening; and finally, the efficacy, if known, of inactivation methods for plasma-derived products. the supplement also included a separate section on pathogen reduction technologies for all blood components using published data. agents were prioritized relative to their scientific/epidemiologic threat as well as their perceived threat to the community including concerns expressed by the regulators of blood. agents given the highest priority due to a known transfusion-transmission threat, and severe/fatal disease in recipients were the vcjd prion, dengue viruses and the obligate red-cell parasite that causes babesiosis (b. microti and related babesia). although the focus of the supplement was the united states and canada, many of the agents (and the process) are applicable worldwide. other experts have prioritized agents differently depending on their local needs with priority given to west nile virus, dengue viruses, leishmania, chikungunya virus, the agents of malaria and lyme disease, and tick-borne encephalitis virus [5] ; of note, several of these agents, including the agent of lyme disease, have never been documented to be transfusion-transmitted. since the publication of the supplement, six new fact sheets were released: (1) yellow fever viruses including vaccine breakthrough infections, (2) miscellaneous arboviruses, (3) xmrv including a comprehensive table of published literature, (4) human parvoviruses other than b19, and bocaviruses, (5) measles due to localized us outbreaks, also, 14 existing fact sheets were updated: (1) human prions other than vcjd, (2) the chronic wasting disease prion, (3) the vcjd prion, (4) bartonella, (5) coxiella burnetii the agent of q fever, which resulted in a massive outbreak in the netherlands from 2007 to 2010 precipitated by high-intensity goat farming, (6) hev due to increasing reports of rna-positive blood donors in japan and europe; of note, in the netherlands, increasing numbers precipitated by high-intensity pig farming, (7) japanese encephalitis (je) complex, (8) tick-borne encephalitis viruses (tbev), (9) dengue viruses with three transfusion-transmission clusters, one each reported in hong kong, singapore and puerto rico, (10) human parvovirus b19, (11) hepatitis a virus (hav) due to a large multi-state outbreak in the us, (12) anaplasma phagocytophilum with eight transfusion transmissions reported in the us, (13) erhlichia including the first report of transfusion transmission in the us, (14) b. microti with data reported for research testing interventions in the us. in addition, tables were released outlining pathogen reduction clinical trials and their published results, and the availability and commercial routine use of such technologies by country for platelets, plasma, red cells and whole blood. of necessity, the list of eid agents is not, and can never be, complete due to the nature of emergence. we recognized that a system of assessing the risk and threat of eids for their potential impact on blood safety and availability must include a process for monitoring, identifying, evaluating, estimating disease severity, assessing risk and development of interventions. thus, the aabb eid group is now developing a 'toolkit' containing the necessary 'tools' starting with links to eid monitoring sites (horizon scanning), risk assessment tools, a flow diagram to follow as one proceeds to determine whether intervention development and implementation are required, and methods to validate and assess the efficacy of any introduced intervention [2] . the goal is 'to develop a systematic approach to risk assessment and intervention development for the impact of emerging infections upon blood safety intended to educate and advise blood systems in a timely and accurate fashion'. when developed, this toolkit may be adapted to the needs of isbt members. the lessons learned from the aabb exercise are that each country or region of the world needs to decide their own priority list of infectious agents that may represent a threat to transfused recipients, as well as processes for surveillance and action should any of those agents pose an immediate safety threat. countries, especially those in the developing world where eid agents have originated in the past such as china, sub-saharan africa, tropical central and south america, may also choose to establish an epidemiologic/molecular surveillance laboratory or collaborations with any such existing laboratories to enable rapid agent identification and disease characterization. some examples of agents internationally recognized as potential or existing threats to blood safety include mers-cov, dengue viruses, chikungunya virus and hev; each will be briefly reviewed. the mers-cov agent is a betacoronavirus first described in september 2012 from a patient in saudi arabia [6] . it is related to the severe acute respiratory syndrome coronavirus (sars) that from 2002 to 2003 resulted in over 8000 infections in 17 countries (including macau and hong kong) with a 10% fatality rate [7] . in the case of sars, bats were confirmed as the natural reservoir although the introduction of the agent to humans was likely through contact with animals (masked palm civets) held in southern chinese markets where the disease was first recognized in nov 2002 [8, 9] . middle east respiratory syndrome coronavirus also likely represents a zoonotic agent that infected humans via viral adaptions or a species jump through an intermediate host. bats are suspected as the ultimate reservoir with camel infection as the most likely link to human infection, but these theories at present are unproven. the majority of cases are associated with severe disease of the lower respiratory tract resulting in pneumonia and multi-organ failure; the most susceptible to clinical disease and death are those with pre-existing co-morbidities [10] . other chronic conditions have been reported in 96% of clinical cases including diabetes, hypertension, heart disease and kidney disease. it is unclear whether persons with these specific conditions are disproportionately infected or have more severe disease. to 30 november 2013, as reported by the us centers for disease control and prevention (cdc) and the world health organization (who), ten countries have reported 160 clinical cases with 68 deaths. all cases and transmissions have been associated with the arabian peninsula where the vast majority of cases and deaths have occurred; travel of infected patients with limited person-to-person transmission are responsible for cases reported elsewhere. although it appears that there is limited person-to-person transmission, clusters of cases have been described. a cluster of 21 human cases occurred associated with four hospitals in eastern saudi arabia that was ultimately sourced to one community-acquired case [11] . updated mers-cov rapid risk assessments are routinely published by the european centres for disease control (ecdc). regarding actions with respect to blood safety, careful surveillance is required as risk involving mers-cov is rapidly evolving at present. to date, only one study has reported testing blood donors including 110 in saudi arabia; all were nonreactive for mers-cov neutralizing antibodies [12] . there is no evidence of transfusion transmission. dengue viruses include four genetically related viruses (65% genetic homology between the four) classified as 'arboviruses' (arthropod-borne viruses). all arboviruses involve replication through an arthropod host (usually a mosquito or tick) have a worldwide distribution and are responsible for huge periodic epidemics. dengue viruses are in the family flaviviridae along with related viruses: west nile virus, yellow fever virus and over 65 other related viruses. arboviruses also include hundreds of agents, of which many are human pathogens in other viral families (bunyaviridae and togaviridae including the alphavirus, chikungunya described below). dengue viruses are the most important of all of these agents due to the number of human infections, clinical disease, associated deaths and ongoing expansion worldwide. it is estimated by the who that greater than 100 countries are endemic in the tropics and subtropics with over 2á5 billion people at risk. it is impossible to know the number of clinical cases that actually occur each year, but one estimate of the global burden of dengue during the 2010 worldwide pandemic was 390 million infections across four continents of which 96 million were symptomatic including 500 000 cases of severe dengue [13] . in asia and latin america, dengue is the leading cause of hospitalization in children. since there is no effective vaccine or specific treatment, vector control is the only intervention. the dengue cycle involves humans and mosquitoes (primarily a. aegypti), but sylvatic cycles also occur and are recently being investigated associated with preliminary reports of a new dengue virus type (denv-5). immunity to a given viral type is considered lifelong, but due to incomplete antibody neutralization between types, secondary infection with a different type can lead to more severe disease referred to 'severe dengue' including haemorrhagic fever and/or shock. the majority of clinical cases are classified as 'fever' defined by who as fever plus two other symptoms most commonly including chills, painful eyes, body aches (that can be severe, 'break-bone fever'), rash, easily bruised or other evidence of haemorrhagic conditions. three transfusion-transmission clusters have been reported [1] ; the first in hong kong, another in singapore (where two clinical cases and one antibody seroconversion were documented) and lastly a case of transfusion-transmitted haemorrhagic fever in puerto rico [14] . infected individuals may donate blood since it has been estimated that 53-87% of dengue-infections are asymptomatic, and even individuals who develop symptoms will have a 1-2 day asymptomatic period. puerto rico, an island in the caribbean and a us territory, experiences annual outbreaks. research blood donation screening in puerto rico has documented the rate of dengue rna in blood donors during the outbreak years of 2005, 2007, 2010, 2011 and 2012 at between 0á03% and 0á31%. the estimated number of viremic donations and the risk of dengue transfusion transmission have been modelled by the cdc. the resulting model output and the observed number of rna-positive donors from blood donation screening show the same trend [15] . although routine screening by the american red cross occurs in puerto rico, the clinical efficacy of the intervention is unknown since the number of transfusion transmissions documented has been few. it is unclear why the number of transfusion transmissions is not higher in endemic areas considering the magnitude of dengue outbreaks and high viral loads associated with infected donors. possible theories include an absence of effective hemovigilance in many of the countries impacted by dengue, inability to differentiate mosquito versus transfusion transmission, protection of secondary infection due to heterologous antibodies present in the transfused unit, immunosuppression in many recipients and possibly different clinical outcomes dependent on the route of infection (i.e. mosquito versus transfusion) [16] . another flavivirus, chikungunya virus, is endemic in many of the same tropical areas as dengue and frequently presents with similar symptoms and during the same epidemic periods but with more severe body aches, especially joint pains and crippling arthritis than that of dengue. the name 'chikungunya' refers to the inability to straighten up due to intense pain ('that which bends up' in the makone language) [1] . an explosive outbreak occurred on reunion island and the islands of the southwest indian ocean off east africa from late 2005 through the beginning of 2007 [17] . during this outbreak, over 300 000 clinical cases on reunion island were documented effecting over 40% of the population with 75% of those infected exhibiting symptoms with another 1á3 million cases during the same outbreak in india. the increased severity and extent of the 2005-2007 outbreak was believed to be due to a mutation affecting the viral envelope protein allowing replication in an alternate mosquito vector, a. albopictus [18] . this resulted in increased viral loads and greater virulence as compared to replication in its primary mosquito vector, a. egypti. transfusion-transmitted risk modelling estimated a risk of up to 1500 per 100 000 donations [17] ; however, transfusion transmission was never documented. what is notable in the islands of this french colony is the number of interventions put into place to prevent theoretic transfusion transmission. these included the suspension of blood collections in areas where risk was estimated to be higher than the risk from hepatitis b virus transfusion transmission, the implementation for platelets of pathogen inactivation and chikungunya virus rna nucleic acid testing, and for continental france, the deferral of those who lived or travelled to an endemic area [19] . subsequently, an outbreak in northern italy occurred in 2007 involving approximately 250 cases; the viral introduction was believed to have been a clinical case in an individual returning from india, and consequently, spread locally via a. albopictus [20] . the data from this outbreak were used to validate a risk assessment tool developed to provide quantitative transmission estimates of eid risks through blood transfusion. the output of the model based on the outbreak peak estimated a prevalence of 1á07 per 10 000 donors, leading to 0á04 infectious donations, 0á13 infectious blood components and a severe outcome in 0á0001 recipients, based on an assumed 0á1% of infected individuals who develop severe disease [21] . subsequently, a case-control study during a chikungunya outbreak in thailand in 2009 investigated the development of symptoms and viremia. of 134 laboratory-confirmed positives, 9% (n = 12) were asymptomatic controls who were rna positive or seroconverted having a median viral load of 3á4 9 10 3 pfu/ml (range: 8á4 9 10 1 -2á9 9 10 5 pfu/ml) versus 91% (n = 122) classified as symptomatic cases with viremia for 8 or fewer days and rna positivity for 17 or fewer days. symptomatic individuals had a median viral load of 5á6 9 10 5 pfu/ml (range: 1á3 9 10 1 -2á9 9 10 8 pfu/ml; p = 0á22 vs. controls) [22] . hepatitis e virus is a single-stranded, rna-containing, small, non-enveloped virus in the genus hepesvirus. there is one serotype but four genotypes with differing geographic distribution, epidemiology and clinical features. hepatitis e virus is globally the most common cause of acute hepatitis with an estimated 20 million incident infections per year, over 3 million acute cases and 700 000 deaths per year [1, 2] . genotypes 1 and 2 are most commonly associated with large, explosive waterborne outbreaks along with some foodborne outbreaks similar to hav. genotypes 1 and 2 also are associated with an overall higher rate of acute disease and for an unknown reason with increased severity in pregnant women (20% mortality especially in the third trimester). in contrast, genotypes 3 and 4 occur in humans but commonly in swine and several other animals. thus, infections with genotypes 3 and 4 have occurred from transmissions through food, most notably via raw or undercooked pork products containing liver and/or blood [1, 2] . generally, less virulent than either genotype 1 or 2, genotype 3 or 4 infection results in severe disease in immunosuppressed individuals, most notably solid organ transplant (sot) recipients leading to chronic infection (>60%), and to cirrhosis and severe liver disease in 14% of those who become chronically infected [23] . of 1200 sot recipients investigated for hev rna in the netherlands, 1% were positive; of those, 11 of 12 hev-rna-positive patients developed cirrhosis [24] . antibody prevalence in blood donors varies greatly (<1% to >50%) depending on geographic location and the test used but is reproducibly seen to increase with age unrelated to a cohort effect but more likely the result of increasing prevalence over time [25] [26] [27] . rna from genotype 3 can be recovered from dutch blood donors at rates of as high as 1:3000 with sequences closely related to patients and pigs in the area [25] . in the same study, igg antibody prevalence was 27%, and 3á5% for igm, with increasing prevalence with age. intensive pig farming in the netherlands, as an example, may be amplifying the virus, which is then spread via contaminated meat and contaminated water used for irrigation. even higher igg seroprevalence rates of 52á5% have been documented in southwest france [26] , linked to the consumption of locally produced pork products containing undercooked pork. in another study, 1939 blood donors in the us had a prevalence of igg of 18á8% (95% ci: 17á0-20á5%) and 0á4% for igm, but no donor had circulating rna [27] . of the 1939 donors, 916 were collected in 2006 (vs. the remainder in 2012 in which prevalence was 16á0%; p < 0á01 vs. 2006); prevalence ranged from 3á4% in those between 18 and 35 years old to 42á2% in those > 65 years old. as part of this study, a donor-recipient-linked repository was tested, in which two suspect cases of hev transfusion transmission were investigated but neither could be confirmed. in one case, passive antibody (and rna) was transfused, but the patient died prior to determining the infectivity of the transfused hev rna from the only rna-positive donor in the study. the other case was one where the level of igg pre-transfusion was just below the assay cut-off, and post-transfusion was just above the assay cutoff. since further antibody evolution did not indicate recent seroconversion, the recipient was likely infected prior to transfusion. since 1979, at least 10 transfusion transmissions have been documented worldwide including cases in japan, the uk, france, saudi arabia, taiwan and india. because hev is nonenveloped, it is very resistant to inactivation; thus, some solvent-detergent plasma manufacturers screen plasma donors and/or pools to reduce the hev rna concentration (i.e. octaplas in the us); commercial tests are being developed to screen plasma for hev rna. each blood system must consider preparing for the unknown including triggers for action (or not and why) and the development of interventions with respect to the safety of blood from eid agents. the examples reviewed represent a wide variety of agents with diverse characteristics, epidemiology, transmission routes and very different risks leading to differing considerations for the introduction of an intervention. the tool-kit development process and final product including methods to monitor eid agent emergence, identify and recognize a transfusion-transmission threat, quantify risk, and appropriately reduce the associated risk to transfusion recipients should be considered for implementation by all blood systems. the author declares no conflict of interests. emerging infectious disease agents and their potential threat to transfusion safety transfusion-transmitted emerging infectious diseases: 30 years of challenges and progress global trends in emerging infectious diseases temporal trends in the discovery of human viruses domanovi c d: commentary; blood supply under threat isolation of a novel coronavirus from a man with pneumonia in saudi arabia epidemic and pandemic alert and response (epr). world health organization bats are natural reservoirs of sars-like coronaviruses severe acute respiratory syndrome coronavirus-like virus in chinese horseshoe bats updated information on the epidemiology of middle east respiratory syndrome coronavirus (mers-cov) infection and guidance for the public, clinicians, and public health authorities for the ksa mers-cov investigation team: hospital outbreak of middle east respiratory syndrome coronavirus lack of mers coronavirus neutralizing antibodies in humans the global distribution and burden of dengue dengue viremia in blood donors identified by rna and detection of dengue transfusion transmission during the 2007 dengue outbreak in puerto rico estimated prevalence of dengue viremia in puerto rican blood donations dengue in the context of "safe blood" and global epidemiology; to screen or not to screen? estimated risk of chikungunya viremic blood donations during an epidemic on reunion island in the indian ocean genome microevolution of chikungunya viruses causing the indian ocean outbreak chikungunya virus: possible impact on transfusion medicine the chikungunya epidemic in italy and its repercussion on the blood system modeling the transmission risk of emerging infectious diseases through blood transfusion viremic profiles in asymptomatic and symptomatic chikungunya fever: a blood transfusion threat? factors associated with chronic hepatitis in patients with hepatitis e virus infection who have received solid organ transplants hepatitis e virus infection among solid organ transplant recipients, the netherlands silent hepatitis e infection in dutch blood donors hepatitis e virus antibodies in blood donors an assessment of hepatitis e virus in us blood donors and recipients: no detectable hev rna in 1939 donors tested and no evidence of hev transmission to 362 prospectively followed recipients key: cord-005033-voi9gu0l authors: xuan, huiyu; xu, lida; li, lu title: a ca-based epidemic model for hiv/aids transmission with heterogeneity date: 2008-06-07 journal: ann oper res doi: 10.1007/s10479-008-0369-3 sha: doc_id: 5033 cord_uid: voi9gu0l the complex dynamics of hiv transmission and subsequent progression to aids make the mathematical analysis untraceable and problematic. in this paper, we develop an extended ca simulation model to study the dynamical behaviors of hiv/aids transmission. the model incorporates heterogeneity into agents’ behaviors. agents have various attributes such as infectivity and susceptibility, varying degrees of influence on their neighbors and different mobilities. additional, we divide the post-infection process of aids disease into several sub-stages in order to facilitate the study of the dynamics in different development stages of epidemics. these features make the dynamics more complicated. we find that the epidemic in our model can generally end up in one of the two states: extinction and persistence, which is consistent with other researchers’ work. higher population density, higher mobility, higher number of infection source, and greater neighborhood are more likely to result in high levels of infections and in persistence. finally, we show in four-class agent scenario, variation in susceptibility (or infectivity) and various fractions of four classes also complicates the dynamics, and some of the results are contradictory and needed for further research. focus on hiv/aids transmission among human groups in order to better understand its dynamical behavior. in epidemic modeling (see, e.g., bailey 1975; anderson and may 1991; murray 2005) , there are two frequently used methodologies: mathematical and simulation methods. for mathematical approaches, a cohort of people is often classified into susceptibles, infectives, and recovereds with (without) immunity (see, e.g., kermack and mckendrick 1927) . systems of differential equations are used to describe the linear (nonlinear) dynamics of epidemics. macroscopically, mathematical models can reveal the relationship among primary factors and describe their effects to epidemic spreading under certain assumptions. as to hiv/aids epidemic, many models have been proposed (may and anderson 1987; hyman et al. 1999; brauer and driessche 2001; wu and tan 2000, etc.) . however, mathematical approaches have some serious drawbacks due to its intractability and the complexity of epidemics. moreover, the complicated nature of hiv/aids transmission makes it even harder to obtain analytical solutions and difficult to study them. in the early 1990s, some researchers started to apply simulation approaches to this field. there is a large literature that addresses the computer simulation of epidemic dynamics (see, e.g., leslie and brunham 1990; atkinson 1996; rhodes and anderson 1996; rhodes and anderson 1997; ahmed and agiza 1998; benyoussef et al. 2003; tarwater and martin 2001) . particularly, cellular automata (ca) method (some literature refers to this as a lattice-based method) has been widely used in modeling complex adaptive systems. despite of its simple structure, ca is well suited to describing the propagation phenomena, such as rumor spreading, particle percolation, innovation propagation, and disease spreading. for instance, in epidemic modeling, fuentes and kuperman (1999) propose two ca models corresponding to the classical mathematical sis model and sis model respectively. ahmed and agiza (1998) develop a ca model that takes into consideration the latency and incubation period of epidemics and allow each individual (agent) to have distinctive susceptibility. gao et al. (2006) put forward a ca model for sars spreading which takes account of social influence. more recently, other methods such as agent-based modeling and system dynamics are introduced to this field (see, e.g., gordan 2003; bagni et al. 2002) . our paper contributes to this filed by developing an extended ca simulation model. we then use the new ca model to investigate some issues in hiv/aids epidemics. most models, including the foregoing ca models, have some limitations that fail to consider the peculiarities of hiv/aids epidemics and are thereby incapable of describing the epidemic accurately and completely (see frauenthal 1980 for more discussion). first, most of the models assume that there is no latent (or incubation) period. however, for some epidemics, especially aids, there are variously lasting periods of latency and incubation as well as behavior-varying infectivity (or susceptibility) during these periods. in fact, the development of aids involves a few stages in which an infected individual can exhibit different behaviors. those diversified behaviors, in turn, have some ignorable effects on the dynamics of hiv/aids. in light of this, we extend the conventional division of epidemic process (i.e., susceptible, infection, and removed) by dividing the infection period into three sub-stages, each corresponding to the clinical stage occurring in the course of aids development. due to the inability of classical ca approaches to accommodate those newly added state transitions events, we also borrow some ideas from discrete-event simulation techniques and make one agent's stage transitions being time-triggered instead of using some state-based transition rules. secondly, it is commonly assumed that individuals in the population are homogenous in the sense that they have equal infectivity and susceptibility, or they can exert the same influence on each other, etc. this assumption may be satisfied in commonly observed epidemics but not consistent with the hiv/aids epidemic. as we know, susceptibility and infectivity heavily depends on individuals' behavior. for examples, safe sex practices such as the use of condom could dramatically reduce the chance of infection. also, the way of hiv/aids transmission for one to another is various, depending on the interactions between people, and thus the probability of getting infected is determined in part by transmission routes and can be quite different between infected-male/susceptible-female and susceptible-male/infectedfemale interactions. under this assumption, the models that are confined to a single high-risk human group are not suitable in overall population cases. new models are needed to explicitly consider the complexity. therefore, we make an extension to the traditional ca model by introducing the extended definition of neighborhood and attaching some attributes to each agent such as infectivity and resistibility. we also define four types of agents that are characterized by different infectivity (and susceptibility) and various forms of neighborhood to represent four types of people in real life. in doing so, we will be able to investigate the dynamics of hiv/aids with heterogeneous groups in a realistic way. thirdly, classical ca models assume that agents in the grid are spatially fixed, that is, once an agent is placed in a cell, it does not move into another cell. this assumption is problematical because people in the real world are migratory. for instance, in china, millions of rural people leave their hometowns and seek jobs in the cities. the migration of population is a driving force for the spread of hiv/aids. ignoring the mobility of agents in epidemic models would jeopardize the creditability of the results obtained. considering this point, we incorporate agents' mobility into their behaviors. in our model, each agent is allowed to move randomly into one of its adjoined and unoccupied cells at random time intervals. recently, agent-based modeling is used in various fields to solve plenty of problems (see, e.g. zhang and bhattacharyya 2007; luo et al. 2007) . some reader might notice that our improved ca model have features that usually found in agent-based methodology. as a matter of fact, our method borrows much from agent-based simulation modeling. to make things simple, we prefer to view this model as being a ca models. this paper is organized as follows. in the next section, we present our extended ca simulation model. section 3 gives a detailed description of simulation results and analyzes some influential factors that affect the dynamical behavior of the model. section 4 concludes and points out some possible extensions and directions for future research. cellular automata have been extensively used as tools for modeling complex adaptive systems such as traffic flow, financial markets, chemical systems, biological groups, and other social systems (see e.g. gerhard and schuster 1989; gerhardt et al. 1990; weimar et al. 1992; karafyllidis and thanailakis 1997; karafyllidis 1998) . usually, a typical ca model consists of a regular two-dimension grid with a certain boundary condition and a swarm of agents living in the grid. 1 the neighborhood of an agent is defined to some (or all) of the immediately adjacent cells and the agents who inhabit in the neighborhood are called neighbors. agents are restricted to local neighborhood interaction and hence are unable to communicate globally. there are several states agents can be in at each time and an agent's state at time t + 1 is determined based on its neighbors' states at time t . the rules used in the determination of next-time states can be written as a mapping: where s is the set of states and t denotes simulation time. the mathematical properties of cellular automata have been studied in martin et al. (1984) . in our model, we consider a population of size n(t) at time t randomly distributed in a two-dimension w × w lattice. population growth rate r is fixed throughout the simulation. at each time, new agents are added to the model, and the dead removed. simulation time advances in a discrete way. the time interval (t, t + 1) is specified to represent one week in real life. this assumption makes the simulations run reasonably fast (with respect to the whole progress of epidemics) without losing any time-specific clinical properties associated with hiv/aids. explicitly modeling the post-infection progression to aids is one feature of our model compared with conventional ca models. classical epidemic models divide the closed population into three subgroups: susceptible, infective, and recovered (removed). this simplified classification is not consistent with the epidemics in real life. particularly, it is well established that an individual, once infected with hiv, undergoes roughly three clinical phrases towards the full-blown aids: (1) infected, not yet infectious, (2) infectious, not yet asymptomatical, and, (3) symptomatical (may and anderson 1987; may et al. 1988) . the lifetime of an individual should cover not only the process from health to infection, but also the sub-stages after infection. thus, we assume that each agent can go through the following states: • s1: healthy state, initially, each agent is set to be in s1 state. healthy agents have no risk of being infected. when a healthy agent moves into the neighborhood of an infectious one or an infectious agent approaches him, the healthy agent's state will change from s1 to s2 because contacts with infectives incur the danger of infection. as for an agent in s2 state, it can transit in two directions: one direction is to change from s2 back to s1, after all its infectious neighbors move away (or its dead neighbors are removed from the grid) or he leaves the neighborhoods of its infectious neighbors; the other direction is to change from s2 to s3 if he unluckily get infected. note that we assume infection is instantaneous, i.e., instantaneous transmission from an infected individual to a susceptible. a newly infected agent is unable to transmit hiv virus until seroconversion. the s3 state corresponds to the early stages of hiv infection. let t 1 denote the duration of this period. empirical works have been done to estimate the parameter. , anderson and medley (1988) report t 1 to lie between 40 and 60 days in transfusion-induced aids cases. in our model we assume that t 1 is a random variable following a normal distribution with the mean μ 1 and the variance σ 2 1 . after t 1 , the infected agent enters s4 state: infectious state. medically, the duration during which an infected is infectious but not yet symptomatic is called incubation period. we let t 2 donate the period. empirical work suggests an average incubation period of around 4 to around 15 years (medley et al. 1987 . a weibull distribution are commonly used to describe this incubation period (see, e.g., anderson 1988; anderson and medley 1988) . furthermore, , anderson and medley (1988) estimated t 2 with a weibull distribution (with a mean of 7.7 years and a median of 7.4 year) based on 545 transfusioninduced aids cases. for simplicity, we take t 2 as a real number drawn from a normal distribution with the mean μ 2 and the variance σ 2 2 rather than a weibull distribution. it should be pointed out that the simulation results generated with the normal distribution here are proved to have little, if any, difference compared with those generated when a weibull distribution is employed. during the t 2 period, hiv viruses in the victim's body are constantly cloning themselves and eventually the immunity system collapses. at this point, the victim starts to show some symptoms and thus transit to s5 state: symptomatic stage. as usually, let t 3 denote the duration of this period. rothenberg et al. (1987) report 5-7 year survival rates among 1660 idus (intravenous drug user) in new york city and find a median time of survival of 282 days. chang et al. (1993) report a median survival time of 10.5 months. empirical work shows that almost all hiv infectives, excluding those who die from other causes, will inevitably develop aids and die of it (may and anderson 1987) . similarly, we assume t 3 follows a normal distribution with the mean μ 3 and the variance σ 2 3 . eventually, the ill agent enters s6 state after t 3 passes by agents in s6 state will be removed from the population at the beginning of the next time and all of their uninfected neighbors will be released from s2 state, back to s1 state. generally, these state transitions take place in the order of s1, s2, s3, s4, s5, and s6. it is impossible for an agent to return from s3 state to s1 or s2 state. this backward transition s2 to s1, demonstrated by a dashed line in fig. 1 , is due to the disappearance of threats posed by infectious agents. moreover, although the transitions among s3, s4, and s5 state are not relevant to the propagation process, this process is closely related to hiv/aids transmission. taking account of this procession is essential for a better understanding of hiv/aids transmission. in the model, all the events triggering transitions could be divided into two categories. one category is a rule-based, such as healthy-to-dangerous, dangerous-to-infected, and dangerous-to-healthy state-transition events. these events occur according to the ca transition rules: an agent's state at time t + 1 is based not only on its own state but also on the states of its neighbors at time t . the other category is time-based, meaning that these events are scheduled at pre-specified times. for instance, an agent entering s3 state will be assigned a time indicating when to change to s4 state. after that amount of time elapse, the transition occurs spontaneously. despite the distinction between these two categories, the subtlety of implementing the two event-triggering mechanisms is very trivial and leaves no further elaboration necessary. actually, these six states can be divided into three "super" classes in terms of the taxonomy used in the classical mathematical models: s1 and s2 states correspond to the susceptible state; s3, s4, and s5 states belongs to the infection state, and s6 state is the removed state. obviously, s1 and s2 state could be treated as one single state without changing any results. the reason why we divide this into two sub-states is that this makes our model easily implemented and our logic decent and legible. hiv/aids epidemic differs from other epidemics in that its dynamics is heavily affected by individual's behavioral patterns and the interactions between them. for example, careful sex practices and sanitization measures in drug taking will make individuals less likely to be infected. behavioral patterns and interactions are mostly determined by individual's life styles, personalities, social networks, etc. however, the majority of models fail to take account of the heterogeneity in agents' behaviors. to capture this, we extend classical ca models by allowing each agent to have its own attributes such as mobility, infectivity, resistibility (susceptibility) 2 and different extent of neighborhood. assume that each cell in the grid can be occupied by at most one agent at a time. at time t , agent i can move from one cell into one of its adjacent cells with probability p m i . here, p m i is a measurement of agent i's activity level. it is a fixed real number, drawn from a uniform distribution (p m min , p m max ) (0 ≤ p m min ≤ p m max ≤ 1). when p m min = p m max , the activity level across agents is equal and therefore agents have the same inclination to move around. one extreme case is p m min = p m max = 0, which corresponds to the situation in which agents stay in their initial places during simulation, whilst p m min = p m max = 1 means that each agent will move into one empty neighboring cell at almost each time (he could get stocked and not move anywhere if it's neighborhood is occupied). it is easy to induce that the average time per move is calculated as 1/p m i . intuitively, high level of activity leads to speedy spreading. our simulation results verify this. besides the heterogeneity in agents' activity, another kind of heterogeneity is introduced when we assign various levels of infectivity and susceptibility to agents. let f i denote the infectivity level of agent i. f i is a real number drawn uniformly from the interval (0, 1). it measures the possibility that agent i transmits hiv viruses to others when they meets. evidently, greater values of f i indicate higher infectiousness of agent i. suppose also that each agent has resistance to being infected. we denote this resistibility as r i for agent i. similarly, r i is also a real number drawn uniformly from the interval (0, 1) and has the property that the greater the resistibility, the less is the chance of getting infected. note that the infectivity of an agent need not be a constant. an agent can have different level of infectivity, depending both on its state as well as on its behavior. it is widely believed that infectives experience two periods of high infectivity (see e.g. may and anderson 1987; may et al. 1988 ), one shortly after being infected and the other at the late stage of his illness. another example is that a patient might have high infectivity during the incubation period and low infectivity owing to good health care during the symptomatic period. although our model allow for various infectivity at different stages for a single agent, we adopt the fixed infectivity for each agent. in doing so, we can focus our attention on some significant issues. we leave various infectivity scenarios for future work. conventional ca models define two types of neighborhoods: moore neighborhood and von neumann neighborhood. in this paper, we extend the concept of ca neighborhood in order to better describe various situations encountered in agent-based modeling. figure 2 illustrates the definition. as we can see in fig. 2b shows the classical moore neighborhood, and fig. 2d classical von neumann neighborhood. figure 2c represents an extended moore neighborhood with the order of 2 × 2, and fig. 2e an extended 2 × 2 von neumann neighborhood. specially, fig. 2a can be simply viewed as an extended 0 × 0 moore (or von neumann) neighborhood. note that fig. 2b , f-i have the neighborhoods with one direction. this directional structure is able to capture the biases or preferences embedded in individuals' behavioral patterns and we can use different directions to represent variable ways of interactions. it is easy to see that the greater is the neighborhood, the larger extent to which an agent can exert its influence to its neighbors. given the above neighborhood definition, a concept of distance is naturally induced. let a pair of integer numbers (x, y) represent an agent's coordinates in the grid. the distance between agent i and j is thus given as (2) next, we specify that the influence indicator m i,j of agent i and j satisfies the following condition: that is to say, the influence intensity is inversely proportional to the distance between them if the influence can be exerted, and zero otherwise. therefore, the infective impact i i,j of agent i on agent j can be expressed as and the probability of an agent infected by one of its neighbors is defined as: where p(·) is a function satisfying the conditions: (1) p(·) is a real-valued function with the value between zero and one; (2) p(·) is increased in i i,j and decreased in r i . in this model, we assume p(·) takes the form of the following equation: here, p(i i,j , r i ) is interpreted as the probability of agent i being infected by its neighbor j . denoted by b i the set of all its neighbors, agent i's overall probability of infection is thus given by it indicates that this overall probability is determined by the most influential neighbor. such specification makes sense in most cases. the model developed in sect. 2 is implemented using java programming language with the repast software package. 3 detailedly commented source code is available from the authors upon request. next, we begin our analysis by considering first a typical simulation run as a benchmark case. 3.1 benchmark case table 1 lists the input parameters chosen for the benchmark case. in this case, the grid consists of 100 × 100 sites (w = 100). the population size n is set to 2000 with the initial infected ratio α = 0.005. all agents are homogeneous in terms of having the same infectivity f i = 0.2, resistibility r i = 0.5, and 1 × 1 moore neighborhood. they are uniformly distributed in the grid. the total simulation time t for each run is set to 2000. figure 3 shows a snapshot of the spatial distribution of the population at some time in a typical simulation. it is commonly believed that as the epidemic develops, its spread ends up with two typical situations: extinction or prevalence. figure 4 depicts these two situations. in fig. 4a , the number of infections 4 climbs early in the process. after about time t = 470, the infection level starts to drop slowly until it reaches zero at time t = 2000, while in fig. 4b , the number of infections increases slowly and reaches an equilibrium level after t = 1600. the intuition . 3 the spatial distribution of the population at time t = 1275 behind is that in the first case, newly infected continuously enter the pool of infectives at a fairly low rate in early stages. after the lengthy incubation period, these infectives begin to develop aids and eventually die. the total number of them drops when the infection rate is very low with regard to the rate at which infectives leave the pool for some reason (dead in the model). while in the second case, healthy agents get infected at a relatively high rate in early stages. the infection level continues to increases because the number of removed agents is relatively small in later stages. thus high infection rate often leads to prevalence as demonstrated in fig. 4b . these two results can be found in real-world situations. notice that fig. 4 the infections vs. time result in two simulations in the parameters setting, the growth rate r is almost zero (r = 0.0001). in next subsection, we will explore the effects of various factors, such as population density, initial infection ratio, and infectivity, on the epidemic. now we keep other parameters constant as before and let the population density β vary to see how β (β = n/w 2 ) affects the dynamics of hiv/aids transmission. tarwater and martin (2001) investigate this issue when studying the outbreaks of measles or measles-like infectious diseases. as one would expect, many common infectious diseases spread more rapidly at a high population density than at a low population density. figure 5 illustrates the time series of the mean numbers 5 of infectives for different population sizes n = 1500, 2000, 2500, 3000, and 3500. we can see that when population density is relatively low (n = 1500, 2000) , the infection levels are relatively low during the entire simulation and decline slowly in the later stages. this suggests that the epidemic died out eventually. in the and 400 at last. for n = 3000 and 3500, the infection numbers reach to a very high level and then drop rapidly. the collapse is because that so many infectives are removed from the model that the pool of infectives shrinks. for clarity, we also plot the fractions of infectives in the population vs. time in fig. 6 . clearly, in late stages of the epidemic, the fractions are greater when β is great than when β is small. in summary, hiv/aids infection is more likely to persist at higher population densities. this is due to that with the population density increasing, the population contact rate rise, leading to increases in the probability of infection. early work (see, e.g., rhodes and anderson 1996) suggests that there is a threshold below which the epidemic would eventually dies out and above which it would persist. due to the limitation of ca methods, it is diffifig. 7 mean number of infections vs. time for different initial infected ratios cult to pin down its exact value. however, with many simulation runs, we still can give an approximate interval in which the threshold lies. an interesting question one may ask is how epidemic spreading is affected by initial configurations of susceptibles and infectives, or whether the multiple infection sources will make the disease more likely to become endemic. with other parameters fixed as before, we run the simulations with α = 0.002, 0.004, 0.006, and 0.008, respectively. figure 7 presents the result. clearly, as α increase, the infection level shifts upwards. in the case of α = 0.004, the infection level reach 300 at time t = 2000, higher than 200 in the case of α = 0.002. by contrast, the level in the case of α = 0.006 climb to around 380 at time t = 1000 and drops slightly to 340 at time t = 2000. the maximum infection is reached in the case of α = 0.008 at time t = 1000, which is more than 440. spatially, more sources of infection imply greater chance of being infected within a certain area, letting hiv/aids epidemics to be more likely to spread out and persist. statistically speaking, an individual's probability of infection is generally proportional to the number of infectious sources. intuitively, the more migratory the population, the more likely that an epidemic is to spread. suppose an agents' activity can be measured by the number of contacts it makes with others within a unit of period of time. as a result, our model assumes that individuals' activity is measured by mobility. later stages. this is, in the case of p m max = 0.005, the level is above 120 and in the case p m max = 0.007, the level is in the range (180, 210). in the last case of p m max = 0.009, the infection level fluctuates above 300, higher than those of other cases. so we conclude that mobility plays a significant role in the dynamics. it could explain why chinese government took rather strong measures to control the migratory people and quarantine the infectives or the suspects during the outbreak of sars in the spring of 2003. as to our model, if agents are configured with higher mobilities, it is more likely that the hiv/aids infection can persist in the population, whilst if configured with lower mobilities, the infection would gradually diminish and eventually die out. next, we are to examine how neighborhood forms affect hiv/aids epidemic dynamics. the parameter sets are kept the same as in benchmark case, except for the adoption of different neighborhood forms. figure 9 illustrates the simulation results generated in two cases: one with 1 × 1 von neumann neighborhood and the other with 1 × 1 moore neighborhood. in the case where the von neumann neighborhood is used, the level of infection goes up to about 140. it is clearly greater than that of the case with 1 × 1 moore neighborhood in which the level only reach about 100. such result suggests that with wider neighborhood, an agent is more likely to get influenced by its neighbors and therefore the likelihood of getting infected increases accordingly. it is easy to induce that the infection level rises with the order of neighborhood. this result also suggests that hiv/aids epidemic dynamics is significantly affected by strong interactions between agents. we now turn to examine heterogeneous mixing, i.e., different at-risk groups coexist. usually, heterogeneous mixing will make the dynamics more complicated and unpredictable. in the following, we assume the whole population is divided into four different groups as shown in table 2 . as shown in table 2 , class p0 has very low infectivity, low susceptibility, and 1 × 1 von neumann neighborhood. it can represents children and (or) elders in the population who hardly infect others and are easy to be infected. class pl refers to ordinary people who have relatively low infectivity and high resistibility (therefore low susceptibility). in our model, this class amounts to a large fraction of the whole population. class ph and class ph+ can represent the two high-risk groups observed in real life. agents of class ph have high infectivity and low resistibility duo to their high-risk behaviors like incautious sex without protection, needle sharing, unhygienic blood transfusion and so on. the biased 0 × 1 (or 1×0) von neumann neighborhood captures their potential oriented or biased behaviors. in contrast, agents of class ph+ with both higher infectivity and higher susceptibility represent those who are, although being in the minority, the most dangerous and malevolent group. such group does exist in real life. for instance, some crimes were reported in china in recent years that a few aids infectives intentionally have sex with innocent people or shot people with contaminated syringes in public places. they blame their infection on the society and the government for not being able to provide necessary health service and compensating too little. the 2 × 2 moore neighborhood indicates their intensive influence on others. we distinguish class ph+ from class ph in order to see whether such malevolent behaviors have significant impacts on the spread of hiv/aids and to what extent. while such rough classification may be incorrect or even erroneous, it surely is reasonable and well supported by our extended ca model. table 3 gives the values of f i and r i used in the following simulations. as we will see later, the results generated with this classification are fairly consistent with those obtained through empirical work. figure 10 gives a typical simulation result in the four-class scenario. the fractions of four classes here are: n p0 = 0.1, n pl = 0.7, n ph = 0.1, and n ph+ = 0.1. the infection curve in fig. 10 is quite similar to that obtained in the single-class scenarios except that its level is fairly higher. in this section, we will investigate the effect of agents' susceptibility on the dynamics of hiv/aids. given n p0 = 0.05, n pl = 0.8, n ph = 0.1, n ph+ = 0.05 and others as before, let r ph+ vary. figure 11 gives the infection levels when r ph+ is set to be 0.6, 0.7, 0.8, and 0.9, respectively. as we can see, these equilibrium infections are almost at the same level, which is inconsistent with our expectation. the differences are so small that we cannot assure with confidence whether changes in susceptibility have impact on the epidemic dynamics. the possible reasons, we believe, are twofold: first, the role played by susceptibility may be not as decisive as the above factors. second, we may describe susceptibility in the wrong way and make it an inessential factor in our model. future work will reconsider this issue and find the better way to describe susceptibility. at last, we will examine whether changes in the fractions of some classes can affect epidemic behaviors. given n p0 = 0.1, n pl = 0.7 and other parameters as before, we take n ph = 0.12 (n ph+ = 0.08), 0.14 (0.06), 0.16 (0.04), 0.18 (0.02), 0.20 (0.0), respectively. figure 12 shows the infection levels in these five combinations. as observed from fig. 12 , we obtained very similar results, compared with those in the foregoing analysis. with n ph+ decreasing, infection level declines. recall that ph+ has more influence on its neighbors than ph, thus leading to greater transmission to others and larger infection rate. this finding suggests the government should pay more attention to those who have high-risk life styles and revengeful behaviors. this makes it the essential issue of how highly infectious and malevolent individuals are restricted and controlled. the focus of the paper is on the modeling of the entire course of hiv/acid epidemics and heterogeneity in agents' behaviors. even though classical ca models are capable of describing the spread of common epidemics but fail to represent the complicated epidemics like hiv/aids disease. the ignorance of heterogeneity gives rise to unacceptable errors in the prediction of the development trends. in addition, components of a conventional ca system such as topological forms of grid, the definition of neighborhood, and state transition rules are simple and unchanged over time. this makes the modeling of complicated dynamics such as hiv/aids transmission difficult and uncontrollable. in this paper, we have developed an extended ca model to capture key epidemiological and clinical features of hiv/aids epidemic. first, we explicitly models and simulates the whole progression of hiv/aids disease (i.e., infected but not infectious, infectious but asymptomatic, symptomatic, and deceased). such improvement can give us a better understanding of the dynamics during the entire hiv/aids epidemics. in order to examine various degrees of influence between agents, we have introduced an extended definition of neighborhood to represent the intensity and bias of influence. this lets us gain insight into how various degrees of interactions affects the hiv/aids epidemics. another type of heterogeneity in disease-related attributes such as susceptibility, infectivity and durations of epidemic phrases is also taken into consideration. moreover, we also consider the effect of agents' mobility on epidemics dynamics. given all the improvements, we have obtained richer simulation results similar to those usually found in the mathematical models or other classical simulation models. we have identified some influential factors that greatly affect the hiv/aids epidemic dynamics. the main findings are that 1) hiv/aids epidemic can end up in the two regimes: extinction and persistence; 2) with these factors such as agents' mobility, population density, initial infection ratio, and the extent of neighborhood increasing, the infection level get higher. after crossing some critical point, the regime generated could change from dying-out to persistence at some point. this result is robust across many of the tested parameter combinations; 3) in four-class scenarios, the great fraction of 'super' infectives (the ph+ class in our model) can also produce higher level of infection. however, our simulation study above is still preliminary. there are some issues needed to be addressed. first, as said before, we should redefine susceptibility in a better way to check its role in the dynamics of hiv/aids epidemic. second, most models posit that a virus carrier's infectivity is constant during the progress of a disease. however, this is not the case for hiv/aids epidemic. various infectivity at different stages could have substantial impact on the dynamics of hiv/aids transmission. this problem needs special attention. additional, further developments of our model, e.g. by adding age-related structure (griffiths et al. 2000) , different subgroup classification, and other heterogeneity, would greatly add to the appeal of this model. with these additions, a better understanding of hiv/aids and thorough empirical work are required. finally, a natural extension of the model is to include the assessment of various control policies and managerial strategies, and this will be a firm support for the decision-making in prevention programs against hiv/aids. on modeling epidemics. including latency, incubation and variable susceptibility the epidemiology of hiv infection: variable incubation plus infectious periods and heterogeneity in sexual activity infectious diseases of humans: dynamics and control possible demographic consequences of aids in developing countries epidemiology of hiv infection and aids: incubation and infectious periods, survival and vertical transmission a simulation model of the dynamics of hiv transmission in intravenous drug users a comparison of simulation models applied to epidemics the mathematical theory of infectious diseases and its applications dynamics of hiv infection on 2d cellular automata models for transmission of disease with immigration of infectives survival and mortality patterns of an acquired immunodeficiency syndrome (aids) cohort in new york state mathematical modeling in epidemiology cellular automata and epidemiological models with spatial dependence a heterogeneous cellular automata model for sars transmission a cellular automaton describing the formation of spatially ordered structures in chemical systems a cellular automaton model of excitable media a simple agent model of an epidemic an age-structured model for the aids epidemic the differential infectivity and staged progression models for the transmission of hiv a model for the influence of the greenhouse effect on insect and microorganism geographical distribution and population dynamics a model for predicting forest fire using cellular automata. ecological modelling a contribution to the mathematical theory of epidemics the dynamics of hiv spread: a computer simulation model flood decision support system on agent grid: method and implementation algebraic properties of cellular automata transmission dynamics of hiv infection the transmission dynamics of human immunodeficiency virus (hiv) [and discussion incubation period of aids in patients infected via blood transfusion the distribution of the incubation period for the acquired immunodeficiency syndrome (aids) mathematical biology i experiences creating three implementations of the repast agent modeling toolkit persistence and dynamics in lattice models of epidemic spread epidemic thresholds and vaccination in a lattice model of disease spread survival with the acquired immunodeficiency syndrome. experience with 5833 cases in new york city effects of population density on the spread of disease diffusion and wave propagation in cellular automaton models of excitable media modelling the hiv epidemic: a state-space approach effectiveness of q-learning as a tool for calibrating agent-based supply network models we would like to express our gratitude to the many people in xi'an jiaotong university who have participated in planning, data collection, and presenting the results. without their efforts, this simulation modeling would not have been possible. this work is supported by nsfc under contract 70601023. we are grateful to ting zhang and jie huang for excellent research assistance and the referees for many helpful comments that greatly improved the presentation. key: cord-017096-pnxjrtgo authors: zhang, pingping title: application of upt-poct in anti-bioterrorism and biosecurity date: 2019-09-20 journal: principles and applications of up-converting phosphor technology doi: 10.1007/978-981-32-9279-6_16 sha: doc_id: 17096 cord_uid: pnxjrtgo with the exception of toxins, bioterrorism agents are mainly microorganisms, many of which cause serious infectious diseases. up-converting phosphor technology-based point-of-care testing (upt-poct) can detect bioterrorism agents from various samples with high sensitivity and specificity, in particular it shows robust performance for complicated samples, such as food, powder, viscera and grains. the tolerance of upt-poct to sample is based on the physical and luminescence stability of ucnps, the stable covalent interaction between ucnps and antibody, as well as the strong buffering capacity of the detection system. reliable results can be obtained in a short time period using a portable biosensor by nonprofessionals owing to the simple nature of upt-poct operation and sample pre-treatment. based on pathogenicity, bioterrorism agents can be categorized into lethal and incapacitating agents. lethal bioterrorism agents have high mortality, for example, the mortality of septicemic plague and pulmonary tularemia can reach 90% and 60%, respectively, in the absence of antibiotic treatment. y. pestis, b. anthracis, f. tularensis type a, b. pseudomallei, yellow fever virus, smallpox virus, rickettsia rickettsii, chlamydia psittaci, and botulinum toxin are all lethal bioterrorism agents. incapacitating bioterrorism agents can make people defenseless, examples include brucella spp. and coxiella burnetii. bioterrorism agents can be transmitted by air, food, and water, and some can be transmitted person-to-person, such as y. pestis (begier et al. 2006) , b. anthracis, and ebola virus. the bioterrorism agents disseminated through air can be made into aerosols and threaten public health on a large-scale, examples include rickettsia rickettsii, y. pestis (agar et al. 2009 ), and b. anthracis (estill 2009 ). some zoonotic pathogens can infect people through contact with infectious animals during slaughter and leather treatment, and animal husbandry, as well as contact between people (begier et al. 2006) . transmission through food includes consumption of the meat or milk of infected animals, for instance, people can be infected by v. cholerae through contaminated seafood (finelli et al. 1992) . water is also an important transmission medium and is the main transmission route for v. cholerae (hill et al. 2011 ). in addition, some agents, for example, b. anthracis spore can survive in the silt at the bottom of a riverbed for decades, and f. tularensis can survive in the cold water of a river for months (chitadze et al. 2009 ), seeking the chance for outbreak. transmission media arthropods such as mosquitos, flies, lice, mites, and ticks, are widespread in nature. the transmission media of y. pestis, f. tularensis, yellow fever virus, and rickettsia przewalskii are flea, tick, mosquito and pediculus humanus corporis, respectively. the suspicious incidents caused by bioterrorism agents can easily lead to public panic because of their perniciousness. the misdiagnosis of the diseases caused by bioterrorism agents, rapid deteriorations for acute and serious infectious diseases, and the limitations of therapeutic means, all highlight the importance of preventing the release of bioterrorism agents. low pathogenic dose, diversity of pathogenic types, high mortality or disability rate, and the potential for widespread dissemination, are all characteristics of bioterrorism agents, as well as strong adaptability to the environment in the exposed zone resulting in long-term threat to public health. several microorganisms can multiply rapidly in vivo, causing serious diseases, for instance, the infectious dose of f. tularensis type a is less than ten live cells. bioterrorism agents attack the respiratory tract, digestive tract, skin, blood, and glands. many of the agents cannot be handled by the immune system, and y. pestis can even survive, proliferate and spread in macrophages (lukaszewski et al. 2005) . acute and serious diseases occur after a short incubation period, often only several days. for instance, the symptoms of pneumonic plague include high fever, cold shivers, cough, chest pain, hemoptysis, and dyspnea, followed by serious poisoning symptoms and death in two to four days. the mortality of pneumonic and generalized plague is up to 30~60%. brucella spp. with anti-phagocytic capsules can proliferate in the lymphatic system and then enter into the blood resulting in toxemia, and its ability to evade elimination by the body results in long-term joint pain, fatigue, and disability, dramatically decreasing the quality of life of the patient. the most catastrophic potential outcome is a worldwide epidemic. three historic large-scale plague epidemics caused 160 million deaths (prentice and rahalison 2007) , seven historic cholera epidemics involved 100 countries, and brucellosis is prevalent in 170 countries. diagnosis can be based on epidemiological history, clinical syndromes, etiology, and serology detection. epidemiological history includes residency in the epidemic areas or entrance into these areas in the last two weeks before morbidity, bites by arthropods, and contact with or consumption of infected products or water. clinical syndromes of some bioterrorism agents appear in isolation, however agents with many infection routes can cause various clinical syndromes. for example, the main clinical syndrome of cholera is diarrhea, however at least six syndromes have been found for tularaemia, including bubonic, pneumonic, gastrointestinal, and systemic (typhoidal and septicemic) tularaemia. the similarity of clinical syndromes between diseases caused by bioterrorism agents and by other factors, makes accurate pathogenesis diagnosis more difficult for doctors when the etiology is unknown. for example, patients with fever caused by f. tularensis are easily misdiagnosed with influenza (simsek et al. 2012) , while there are no differences the in clinical syndromes between pneumonia caused by f. tularensis or that arising from other causes (stralin et al. 2002) . misdiagnosis and delayed treatment are responsible for lack of safeguard implementation and the subsequent dissemination of bioterrorism agents in medical institutions. therefore, etiology detection is critical for determining and preventing infectious disease. however, the low numbers of pathogens in the initial stage of a disease are difficult to find by etiology detection, fortunately early diagnosis based on the detection of antibodies in blood is a plausible approach in practice. because of the great threat of bioterrorism agents to the lives of patients, essential therapy must be administered in time to avoid death and poor prognosis. for instance, water and electrolytes must be administered to patients infected by v. cholerae in time because severe dehydration and failure of microcirculation caused by diarrhea often occur during the progression of cholera. the prognosis is usually poor owing to the limited therapeutic means. antibiotics are often excessively applied for saving the lives of patients infected by bacterial bioterrorism agents, such as streptomycin specific for plague and many antibiotics that are effective for tularaemia, however the side effects of this therapy method are osteoporosis and joint injury. eliminating infection sources, cutting-off transmission routes, and protecting susceptible patients, are all methods for prevention and control of the spread of bioterrorism agents. quarantine is essential for infectious bioterrorism agents, including isolation of prime areas of disease outbreak, quarantine of patients and people who have gone to countries with epidemics, conflagration and deep interment of cadavers, use of disinfectant, and correct treatment of material from patients with high-temperature and high-pressure. individual and environmental defenses should be enhanced when nursing and treating patients or infected animals, as well in the resulting treatment of cultures in scientific research. pre-inoculation of vaccines is a good prevention method for protection of people in natural foci, doctors and scientific staff, for example, vaccines or attenuate strains of y. pestis, f. tularensis, and b. anthracis can be inoculated by cutaneous scarification. however, no effective vaccine has been obtained for some bioterrorism agents, such as chlamydia psittaci. there are some effective virus vaccines, such as vaccinia vaccine (against smallpox virus), rabies vaccines, hepatitis b vaccines, and hantavirus vaccine. however, mutated viruses often emerge as rna viruses, especially for retroviruses such as human immunodeficiency virus, leading to some vaccine failure and the necessity for preparation of new vaccines against the mutated virus. owing to the high pathogenicity of bioterrorism agents, as well as the high transmission capacity of microorganisms, essential isolation and protection measures must be carried out to ensure biosecurity during the process of detection. prompt and accurate testing favors the detection of suspicious substances, so that an emergency signal can be issued or a false alarm can be revealed. in addition, multiplexed detections can be applied to unknown pathogens to improve efficiency and reduce environmental contamination caused by excessive operation. protective measures during the detection process are essential to ensure individual and environmental security and prevent the spread of bioterrorism agents. specialized laboratories for pathogenic microorganisms, or detection vehicles on site equipped with such laboratories are necessary for lethal bioterrorism agents. for detection on site, operators should wear rubber gloves, surgical masks, hats, and protective clothes. isolation for patients and infected animals, and blockades of the contaminated region are essential. the screened suspicious matter should be sent to the laboratory for further identification in a biohazard marked container that is waterproof, breakage-proof, and high temperature and pressure resistant. according to the national standard for laboratory security, the defense levels of bio-laboratories are from 1 to 4. experiments involving the most infectious bioterrorism agents must be performed by professionals using a biosafety cabinet in a level 3 or 4 laboratory. the measures for infection prevention in the laboratory include pre-inoculation with vaccine, wearing protective clothes, using biosafety cabinets for culture and infective material, high-pressure treatment for growth medium and contaminated gloves, as well as sprinkling disinfectant in the biosafety cabinet. the detection limit, specificity, and tolerance of detection methods determine their accuracy. the detection limit for bioterrorism agents must be very low, ideally single cell, because of their high pathogenicity in low dose and the high capacity for proliferation of some microorganism bioterrorism agents. pre-incubation of microorganism bioterrorism agents to increase detection rates is not permitted. specificity is particularly critical for a detection method because there are incalculable microorganisms and other organisms in nature. no specific reaction should occur for substances or strains that share structural similarity, close genetic relatedness or similar transmission routes with the targets. a detection method must be available for different specimens within appropriate operational-error to ensure the stability. various fresh or decomposed animal tissues obtained in natural foci and the white powders used by terrorists for concealment, such as flour, milk powder, and putty powder, are representative of the complicated specimens that the detection method will be confronted with. the errors brought by non-professional operation are also considered to ensure the detection accuracy in practice. in addition to safety and accuracy, detection time should also be shortened. a short detection time is beneficial for therapy, cutting-off transmission routes, preventing the dissemination of contamination, and promptly eliminating negative social effects. multiplex detection has been developed to improve detection efficiency, and reduce sample volumes and operational handling. compared with the detection of individual targets, multiplex detection reduces the workload and shortens detection times. in addition, multiplex detection can give results for many targets at once, therefore smaller sample sizes are sufficient, which is especially important for precious low-volume samples. less operational handling minimizes the possibility of contamination of the inspector and environment with bioterrorism agents. because of difficulties arising from multiple reactions in one system and simultaneous multiple signal extraction, multiplex detection still shows unsatisfactory performance in stability, anti-interference, and reproducibility. despite these shortcomings, multiplex detection is urgently required for identification of unknown pathogens and simultaneous detection of multiple pathogens in one sample because multiple bioterrorism agents can be released in one sample, for example, coxiella burnetii, chlamydia psittaci and influenza virus have been combined into aerosols and applied for biowarfare. bio-threats are becoming increasingly serious with developing technology because of reductions in cost, and improved transmission and operation. many bioterrorism agents can be easily obtained from nature, such as zoonotic microorganisms from natural hosts, and ricin and abrin, which are easily prepared from plants. modern fermentation technology promotes mass-production of microorganism bioterrorism agents (yang 2008), while many toxins can be synthetized using chemical methods. water, air conditioning systems, food, and letters have been used as the vectors for bioterrorism agents in bio-attack incidents, and the agents can appear in the form of powders and aerosols, amongst others. the detection of bioterrorism agents includes surveillance in natural foci and treatment for public health emergencies. surveillance, especially for common infectious disease, is an efficient measure for preventing disease outbreaks. the objects of the monitoring are different depending on the sort of disease. for example, rodents and fleas are natural reservoirs for y. pestis (prentice and rahalison 2007) ; water and plankton are the natural environment and reservoir of v. cholerea respectively (huq et al. 1995) ; and b. pseudomallei can survive in tropical and subtropical natural foci (draper et al. 2010) . public health emergencies, such as bioterrorism attacks, laboratory releases, collective food poisoning, and concentrated outbreaks of cases, require prompt responses based on the detection results. according to the definition specified by the committee of poct, chinese association of medical equipment, point of care testing (poct) is a detection method implemented on site, and it reports results in a short time period using portable analytical instruments and associated reagents. to satisfy the strict requirements of poct, a detection method must be sufficiently rapid, sensitive, and specific. screening detection on site mainly relies on poct methods. the short response time for surveillance in natural foci and during public health emergencies has provided powerful support for the prevention of disease outbreaks. the minimal operation procedures that are a feature of poct methods make poct applicable for detecting bioterrorism agents to reduce the release of the agents. immunochromatography is well known as a poct method for onsite screening of bioterrorism agents because it can give results from simple sample loading and the used strip can be directly disposed of after treatment with high temperature and pressure. the sensitivity and specificity of traditional immunochromatography methods are often too low because the physical interaction between antibodies and gold particles is fragile, and the results are analyzed by the naked eye. up-converting phosphor technologybased point-of-care testing (upt-poct) as described below are based on the covalent conjugation of upconversion nanoparticles (ucnps) and antibodies, and then the emission signal at 540 nm resulting from excitation at 900 nm can be transmitted into readable electrical signals by biosensors. therefore, the upt-lf assay can detect various samples with high sensitivity and specificity, in particular it shows robust performance for complicated samples. the detection mode of upt-poct for bioterrorism agents hinges on the molecular size of the detection target. double-antibody sandwich mode is used for detection of macromolecules, while competition mode is used for micro-molecules. in addition, the principle of the double-antigen sandwich mode is in parallel with that of the double-antibody sandwich mode for utilization of macromolecules, mainly antibodies. upt-poct based on the double-antibody sandwich mode is mainly applied for the detection of bacterial antigens or protein toxins, such as y. pestis (yan et al. 2006) , brucella spp. (qu et al. 2009 ), b. anthracis spore (li et al. 2006) , f. tularensis (hua et al. 2015b) , b. pseudomallei (hua et al. 2015a) , v. cholerae (hao et al. 2017 ), e. coli o157 (wang et al. 2007 ), ricin , and abrin . while upt-poct based on the double-antigen sandwich mode is applied for the detection of antibodies against bioterrorism agents, such as antibodies against y. pestis (hong et al. 2010 ) and hepatitis b surface antibody (li et al. 2009 ). for example, in the detection of y. pestis ( fig. 16.1) ; mouse anti-y. pestis antibody 1 is immobilized on the nitrocellulose membrane as the test (t) line, while mouse anti-y. pestis antibody 2 is combined with ucnps covalently. for positive detection, y. pestis in samples is captured by ucnps-mouse antibody 2 conjugates, and then flows forward and is captured by mouse antibody 1 at the t line of the nitrocellulose membrane, forming t line-mouse antibody 1-y. pestis-mouse antibody 2-ucnps conjugates. ucnps can emit visible light signals after excitation by infrared light, and the intensity of the signals is in direct proportion to the concentration of y. pestis. whether there are y. pestis present or not, goat anti-mouse antibody-mouse antibody 2-ucnps conjugate will be formed on the control (c) line on the nitrocellulose membrane for quality control. the signal ratio between the t and c lines, the t/c ratio, is defined as the detection result, and t/c values increase with the increase in quantity of y. pestis in the sample. in the double-antigen sandwich mode, bioterrorism antigens, such as f1 antigen of y. pestis, are dispensed in the t line and combined with ucnps respectively, and then the corresponding antibody is detected. competitive mode is used to detect micro-molecular matter that is too small to be detected by double-antibody sandwich mode. upt-poct based on competitive mode is applied for the detection of mycotoxin, such as aflatoxin b1 (afb1) , aflatoxin m1 and t-2 toxin. using the detection of aflatoxin b1 (afb1) as an example (fig. 16.2) , afb1-bsa cross-linking agent is immobilized on the nitrocellulose membrane as the t line, while mouse anti-afb1 antibody is combined with ucnps covalently. for positive detection, ucnps-anti-afb1 antibody is combined with afb1 in samples and cannot be further captured by afb1-bsa crosslinking agent on the nitrocellulose membrane, leading to the p. zhang the upt-poct assay for multiplex detection is based on a 10-channel upt-lf disc or multiple t lines for one strip. the upt-lf disc is composed of ten one-target strips, therefore ten targets can be detected from the loading of one sample. for instance, a upt-lf disc prepared with ten proteins of y. pestis based on the doubleantigen sandwich mode has been realized for the detection of antibodies against y. pestis, providing a clue for seeking diagnosis biomarker of y. pestis (hong et al. 2010) . a strip with multiple t lines can detect multiple targets, such as the upt-lf assay developed for simultaneous detection of v. cholerae serogroup o1 and o139 (hao et al. 2017 ). sensitivity and specificity are crucial to the performance evaluation for a detection method, and the evaluations of upt-poct for detection of different bioterrorism agents are shown in table 16 .1. the sensitivity of upt-poct for bacterial bioterrorism agents can reach 10 3 cfu/ml (namely 10 cfu for each test based on ten-fold sample dilution and 100 âµl of sample loading volume), and that for toxin can reach 0.03 ng/ml. the quantitative range covers four to five orders of magnitude, even six for detection of b. anthracis spore (zhang et al. 2014 ). the coefficients of variation for detection are all below 15%. upt-poct shows excellent specificity to the bioterrorism agents that share genetic relatedness, similar transmission routes, or similar structure with the targets. the tolerance to biochemical agents (table 16 .2) and operation error of upt-poct for the detection of five bacterial bioterrorism agents has been evaluated (zhang et al. 2014; hua et al. 2015a, b) . the high concentration of agent that the detection method can tolerate is defined as the tolerance limit. the tolerance limits of upt-poct for ph, ionic strength, viscosity, and bio-macromolecule concentration can reach ph 1 -13, â�¥ 4 mol/l of nacl and kcl solution, â�¤ 100 mg/ml peg 2000, â�¥ 20% glycerol, â�¥ 400 mg/ml of bsa and â�¥ 80 mg/ml of casein, respectively. at some tolerance limits, the sensitivity could be improved by one order of magnitude. the operation error, including the volume variation of the sample (from â��50 to 200%), sample treating buffer (from â��22 to 44%), and loading mixture (from â��30 to 30%), has little influence on the sensitivity and specificity of upt-poct. upt-poct shows excellent performance for detection of bacterial bioterrorism agents in simulated samples, such as food, powder, and viscera (zhang et al. 2014; hua et al. 2015a, b) (table 16. 3), and it can also effectively detect abrin and afaltoxin b1 in food and grains zhao et al. 2016) (table 16 .4). the highest tolerance limit of upt-poct to simulated samples could reach 400 mg/ml. for detection of v. cholerae in 102 field water samples obtained from sample collection sites in guangzhou city (china), upt-poct is more sensitive than the isolation-culture method and colloidal gold immnochromatography assay, and its sensitivity could match that of real-time fluorescent pcr with fewer false positive results (hao et al. 2017) . the sample pre-treatment for detection by upt-poct is merely grinding, or supernatant extraction by centrifugation after grinding and shaking on a vortex shaker for 15 min (or ultrasonicating for 10 min), and then the sample can be directly mixed with sample-treating buffer for detection. the tolerance of upt-poct to sample is based on the physical and luminescence stability of ucnps, the stable covalent interaction between ucnps and antibody, as well as the strong buffering capacity of the detection system. detection methods for bioterrorism agents include the isolation-culture or animal inoculation method, biochemical method, nucleic acid method, and immunization method. â�¥ 400** â�¥ 400 ** the sensitivity of upt-lf strip improved by one order of magnitude at that tolerance limit microorganism bioterrorism agents can be identified by the isolation-culture method or inoculation of animals, and toxin can also be injected into animals for species identification. selective culture medium, as well as inoculation and dissection of susceptible animals, are both common experiments. the culture methods for the various bioterrorism agents are different. (1) bacteria can be identified by selection in selective medium, for example, alkaline peptone water is the medium for selective culture of v. cholerea, while y. pestis can be identified by culturing with bacteriophage lysis. (2) isolation of viruses by cell culture is the main method for virus detection because viruses can only survive in live cells. after virus infection the mutated cell can be directly detected by microscopy, or observed through the change in ph of the medium, hemadsorption or hemagglutination. culture in chick embryo is also common for viruses such as influenza virus. animal incubations are better than cell incubation for some viruses, for example, inoculation of mice is the best culture method for rabies virus. (3) rickettsia bioterrorism-associated agents are cultured and isolated by guinea pig and chick embryo. (4) the types of toxin can be identified by lethality or the animal response after inoculation of susceptible animals or cells, such as vomiting and diarrhea caused by s. aureus enterotoxin b. the susceptible animals for toxins are different, for instance, mouse and cat are sensitive for botulinum toxin and s. aureus enterotoxin, respectively. the isolation-culture method is the basic detection method-even gold standard detection method-for bacterial bioterrorism agents. however, it must be conducted by professionals in particular institutes equipped with biosafety facilities, and bioterrorism agents can be easily transmitted owing to improper operation or defense. because of its low sensitivity, the isolation-culture method is often combined with other methods to improve the accuracy of detection, such as the bacteriophage lysis method (zhao et al. 2013) . for example, the bacteriophage lysis method for identification of y. pestis is realized through adding bacteriophage into cultured bacteria at 18-20â°c. the low sensitivity and specificity of the animal injection method for detection of toxins are caused by the diversity of toxins and individual differences between animals. the biochemical method is based on the properties or metabolism characteristics of the microorganism or toxin, and is in fact a detection system because the species cannot be determined by one property, such as the systemic biochemical detection according to diagnostic criteria for cholera (ws 289-2008 , healthy industry standards of the people's republic of china). the poor sensitivity and specificity of the biochemical method, as well as the complicated operation, are a result of the property similarities between microorganisms or biological substances. the nucleic acid method is based on the principle of dna replication in vitro, such as the polymerase chain reaction (pcr) and loop-mediated isothermal amplification (lamp). pcr is a laboratory detection method that parallels the dna replication process in vivo. the single strand dna template is formed at 95â°c, and then it can match with a primer at annealing temperature (about 55â°c) based on their complementary sequences, subsequently a new complementary dna strand can be obtained using dntp as a reactive material catalyzed by taq dna polymerase. these strands can be further used as a template for the next cycle, therefore the target dna can be multiplied millions of times by dozens of cycles. the products can be determined by dna electrophoresis for common pcr, while for real-time quantitative pcr the amplification process can be monitored by the change of fluorescence signals. in addition to dna as a template, rna could also be used for amplification, and this is realized using the reverse transcriptional pcr (rt-pcr) method. the target genes of pcr for y. pestis include the caf 1 gene that encodes f1 antigen, ymt gene that encodes murine toxin, pla gene that encodes plasminogen activator, hms gene that is related to pigmentation, and specific segment 3a in chromosome (qu et al. 2010) . segment 3a is the main target gene because it is not as easily lost as the genes of plasmid such as pla and caf 1. the target genes of pcr for f. tularensis include fopa gene (ay579741) that encodes outer membrane protein, and the akr gene (am286280, 959924-960988) in chromosomes that encodes aldo/ketoreductase. the genes at two specific toxin plasmid for b. anthracis, including cya, lef, paga at plasmid pxo1 and capa, capb, capc at plasmid pxo2, are often used for species identification (koehler 2009 ), while further identification by the detection of the genes in chromosomes, such as gs sequences, is necessary because of the inaccuracy of detection results caused by the deletion or change of plasmids. the pcr method has higher sensitivity than other current methods, but its application on site as a poct method is limited by its dependency on expensive equipment, sample pre-treatment (particularly difficult for dna extraction from complicated samples) and professional operation, as well as its high false positive results. for detection of bioterrorism agents, the special biosecurity facilities for dna extraction (even in an equipped laboratory the infection and contamination in the dna extraction procedure must be paid particular attention) were the major obstacle for utilization of pcr on site. currently, an instrument based on pcr, called a filmarray (seiner et al. 2013) , demonstrates a promising prospect for application of pct in the field by integration of the sample treatment, amplification, and result analysis in airtight system. however, the complexity of pre-treatment of complicated samples in routine surveillance and public health emergencies is still a significant limitation of application of pcr on site. loop-mediated isothermal amplification (lamp) is a new method for gene diagnosis invented by a japanese professor, in which a dna strand that is complementary to template dna can be synthesized at a determined temperature through strand displacement reaction. using four primers designed according the six segments of the target dna, lamp detection can be realized through one procedure following mixing of the template, primer, strand displacement type dna polymerase and other substrates. the pyrophosphate isolated from dntps in the dna synthesis process can react with mg 2+ ions, resulting in the formation of a white precipitate. lamp has been developed for detection of some bioterrorism agents, such as y. pestis (feng et al. 2017 ) and b. anthracis (qiao et al. 2007) , and it is very suitable for rapid detection on site because only a thermostat is needed and the results can be observed by the naked eye. however, it is not suitable for long strand target dna, because sequences longer than 500 bp are difficult to amplify based on the strand replacement reaction. because lamp is an amplification reaction similar to pcr, a high frequency of false positive results are often generated by lamp because of contamination. the immunization method is based on the reaction between antigens and antibodies, such as enzyme linked immunosorbent assay (elisa), immunochromatography, immunodiffusion, and immunoprecipitation. immunodiffusion and immunoprecipitation were developed in the 1950s, and are not widely used at present owing to poor sensitivity. elisa is a routine laboratory method with excellent sensitivity and specificity owing to signal amplification by enzymes and several rinse procedures. however, a high number of rinse procedures increases the complexity of the operation and the operation error, as well as the potential for spread of bioterrorismassociated agents. immunochromatography was described in sect. 16.4.2. many upt-lf assays have been developed for detection of several bioterrorismassociated agents, including detection of bacteria (y. pestis, brucella, b. anthracis spore, f. tularensis, b. pseudomallei, v. cholorae, e. coli o157:h7, antibody against y. pestis), viruses (hepatitis b surface antibody) and toxins (abrin, ricin, aflatoxin b1, aflatoxin m1 and t-2 toxin). upt-poct has been provided at many centers for disease control and prevention, as well as entry-exit inspection and quarantine bureaus, and it also provides technology support for biosecurity at large events, such as the games of the 2008 olympiad in beijing, shanghai world exposition, and asia games in guangzhou. in 2011, upt-poct was integrated as a mobile biological rapid detection instrument in the stand for construction of city fire station (152-2011, issued by the ministry of housing and urban-rural development of china and the national development and reform commission of china). upt-poct is suitable for detection of bioterrorism agents on site as a poct method because of the integration of the benefits of immunochromatography, upconverting phosphor particles, and portable biosensing. first, upt-poct can give reliable detection results. given the extremely low pathogenic dose and high social perniciousness of bioterrorism agents, particularly the high transmission capacity of microorganism agents, detection sensitivity and reliability are important for detection methods. compared with the isolation-culture method, biochemical method, and colloidal gold method, upt-poct can realize sensitive and quantitative detection based on immunological interactions, and shows excellent performance that is comparable with that of real-time quantitative pcr in some applications (hao et al. 2017) , resulting from the merits of the immunological recognition mode, up-converting phosphor, and biosensor. first, immunological recognition between antigens and antibodies is highly sensitive and specific. in addition, the infrared excitation light for up-converting phosphor particles avoids the interference from other biomaterial in the samples, resulting in a more efficient signal isolation rate than other luminous detection methods. the efficient signal extraction and quantitation calculation by the biosensors also facilitate the recognition of the weak positive signal, which is superior to observation with the naked eye for colloidal-gold immunochromatography assays. upt-poct can be tolerant to a great diversity of complex samples in the bioterrorism and biosecurity fields (such as meat, decomposed viscera, and flour). many detection methods are limited by the complicated pre-treatment of samples. for example, repeated selection and identification is required for the isolation-culture method, and the complex composition of samples can easily influence the result of biochemical detection. the pretreatment of complex samples is also a significant challenge for laboratory methods, such as the extraction of dna in complicated samples in pcr that could seriously influence the detection rate. in detection of many bioterrorism agents, upt-poct shows robust performance for various samples with simple pretreatments, such as grinding, or supernatant extraction by centrifugation after grinding and shaking (ultrasonic). this robust performance derives from the physical stability, luminous stability, and up-converting capacity of ucnps, the solid covalent combination between the ucnps and antibody, and the excellent buffering capacity of the detection system. upt-poct is also safer than other detection methods. there are some operations that are unfavorable for the control of bioterrorism agents during detection using other methods, such as repeated proliferation for the isolation-cultured method, various rinses for elisa, complex pre-treatment of sample for the pcr method, and multiple tests for bio-chemical detection. compared with these methods, the simple sampletreatment process based on its high tolerance, and the simple sample-loading manner of upt-poct, reduces the potential for the spread of bioterrorism agents in the detection process. the short acquisition time for upt-poct facilitates rapid response to disease outbreaks in surveillance and public health emergencies, and is derived from the 15 min reaction process of the immunochromatography detection mode and the simple sample pretreatment. as a quantitative detection method, the detection time needed for pcr is more than that for upt-poct owing to the time required for dna extraction. portability is the main obstacle to many laboratory detection methods for application on site, for example the expensive and cumbersome equipment for the pcr method. the portability of upt-poct derives from the small size of the strips and biosensors. in addition, the biosensor could work with a standard mains electricity source or battery. for upt-poct, reliable results can be obtained by nonprofessionals owing to the simple nature of upt-poct operation and sample pre-treatment, making it possible to treat incidents rapidly for surveillance and in public health emergencies. characterization of the rat pneumonic plague model: infection kinetics following aerosolization of yersinia pestis co92 pneumonic plague cluster water-borne outbreak of oropharyngeal and glandular tularemia in georgia: investigation and follow-up association of the melioidosis agent burkholderia pseudomallei with water parameters in rural water supplies in northern australia recovery efficiency and limit of detection of aerosolized bacillus anthracis sterne from environmental surface samples yersinia pestis detection by loop-mediated isothermal amplification combined with magnetic bead capture of dna outbreak of cholera associated with crab brought from an area with epidemic disease the far-reaching impact of bioterrorism: what the european union is doing regarding deliberate releases of biological/chemical agents based on the events in the united states development and evaluation of an up-converting phosphor technologybased lateral flow assay for the rapid, simultaneous detection of vibrio cholerae serogroups o1 and o139 toxigenic vibrio cholerae o1 in water and seafood development of an up-converting phosphor technology-based 10-channel lateral flow assay for profiling antibodies against yersinia pestis development and systematical evaluation of an up-converting phosphor technology based lateral flow assay for quantitative detection of burkholderia pseudomallei development and evaluation of an up-converting phosphor technology-based lateral flow assay for rapid detection of francisella tularensis coexistence of vibrio cholerae o1 and o139 bengal in plankton in bangladesh bacillus anthracis physiology and genetics development of up-converting phosphor technology-based lateral-flow assay for rapidly quantitative detection of hepatitis b surface antibody development of up-converting phosphor immunochromatography for fast and quantitive detection of bacillus anthracis spore rapid detection of aflatoxin m1 in milk powder and milk based on up-converting phosphor technology rapid detection of abrin in foods with an up-converting phosphor technology-based lateral flow assay pathogenesis of yersinia pestis infection in balb/c mice: effects on host macrophages and neutrophils biological and chemical bioterrorism agents loop-mediated isothermal amplification for rapid detection of bacillus anthracis spores rapid and quantitative detection of brucella by up-converting phosphor technology-based lateral-flow assay ambient stable quantitative pcr reagents for the detection of yersinia pestis evaluation of the filmarray(r) system for detection of bacillus anthracis, francisella tularensis and yersinia pestis identification of francisella tularensis by both culture and real-time taqman pcr methods from environmental water specimens in outbreak areas where tularemia cases were not previously reported an outbreak of primary pneumonic tularemia development of ucp-immunochromatography test for rapid detection of e. coli o157 toxin warfare agents: recognition molecules and drugs for control preparation of monoclonal antibodies against ricin toxin and development of up-converting phosphor technology-based lateral flow assay for its quantitative detection rapid quantitative detection of yersinia pestis by lateral-flow immunoassay and up-converting phosphor technology-based biosensor evaluation of up-converting phosphor technology-based lateral flow strips for rapid detection of bacillus anthracis spore, brucella spp., and yersinia pestis outer membrane proteins ail and ompf of yersinia pestis are involved in the adsorption of t7-related bacteriophage yep-phi development and evaluation of an up-converting phosphor technologybased lateral flow assay for rapid and quantitative detection of aflatoxin b1 in crops key: cord-266189-b3b36d72 authors: dignum, frank; dignum, virginia; davidsson, paul; ghorbani, amineh; van der hurk, mijke; jensen, maarten; kammler, christian; lorig, fabian; ludescher, luis gustavo; melchior, alexander; mellema, rené; pastrav, cezara; vanhee, loïs; verhagen, harko title: analysing the combined health, social and economic impacts of the corovanvirus pandemic using agent-based social simulation date: 2020-06-15 journal: minds mach (dordr) doi: 10.1007/s11023-020-09527-6 sha: doc_id: 266189 cord_uid: b3b36d72 during the covid-19 crisis there have been many difficult decisions governments and other decision makers had to make. e.g. do we go for a total lock down or keep schools open? how many people and which people should be tested? although there are many good models from e.g. epidemiologists on the spread of the virus under certain conditions, these models do not directly translate into the interventions that can be taken by government. neither can these models contribute to understand the economic and/or social consequences of the interventions. however, effective and sustainable solutions need to take into account this combination of factors. in this paper, we propose an agent-based social simulation tool, assocc, that supports decision makers understand possible consequences of policy interventions, but exploring the combined social, health and economic consequences of these interventions. in order to handle crises like the covid-19 crisis, governments and decision makers at all levels are pressed to make decisions in a very short time span, based on very limited information (rosenbaum 2020) . where italy and later spain are criticized by not being quick enough to react on the spread of the sars-cov-2 virus, one wonders whether others would have made different decisions at that time. although health is considered of the prime importance, the interventions from governments are having huge economic impacts and possibly even bigger socio-psychological impact. because the incubation time of the coronavirus is around two weeks it will take at least two weeks before the effects of any restriction on movements is visible. that is, decisions will everytime be two weeks behind the actual situation. given this fact, many countries made a jump start and installed possibly more severe restrictions than necessary in order to stay ahead of the effect. however, it also means that in many countries the restrictions will last for several more months. extended lockdown and restrictions on movement are leading to social stress and the economic effects will have lasting consequences too. it can be expected that people will find unforeseen ways to circumvent the restrictions and there will be increased attempts to violate the restrictions, all potentially undoing the effects of the restriction. at the same time lockdowns are causing social stress, unemployment in several countries is rising at an incredible speed, causing even more social unrest. this is pushing the timing and type of strategies to exit lockdown and lessen the current restrictions, leading many governments to consider the introduction of tracking apps or other means to limit spread once restrictions are (partially) lifted, in order to limit the risks of a second or third pandemic wave. how and when should restrictions be lifted? is it better to first re-start schools, or should transport and work take priority? what will be the effect of these strategies on the pandemic but also on the economy and social life of people and countries? it is also clear that interests and possibilities are not equal for all countries. thus a preliminary lifting of restrictions in one country might adversely affect the spread in the neighbouring countries. unless all cross border transport is halted, which is almost impossible given the economic and food dependencies between countries worldwide. the above considerations make clear that health, social-psychological and economical perspectives are tightly coupled and all have a huge influence on the way the society copes with the covid-19 crisis. epidemiology (the study of the distribution and determinants of diseases in humans) is a cornerstone of public health, and shapes policy decisions and evidence-based practice by identifying risk factors for disease and targets for preventive healthcare. as such, it is not strange that during the initial times of the covid-19 pandemic, decision makers got their main advice from epidemiologists. epidemiology assumes that risk factors are general, abstract and difficult for an individual to control. however, individual behavior has a direct influence on these risks, as it became evident in the spread of the coronavirus. the study of individual and societal behaviors are therefore necessary to the taken into account next to epidemiological studies, in order to be able to understand the combined effect of any policy measure across all aspects. this is especially serious if the effects of a restriction (or the lifting of a restriction) have an effect in one perspective that invalidates the assumptions made from another perspective (bénassy-quéré et al. 2020) . there is thus a need for ways to couple the different perspectives and model the interdependencies, such that these become visible and understandable. this will facilitate balanced decision making where all perspectives can be taken into account. tools (like the one we propose) should thus facilitate the investigation of alternatives and highlight the fundamental choices to be made rather than trying to give one solution. in this paper, we present assocc (agent-based social simulation for the covid-19 crisis), an agent-based social simulation tool that supports decision makers gain insights on the possible effects of policies, by showing their interdependencies, and as such, making clear which are the underlying dilemmas that have to be addressed. such understanding can lead to more acceptable solutions, adapted to the situation of each country and its current socio-economic state and that is sustainable from a long term perspective. in the next section we will briefly discuss the current covid-19 crisis situation. in sect. 3, we will discuss the different perspectives of the consequences of the crisis, and show what is needed to connect the different perspectives. in sect. 4, we describe the agent architecture at the core of the assocc approach, which brings together the epidemiologic, social and economic perspectives, and show how this is implemented in a workable agent architecture that can be used in a social simulation framework. in sect. 5, we describe the practical application of the social simulation framework assocc by exploring different example scenarios where the different perspectives are combined. conclusions are drawn in sect. 6. the covid-19 crisis is characterized by very emotional debates and an atmosphere of crisis and panic (khosravi 2020) . when the pandemic spread from asia to europe it took some time to realise its possible consequences and what would be appropriate measures to prevent these consequences. also the usa seemed to ignore what was happening in asia and europe for some time, causing a considerable delay in introducing defensive policies when the pandemic finally reached the country. in a country like the usa where little social security exists and government is not prepared to invest money in preparing for disasters the societal consequences are possibly even larger than in europe (hick et al. 2020) . this was also one of the lessons learned from the storm katrina (executive-office 2006), but the lack of preparedeness still remains. in the usa alone already by end march 2020, over 6.6 million americans had filed for unemployment. 1 as the number of covid-19 cases increased, policies had to be made quickly in order to prevent the rapid spread of the virus resulting in an overload of hospital and ic capacity. given the lack of data about the coronavirus at the time of decision making, epidemic models based on those modeling earlier influenza epidemics were leading on making decisions. there was simply not enough information and not enough possibilities to take more focused measures that would target the right groups and still had the desired effect. the early days of the pandemic show a quick change of approaches as more became known about the coronavirus. e.g. the fact that in early stages it was believed that asymptomatic carriers were not able to transmit the virus, was influential in initial decisions concerning testing and contact tracing, and possibly to the high speed at which the virus spread initially. currently many countries have introduced severe movement restrictions and full lockdown policies, with potentially huge social and economic consequences. however, also in these cases it is unclear what are the factors and motivations leading to the policy. in fact, an initial study by oxford university shows that there is little correlation between the severity of the spread of the coronavirus and the stringency of the policies in place in different countries. 2 the initial idea was that these measures would last for one or two months. however, at the moment there are already several countries speaking about a period of several months extending at least until mid 2020 or even longer. based on data from previous pandemics, initial economic policies were based on the expectation of getting back to normal within a limited amount of time, with many governments soldering the costs for the current period, it is increasingly clear that impact may be way above what governments can cope with, and a new 'normal' economy will need to be found (bénassy-quéré et al. 2020) . and above that, the international dependencies of the world economy require that countries should coordinate their policies in order to sort maximum effect. a thing which is notoriously difficult and has not been improved by the recent attitude of the usa to go for its own interests first and show little solidarity with other countries. from a sociological perspective there are not many theories that can be used to predict how the current situation will develop. however, some principles are clear. people have fundamental needs for affiliation and thus need to socialize. people can use the internet for some of these needs, but physical proximity to other people is a basic need and cannot be withheld for a long period without possibly severe consequences. keeping families locked up in their homes for long periods also will create problems of its own, even without considering the particular dangers for disfunctional families and domestic violence victims. new practices need to be formed where all members of the family will experience less privacy and autonomy and have to adjust their daily practices to accommodate the new situation. this is possible for short periods as tensions can be kept in reign. however, over long periods this might lead to conflicts and consequent problems of distress, increased domestic violence, suicides, etc. cluver et al. (2020) . these experiences might also affect families and society on the long term. people will change their behavior permanently to avoid similar situations. thus e.g. people might be less inclined to travel, get close to other people, etc. the effects of these changes can be more subtle, but have a long lasting effect on the well being of society. from the considerations above may be clear that policies impact epidemics, economics and society differently, and that a policy that may be beneficial from one perspective may lead to disastrous consequences from another perspective. as the crisis progresses and with it its impact increases, decision makers need to be aware of the complexity of the combined impact of policies. means to support understanding this complexity are sorely needed as are tools that enable the design and analysis of many 'what-if' scenarios and potential outcomes. in this section, we describe the epidemics, economics and social science models that are needed to support decision makers on policies concerning the covid-19 crisis and the complexity of combining these models. epidemiological models are dominated by compartmental models, of which sir (cope et al. 2018) , or seir, formulations are the most commonly used. in compartmental models people are divided into compartments, as depicted in fig. 1 . seir shows how people start as being susceptible to a virus (s), then can become exposed (e) to the virus. from that state they can become infected ( i 1 ). in that condition they can visit a doctor ( o 1 ) or just stay home ( i 2 ). from those states they can still visit a doctor and in the end they are either recovered (r) (or passed away). once recovered people can become susceptible again if immunity does not last. the spread of the virus is determined by the probabilities with which persons move from one state to the next. of course this figure gives a very simple picture of all complexities involved in epidemic models. however, what is clear from this picture is that people are a kind of passive containers of the virus that can infect others with a probability and can recover and become immune. there are no explicit actions that people undertake. a model like the above is often combined with a social network model that shows the speed of the spread along certain dimensions. when a person gets infected that is central in the network he will spread the virus in all directions and the epidemic spreads quick as well. cope et al. (2018) these models are mathematical simplifications of infectious diseases, and allow for understanding how different situations may affect the outcome of the epidemic, by assuming that every person in the same compartment exhibit the same characteristics. that is, the model does not consider explicit human behaviour (as is also explained in the media nowadays 3 ) or consequences of interventions on actions of people (heesterbeek et al. 2015) . these are either considered outside of the model, or are transformed into an estimated effect on the parameters of the seir model that is uniform for all individuals in one compartment. e.g. the effect of closing schools on the spread of the corona virus can be interpreted as a lowering of factor in fig. 1 (meaning a lower number of places where the spread of the virus is possible and thus a lower probability for the virus to spread). however, what if the consequence of closing schools would be that children are staying with their grandparents or that they are brought together at the homes of other children rotating caring between families? in these cases, the number of places where the virus can spread may actually increase, which might outweigh the effect of closing the schools. it is possible to include all these factors in to the probability factor. however, by doing so, we loose the actual causal links between the different factors as these are not explicit part of the model and therefore cannot be easily identified and adjusted. in economics there are many competing models and theories. without singling out a specific model, it can be said that in general economic models have difficulties in times of crisis (kirman 2010; colander et al. 2009 ). competing theories focus on different aspects of the economy and make different assumptions about the rest. the main issue all models struggle with (exactly like epidemiological models) is that of human behavior. often, economic models take a 'homo economicus' view on human behavior: a 'rational' individual that always goes for maximum utility or profit in all circumstances. however, we all know this is not always true, as can be easily illustrated by the ultimatum game, an experimental economics game in which one player proposes how to divide a sum of money, e.g. 100 dollars, with the second party. if the second player rejects this division, neither gets anything. rationally, the agent that gets offered should accept any offer as it is more than nothing (what he gets when refusing). however, empirical studies show that people only accept what they perceive as fair offers (andersen et al. 2011) . in our example, when more than 30-40 dollars are offered. so, fairness apparently also is worth something! it neatly illustrates that people have more motivations to take actions than mere utility of that action. again, such motivations and values can be somehow incorporated in economic models, but such interpretation is not part of economic theory. i.e. economics does not directly provide an answer on how much fairness is worth. this will depend from the context and history. e.g. suppose an agent is pretty fair and offers the other agent 49 dollars and the other agent accepts. the next time the same agent offers the other agent 45, but now the other agent refuses due to the fact he feels he will get offered less and less. however, he might very well have accepted the 45 if it was offered the first time. the above is just one example to show that economic models are very good to explain and predict some behavior of people, but do not include all motivations and behaviors following from those. the third perspective that is relevant for modeling of a pandemic is the social science. in particular, social network analysis is often used to understand the possible ways a virus might spread (firestone et al. 2011) . nowadays much of the work in this area is related to online social media networks, because a lot of data is easily available on such networks. however, for the spread of a virus the physical social networks (between friends, family, colleagues, etc.) are those of main interest. given that there is enough data to construct the physical social networks, they are a very good tool to determine which people might potentially be big virus spreaders. this can be due to their role (e.g. nurses in elderly care) or due to the fact that they have many contacts in different contexts (e.g. sports and culture) that are otherwise sparsely connected and they form bridges between densely connected communities. however, knowing which people or which roles one would like to target for controlling the spread does not mean one can device effective policies. e.g. in the current pandemic it is clear that closing schools as being a possible bridge between communities is acceptable, while closing hospitals or elderly care centres is not. so, again, additional aspects have to be included in these models in order to use the theory in practice. more semantics for the nodes in the networks are needed as well as what the links between the nodes are constituted of. do the links indicate the number of contacts per day? can people also change the network based on their perception of a situation? i.e. avoid contacts, have different types of contacts, establish new contacts, etc. so, where social network analysis looks at people as nodes in a network, people are the ones that actually create, maintain and change the social network. when a government tries to contain the spread of a virus the social networks give a good indication where that might be done most effective, but how the people constituting the network will react to the policies is not included in the social network theory. thus whether new links will arise bypassing previous links or other persons will take the place of so-called super spreaders cannot be derived from this theory. given the above short discussions of the different modeling perspectives of the crisis one can conclude that all perspectives include some assumptions about human behavior in their models. however, this behavior, and especially how people influence each other's behavior is not part of any of these models. we propose to take the human behavior as the central perspective and use it as a linking pin to connect the different perspectives as illustrated in fig. 2 . in the next section we will discuss how our agent model from the assocc framework can be used to fulfill this central role to couple the different perspectives. the design of the assocc framework is based on the fact that individuals always have to balance their needs over many contexts. in the research that we have done in the past twenty years we have come to the the sketch in fig. 3 that illustrates how people manage this balancing act in their daily life. we assume that people have a value system that is reasonably consistent both over times, contexts and domains. the value system is based on the schwartz value circumplex (schwartz 1994 ) that is quite universal. it depicts a number of basic values that everyone possesses and their relations. people differ in how they prioritize values rather than on which values they have. although, priorities can differ between individuals they are reasonably consistent within cultural groups. therefore the values can also be seen as linking individual drivers to the social group. we have used this abstract social science framework already in vanhee et al. (2011), vanhée (2015) , cranefield et al. (2017) , heidari et al. (2018) . in order to use it in these simulations we have formalized and extended the framework such that it can be coupled to concrete behaviour preferences in each context. thus values give a stable guideline of behavior and they will be kept satisfied to a certain degree whenever possible. thus, if "conservatism" is important to a person, she will, in general, direct her behavior to things that will benefit the preservation of the community. the second type of drivers of behavior in fig. 3 are the motives that all people have in common. this is based on the theory of mcclelland (1987) . the four basic motives that are distinguished are: the achievement motive drives us to progress from the current situation to something better (whatever "better" might mean). the affiliation motive drives us to be together with other people and socialize. thus we sometimes do things just to be doing it together with our friends and family. the power motive actually does not mean we want power over others, but rather that we want to be autonomous. i.e. being able to do tasks without anyone's help. finally, the avoidance motive lets us avoid situations in which we do not know how to behave or what to expect from others. each of these motives is active all the time and whenever possible it will drive a concrete behavior. thus, i might go to my grandparents to ask a question rather than text them, just because i want to have a chat. the third type of elements that determine behavior are the affordances that a context provides. these affordances determine what kind of behavior is available and also what type of behavior is salient. e.g. in a bar one often drinks alcohol. even though it is not obligatory it is salient and also afforded easily. individuals have to balance between their values, their motives and the affordances to determine what behaviour would be more appropriate in each situation. as one can imagine this is quite tricky and will take too much time and energy if done in every situation from scratch. therefore, in human society we have developed social structures to standardize situations and behaviors in order to package certain combinations that will be acceptable and usually good (even if not optimal). these social constructs are things like: norms, conventions, social practices, organizations, institutions. note that these constructs give general guidelines or defaults of behavior, but are no physical restrictions on what is possible! implementing this whole architecture would be too inefficient for any social simulation. therefore we use this as theoretical starting point, but translate it into a simpler model that is more efficient and scalable. thus for the assocc simulation framework we fix the the most important aspects of the values and motives described in the above architecture, into a set of needs, illustrated in the following fig. 4 . from the figure it can be seen that we model the values and motives as needs that deplete over time if nothing is done to satisfy them. the model prevents that an agent will only look at the need with the highest priority and only at other ones when that need is completely satisfied. by calibrating the size and threshold and the depletion rate of each need we can calibrate and balance all the needs over a longer period, between different contexts and over several domains. e.g. using this model it becomes possible to decide for an individual whether it is more important to work a bit more or go home and be with the family. this simple model is the crux behind combining health, wealth and social wellbeing in a simulation model (fig. 5) . there is not a complete mapping as not all values are of importance given the type of activities and choices that the agents can make. moreover, the achievement motive is not included explicitly. individual agents inherently are driven to take actions because they need to fulfill their needs. longer term kinds of achievement, like getting educated or getting rich are not important for the purpose of this context. some of the needs are subdivided. safety has the more concrete needs of foodsafety, financial-survival, risk-avoidance, compliance, and financial-safety. food safety and financial survival represent that individuals have to have enough food and money to survive. financial safety means that one has some buffer to pay other than the basic necessities. compliance indicates whether complying to norms is important or not. risk-avoidance in this context indicates whether people take actions that might get them infected or avoid any of those at all costs. given this core model of the agent decision model for behavior we can now describe the actual agent architecture as we have implemented it in assocc. we have developed a netlogo simulation consisting of a number (between 300 and 2500) of agents that exist in a grid. agents can move, perceive other agents, and decide on their actions based on their individual characteristics and their perception of the environment. the environment constrains the physical actions of the agents but can also impose norms and regulations on their behavior. e.g. the agents must follow roads when moving between two places, but the environment can also describe rules of engagement such how many agents can occupy a certain location. through interaction, agents can take over characteristics from the other agents, such as becoming infected with the coronavirus, or receive information. the main components of the simulation are: • agents: representing individuals. agents have needs and capabilities, but also personal characteristics such as risk aversion or the propensity to follow the law and recommendations from authorities. needs include health, wealth and belonging. capabilities indicate for instance their jobs or family situations. agents need a minimum wealth value to survive which they receive by working or subsidies (or by living together with a working agent). in shops and workplaces, agents trade wealth for products and services. agents pay tax to a central government that then uses this money for subsidies, and the maintenance of public services such as hospitals and schools. • places: representing homes, shops, hospitals, workplaces, schools, airports and stations. by assigning agents to homes, different households can be represented: families, students rooming together, retirement homes, three generation households and co-parenting divorced agents. the distribution of these households can be set in different combinations to analyse the situation in different cities or countries. • global functions: under this heading we capture the general seir model of the corona virus which is used to give agents the status of infected, contagious, etc. this model also determines the contageousness of places like home, transport, shops, etc. based on a factor that represents fixed properties of a place (like size, time people spend there on average, whether it is indoor or outdoor) and density (how many people are there at the same time). under this global functions we also capture economic rules that indicate tax and subsidies from government. finally we also include the social networks and groups that exist under this heading. the social networks give information about normal behavior and also provide clusters of agents performing activities together. • policies: describing interventions that can be taken by decision makers. for instance social distancing, testing or closing of schools and workplaces. policies have complex effects for the health, wealth and well-being of all agents. policies can be extended in many different ways to provide an experimentation environment for decision makers. agents can move between places and take the policies into consideration for their reasoning. as described in sect. 4, agents' decisions are based on the container model of needs. these needs are satisfied by doing activities and decay over time. needs may have different importance to each agent but the overall assumption is that agents will try to satisfy their most important need that is least satisfied at a given moment given the context. the context determines which choices are available at any given moment. thus e.g. if agents have to work in a shop they will (normally) go to work even if the need for safety is high. but if they have work that can be done at home as well, they have a choice between going to work or staying home to work. in that case, their need for safety can make them decide to stay home. most needs are composite. for instance, safety is built up of food-safety, financial-survival, risk-avoidance, and compliance, listed here in order of importance, i.e. relevance to the agent's direct need to survive. e.g. the safety need is defined as the minimum of the first two, and a weighted mean of the rest, where the weights are the importances assigned to each subneed. here the satisfaction of food-safety is defined as having enough essential resources stocked up at home, such as food and medicine, measured over a two week period (i.e. the need is fully satisfied if the agent has enough supplies for the coming two weeks and decays from there). the only way to increase this need, is by going shopping for essential resources. however, going shopping, i.e. leaving the house may conflict with the need for safety, so the agent will need to balance these two needs in its decision to go shopping. agents with a high level of risk-avoidance will be more likely to try to avoid the disease and thus want to stay away from large groups of people. the need for belonging includes conformity, i.e. the need to comply with norms and regulations, which is satisfied by taking actions that conform to the rules, such as staying inside during lockdown, or going to school or work if that is requested from them. conformity can have a negative value, in the case that the agent decides to break a rule. the need for autonomy is satisfied when agents are able to follow their own plans. agents satisfy their need for autonomy when they are able to make an "autonomous" decision. lockdown policies block many of these actions, which means that when an agent reaches a too low level of autonomy it may decide to break lockdown. however, to regulate this effect and not provide agents with too strong incentives to break the lockdown, the 'compliance' is used as a regulating factor. the need for self-esteem is satisfied when the agent believes that its actions are accepted and even followed by other agents. finally, the need for survival represents an agent's need to rest if it is sick. this need can be satisfied by resting at home if the agent believes it is sick, and depletes if it does anything else while it believes to be sick. under this need we also fitted the conformity need. people will conform to what other people do if they are uncertain about the context and which action is the best to take. conforming to others is safe as it is usually good to do what others do. thus it contributes to (social) survival. as can be seen above, agents in assocc have a wide range of needs. they range over the social, health and economic dimensions of society and can therefore also show how interventions intended to remedy e.g. the spread of the virus can influence other dimensions and due to these dependencies do not have the intended effect. in the next section we describe some examples of scenarios we have simulated in assocc to illustrate the use, the range of possible scenarios and importance of the approach. in this section we briefly discuss two scenarios we have simulated with assocc. they are about quite different aspects of the covid-19 crisis. the first is on the question whether schools should be closed and employees work at home during the pandemic. the second investigates some economic effects of having a lockdown for several weeks or even months. both scenarios are built on the same model as described above, which shows the richness of the model and possibilities for exploration of new policies. typically, schools are places where both children and parents from a community gather, which potentially leads to spreading of the virus. when epidemics emerge, one common measure to be taken is the closure of schools. in case of influenza-like illnesses it has been proven that this is an effective measure, since children are highly susceptible for that disease. however, the covid-19 pandemic is different, as it seems to be less contagious for children. sweden, as one of the few european countries, has kept primary and junior school (0-16 years old) open, based on the premise that closing schools would mean many parents with jobs that are vital to society, like healthcare personnel, would have to stay home and take care of their children instead of going to work. this would delay efforts to stem the spread of the virus. in chang et al. (2020) it is already indicated that closing schools would not have much effect in the australian situation. in this scenario, we explore potential consequences of keeping schools open on the spread of the virus but also on social and economic aspects. we model the direct and indirect effect on the spread of the virus when schools are closed and people work from home. we assume that schools will be closed as soon as a certain amount of infected people within the city has been reached. different thresholds have been tested, but in this paper schools are closed whenever one infected person is detected. the scenario assumes that when children are staying at home, at least one adult should be at home to take care of them. this caregiver is assumed to work from home for the duration of the school closure. we will in particular look at the effect on the spreading of the coronavirus when schools are closed and thus parents working at home compared to parents already working at home with children still going to school. in fig. 6 the results of closing the schools and working at home are shown. we show a graph with the health status of the population and the activities taking place. when only the schools are closed (the left two graphs), some people are forced to work at home too, as they have to take care of their children. the scale on the horizontal axis denotes time. four ticks make up one day. so, the runs cover about 2-4 months (depending on whether any differences still appear after 2 months). the vertical axes show the number of agents which is around 330. this number can vary slightly as it depends on the household distribution and the way the households are generated. the results are based on 40 runs. although the number of people at home is of course much higher than people in other places (the green line in the bottom plots) compared to the baseline simulation, which implies less spreading of the virus, the peak of infected people is higher (around 210 and 215 resp.). the peak is reached around the same time, i.e. 40 ticks (or 10 days). the number of people that died is 80 and 78 resp. this means that the differences are not significant. we also see a difference in people going to non-essential shops (the black line in the activity graphs). the number of people at non-essential shops, on the right where the last two peaks (saturday and sunday) show that people go out shopping in the weekend. the above findings show that the measures do not lower the peak as expected, but even increase that peak. this can be explained by the fact that more people want to leave the house during the weekend, as they have worked at home during the week with less social interaction. at these public places typically people of different communities are gathered, while at work or school, more or less the same group of people are present. thus closing the schools only increases the spreading of the virus. thus without additional restrictions, like social distancing and closing restaurants, the effect of working at home or closing schools is surpassed by the side effect, namely people wanting to go out. the covid-19 pandemic has severe medical and social consequences. but it has vast economic consequences as well. a recession is expected, and predictions show that it can be way more severe than the banking crisis (2009) recession. thus governments are trying to minimize the economic impact of the pandemic and try to minimize the "financial issues" people and companies have by giving massive a b c d fig. 6 results financial support. also, once the pandemic has been dealt with, "restarting" the economy is a big challenge as many people have lost their jobs and economic activity has slowed to a minimum. with this scenario we demonstrate the relation between the pandemic, health measures and the economy. this shows the complex and interconnected nature of the situation. in many countries the government has locked down the country and closed down business operations. this cuts the cycle of income of companies and subsequently either people become unemployed or governments temporarily take over paying the wages of the employees. in this scenario we show two different situations. the first shows what happens with a lockdown and no government support, the second shows the situation when government supports companies by taking over wages. when government closes all non-essential shops to prevent spread and orders a lockdown the curve of the pandemic is really flattened. within the time span of our simulation we see that we are still not experiencing many infections and hardly any deaths. note that lockdown here starts when only one person was infected. so, much earlier than in real life! people are almost all working at home. the essential shops seem to be doing well. this can be seen as hoarding behavior of people that want to make sure they have enough essential products and have still money to buy things. they do not buy anything non-essential as those shops are all closed (and we have no on-line shopping in our simulation!). the non-essential shops do not go bankrupt in the simulation because we did assume here that they could postpone all fixed costs, like loans, rent, etc. if these fixed costs are taking into account all the nonessential shops will go bankrupt pretty quick without government intervention. the velocity of the money flowing through the system decreases due to the lockdown (fig. 7) . in the second situation government takes over wages from all people working in the non-essential shops. overall we see capital increasing slightly. this can be explained due to the fact that people cannot spend money on non-essential products and leisure activities (going to cafes, sport events, etc.) while the number of unemployed/non-payed people is relatively small (fig. 8) . the economic activity is kept up by the government. however, the government reserves are quickly depleting as less tax is coming in and more subsidies are paid. this situation is not sustainable! in this paper we have shown that in crises like the covid-19 crisis the interventions of government have to be made quick and are often based on too little information and only partial insights of the consequences of these interventions. in most cases the health perspective is the leading perspective to inform government decisions. however, the models used by epidemiologists lack the part on human behavior. therefore they have to translate government interventions and the expected reactions from citizens to parameters into parameters of the epidemic model without being able to check for interdependencies. we have shown how the assocc platform can be a good addition to current epidemiological and economic models by putting the human behavior model central and from that core connect the health, wealth and social perspectives. in general there are several advantages of using an agent based tool like assocc. first a tool like assocc explicitly avoids providing detailed predictions, but rather supports investigating the consequences in all perspectives from government policies. by comparing variations of policies (like government subsidizing people or not in the economic scenario) it becomes clear what are the consequences and which are the fundamental choices that have to be made. taking this stance avoids politicians being able to just blame the model for giving a wrong advice. technology should not give single solutions to decision makers, but should support the decision makers to have a good insight in the consequences of their decisions and which are the fundamental priorities that have to be weighed. e.g. the joy of life of elderly people against the risks of meeting their family and becoming infected. or the risk of contagion happening through schools vs. the social disruption when schools are closed. due to the agent based nature it is also possible to explain where results come from. e.g. closing schools might not have any positive effect on the spread of the virus due to all kinds of side effects. in the scenario we could see that people were going more to non-essential shops in the weekend because they had to be more at home to take care of the children during the week. their need for belonging went up and going to the shops was the only option left as other leisure places were closed. tracing this behavior back in this way helps to explain what is the basic cause and where in the chain something can be done if one wants to avoid this behavior. each of the steps in these causal chains is based on a solid theory (social-psychological, economic or epidemiological in our case). and thus can be checked both on its plausibility as well as on its theoretical foundation. using the agent based models makes it easier to adjust parameters in the simulation based on differences in demographics and cultural dimensions in different countries, but also in different regions or even neighbourhoods in big towns. therefore more finegrained analysis can be performed on the impact of government interventions on different parts of the country. e.g. rich and poor neighbourhoods or urban and rural areas can be affected in different ways by the same measure. finally, we should stress that we do not claim that agent based models like used in assocc should replace domain specific models! actually one could use assocc to simulate intervention scenarios to study the human reactions to it and the consequences. having a good insight in these dependencies one can feed the domain specific models with better information to make precize optimizations or predictions for the effect of that intervention. thus in this way the strength of the different types of models can be combined rather than seen as competing. supplementary information the full project is available at https ://simas socc. org. the netlogo and gui code is available at https ://githu b.com/lvanh ee/covid -sim including a few runnable scenarios, described in the website. stakes matter in ultimatum games europe needs a catastrophe relief plan. mitigating the covid economic crisis: act fast and do whatever modelling transmission and control of the covid-19 pandemic in australia the financial crisis and the systemic failure of the economics profession characterising seasonal influenza epidemiology using primary care surveillance data no pizza for you: value-based plan selection in bdi agents the federal response to hurricane katrina: lessons learned. government printing office the importance of location in contact networks: describing early epidemic spread using spatial social network analysis. preventive veterinary medicine modeling infectious disease dynamics in the complex landscape of global health simulations with values duty to plan: health care, crisis standards of care, and novel coronavirus sars-cov-2 perceived risk of covid-19 pandemic: the role of public worry and trust the economic crisis is a crisis for economic theory human motivation facing covid-19 in italy-ethics, logistics, and therapeutics on the epidemic's front line beyond individualism/collectivism: new cultural dimensions of values implementing norms? utrecht universiteit publisher's note springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations acknowledgments open access funding provided by umea university.open access this article is licensed under a creative commons attribution 4.0 international license, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the creative commons licence, and indicate if changes were made. the images or other third party material in this article are included in the article's creative commons licence, unless indicated otherwise in a credit line to the material. if material is not included in the article's creative commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. to view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. key: cord-332583-5enha3g9 authors: bodine, erin n.; panoff, robert m.; voit, eberhard o.; weisstein, anton e. title: agent-based modeling and simulation in mathematics and biology education date: 2020-07-28 journal: bull math biol doi: 10.1007/s11538-020-00778-z sha: doc_id: 332583 cord_uid: 5enha3g9 with advances in computing, agent-based models (abms) have become a feasible and appealing tool to study biological systems. abms are seeing increased incorporation into both the biology and mathematics classrooms as powerful modeling tools to study processes involving substantial amounts of stochasticity, nonlinear interactions, and/or heterogeneous spatial structures. here we present a brief synopsis of the agent-based modeling approach with an emphasis on its use to simulate biological systems, and provide a discussion of its role and limitations in both the biology and mathematics classrooms. agent-based models (abms) are computational structures in which system-level (macro) behavior is generated by the (micro) behavior of individual agents, which may be persons, cells, molecules or any other discrete quantities. typical abms contain three elements: agents, an environment, and rules governing each agent's behavior and its local interactions with other agents and with the environment. decades of advancement in computer power has made agent-based modeling a feasible and appealing tool to study a variety of complex and dynamic systems, especially within the life sciences. as the use of abms in research has grown, so too has the inclusion of abms in life science and mathematical modeling courses as a means of exploring and predicting how individual-level behavior and interactions among individuals lead to system-level observable patterns. abms are now one of the many types of models students studying the life sciences or applied mathematics should encounter in their undergraduate education. prior to the introduction of abms into biological and applied mathematics curricula, the clear model format of choice was the ordinary differential equation (ode), or maybe a pair of them; occasionally, discrete difference equations and/or matrix equations would also be introduced. exponential growth and decay were ready examples, paving the way for extensions of the exponential growth process toward a carrying capacity in the form of the logistic growth process (voit 2020) . this logistic process was easily generalized to two populations, which were at first independent, but then allowed to interact. depending on these interactions, the result was a pair of two populations competing for the same resource or a simple predator-prey model in the format of a two-variable lotka-volterra system. although odes and any other types of "diff-e-qs" are a priori dreaded by almost all but mathematicians and physicists, the concept of an ode, if adequately explained, becomes quite intuitive. for instance, one may ease a novice into the world of odes by considering changes in the water level w of a lake over time (ayalew 2019) . whereas these dynamics are difficult to formulate as an explicit function w (t), newcomers readily understand that changes in the water level depend on influxes from tributaries, rain, and other sources on the supply side, and on effluxes, evaporation and water utilization on the side of reducing the amount of water. just putting these components into an equation leads directly to a differential equation of the system (weisstein 2011 ). on the left side, one finds the change over time as dw /dt, and this change is driven, on the right-hand side, by a sum of augmenting and diminishing processes. there is hardly a limit to what can be achieved with odes in biology, with the very important exception of processes that have genuine spatial features. and while it is not difficult to ease a biology undergraduate into ordinary differential equations, the same is not necessarily true for partial differential equations (pdes). however, spatial phenomena in biology seldom occur in homogeneous conditions. as examples, consider the formation of tumors with angiogenesis and necrosis; the local patterns of cell-to-cell signaling that governs the embryonic development; the spread of the red fire ant (solenopsis invicta) from mobile, al, its alleged port of entry into the usa, all along the gulf and east coasts; or the population size and dynamics of the santa cruz island fox (urocyon littoralis santacruzae) being driven by territory size which in turn depends on local vegetation (scott 2019) . until relatively recently, the conundrum of space was often dealt with in the final chapter of mathematical modeling in biology. a sea change came with the development of abms, which are natural formats for both stochasticity and spatial phenomena. by their nature, these models are computationally expensive, which initially prevented their use in most classrooms. however, this situation has obviously changed. as the bio2010 report (2003) stated: "computer use is a fact of life of all modern life scientists. exposure during the early years of their undergraduate careers will help life science students use current computer methods and learn how to exploit emerging computer technologies as they arise." classroom use of abms has thus become not just logistically feasible, but also very appealing for demonstrating spatial dynamics in a wide range of biological systems (kottonau 2011; triulzi & pyka 2011; shiflet 2013; pinder 2013) . supporting this appeal is a repertoire of software tools, such as simbio and netlogo (see sect. 3), that contain predefined examples and require minimal computer coding skills for model analysis. here, we present a brief synopsis and history of this modeling approach with emphasis on life science applications (sect. 2), describe some of the software tools most frequently used in the classroom (sect. 3), and then focus on some of its roles and limitations in the classroom (sect. 4). the origins of agent-based modeling. the true origins of any method or procedure are seldom identifiable in an unambiguous manner. in the case of agent-based modeling, one could think of craig reynolds' 1987 seminal article on the formation of bird flocks (with the agents denoted as boids, short for "bird-oid object"), which he was able to represent with just three rules of behavior: (1) avoid collisions with nearby birds; (2) attempt to match the velocity of nearby birds; and (3) attempt to stay close to nearby birds in the flock (reynolds 1987; gooding 2019) . the result of simulations with this simple abm was very realistic-looking flocking behavior. particularly intriguing in this study was the fact that there was no leader or a global organizing principle. instead, the virtual birds were truly individual agents that self-organized locally, thereby generating a globally coherent flight pattern. while reynolds' work was a milestone, key concepts leading to modern abms can be found much earlier. one notable contributor of ideas was nobel laureate enrico fermi, who used mechanical addition machines to generate probabilities for stochastic models with which he solved otherwise unwieldy problems (gooding 2019) . this procedure was an early form of the method of a monte carlo simulation, which was later independently developed and published by stanislav ulam, like fermi a member of the manhattan project (metropolis & ulam 1949; metropolis 1987) . another very important contribution to the budding development of abms was the turing machine (turing 1936) , which is a mathematical model of computation that uses a set of rules to manipulate symbols in discrete cells on an infinite tape. much closer to abms were ideas of ulam, who was fascinated by the "automatic" emergence of patterns in two-dimensional games with very simple rules (ulam 1950; metropolis 1987) . together with the new concept of game theory (von neumann & morgenstern 1944) , all these ideas were developed into the concept of cellular automata, which are direct predecessors of abms (ulam 1950; ulametal 1947; von neumann & morgenstern 1944) . a very appealing implementation of a cellular automaton was john conway's famous game of life (gardner 1970) . the social sciences adopted computational methods in the 1960s for microanalytic simulations or, simply, micro-simulations (gilbert & troitzsch 2005; gooding 2019 ). in contrast to today's abms, which use simple rules to recreate observed or unknown patterns, the agents in the original microsimulations acted according to empirical data (bae 2016). the advantage of this strategy is that model analysis can reveal systemic behaviors under different realistic scenarios (gooding 2019) . a seminal paper in this context described generic segregation processes (schelling 1971) . other agent-based modeling work in sociology and economics was gleaned from the biological abm work of smith (1982) , who formulated darwin's ideas of evolution as a computer simulation. this idea inspired nelson & winter to apply similar concepts and implementations to market studies, where firms were modeled like animals that followed specific routines. in particular, the firms were in competition, and the market weeded out bad routines while rewarding the fittest. nelson & winter's influential book an evolutionary theory of economic change (nelson & winter 1982) strongly proposed the use of computer simulations, which today would fall within the scope of agentbased modeling. their ideas led to a school of thought called evolutionary economics (hanappi 2017 ). an early and particularly influential paper in this context tried to shed light on the stock market (palmer 1994) . initially, abms of artificial life simulated simple homogeneous agents that acted like huge colonies of ants that could just move and eat in the pursuit of food. somewhat more sophisticated, economic simulations used as the main agent homo economicus, a consistently rational human pursuing the optimization of some economic goal or utility with exclusive self-interest (persky 1995) . based on these humble beginnings, sophistication in computing soon permitted heterogeneous agents and much more complicated landscapes than before. the successes in economics were so tantalizing that simulation studies eventually reached the most prestigious journals of economics and the social sciences geanakoplos 2012; hanappi 2017) . modern abms in economics are capable of capturing much of the complexity of macroeconomic systems (e.g., caiani (2016) ). following directly the principles of cellular automata, kauffman studied large grids with elements that changed features in a binary fashion (kauffman 1993) . for instance, a white agent could turn black, and this shift occurred according to boolean rules that usually involved some or all neighboring grid points. kauffman was able to demonstrate the emergence of complex patterns, such as oscillations and percolation. starting in the 1980, wolfram performed systematic studies of cellular automata, which led to his influential 2002 book a new kind of science (wolfram 2002 ) that assigns cellular automata a wide range of applications in a variety of fields. biological applications of agent-based modeling. abms have been constructed to study a wide range of biological phenomenon. numerous reviews and research efforts using abms have focused on specific biomedical systems. issues of gene expression were modeled by thomas (2019) . morphogenetic and developmental processes were discussed in grant (2006) , robertson (2007) , , tang (2011), and glen (2019) . models of tissue mechanics and microvasculature are captured in and van liedekerke (2015) . inflammation, wound healing and immune responses were addressed in , ), chavali (2008 , and castiglione & celada (2015) . other authors (wang 2015; segovia 2004; cisse 2013 ) used abms to model cancer growth, tuberculosis, and schistosomiasis, respectively. lardon and colleagues (lardon 2011) used abms to analyze biofilm dynamics. butler and colleagues described the use of abms in pharmacology (butler 2015) . abms studying multicellular systems provide a unique capability to examine interactions and feedback loops across the different hierarchies of the system (hellweger 2016) . reviews of abms in the context of ecology, environmental management, and land use include (bousquet 2004; matthews 2007; grimm & railsback 2005; caplat 2008 ; deangelis & diaz 2019). in some applications, interventions or treatments were addressed and therefore required the adaptation of agents to changing scenarios (berry 2002) . abms have also been used to simulate epidemics with analyses examining the impact of implemented or potential intervention measures (e.g., quarantining/physical distancing, mask wearing, and vaccination) (mniszewski 2013; perez & dragicevic 2009; tracy 2018) . visual representations of epidemiological abms have even been used by news outlets during the covid-19 pandemic to help explain to the public how various intervention methods change the shape of an epidemic (i.e., "flatten the curve) or the basic reproduction number (r 0 ) of an epidemic; see, for example, fox (2020) and stevens (2020). rationale & pitfalls of abms. the two most frequent goals of an abm analysis are (1) the elucidation and explanation of emergent behaviors of a complex system and (2) the inference of rules that govern the actions of the agents and lead to these emerging system behaviors. this type of inference is based on large numbers of simulations, i.e., replicate experiments with the abm using the same assumptions and parameters, and different experiments over which assumptions or parameter values are systematically changed. simulations of abms essentially always yield different outcomes, because movements, actions and interactions of agents with each other or with the environment are stochastic events. the inference of rules from simulation results is an abductive process (voit 2019) that is challenging, because one can easily demonstrate that different rule sets may lead to the emergence of the same systemic behaviors, and because even numerous simulations seldom cover the entire repertoire of a system's possible responses. in fact, hanappi (2017) warned: "assumptions on microagents that play the role of axioms from which the aggregate patterns are derived need not-and indeed never should-be the end of abm research." arguably the greatest appeal of abms, and at the same time a treacherous pitfall, is their enormous flexibility, which is attributatble to the fact that any number of rules can be imposed on the agents, and that the environment may be very simple but can also be exceedingly complicated. for instance, the environment may exhibit gradients or even different individually programmable patches (barth 2012; gooding 2019) including importing geographic information systems (gis) data to define the characteristics of each patch (see scott (2019) for an example with a detailed explanation of how gis data are incorporated into an abm). in particular, the repeated addition of new elements to a model can quickly increase the complexity of the model, thereby possibly distracting from the core drivers of the system's behavior, obscuring the importance of each of the rules the agents must follow, and generally making interpretations of results more difficult. related to this option of adding elements is the critique that abms can be 'tuned' by researchers to create outcomes that support the researcher's narrative (gooding 2019) . to counteract increases in complexity, some authors have begun to develop methods of model reduction that retain the core features of a model, but eliminate unnecessary details (zou 2012). another critique of abms is that simulations are not readily reproducible, because the rules can be complicated and stochasticity rarely repeats the same model trajectories. a strategy to increase the reproducibility of abms was the establishment of protocols for the standardized design and analysis of abms (lorek & sonnenschein 1999; grimm 2006 grimm , 2010 heard 2015) . (2017). in a slight extension of abms, lattilä (2010) and cisse (2013) described hybrid simulations involving abms and dynamic systems, and heard (2015) discussed statistical methods of abm analysis. these introductory tutorials and reviews, however, are typically not designed for undergraduates with limited mathematical or computational modeling experience. the scientific community has also worked hard on facilitating the use of abms by offering software like netlogo, swarm, repast, and mason. summaries and evaluations of some of the currently pertinent software are available in berryman (2008) and abar (2017) , and netlogo is described more fully in sect. 3. a variety of software tools can be used to construct, simulate, and analyze abms. when abms are taught or used in biology or mathematics courses, software should be chosen to align with pedagogical objectives. note that the pedagogy of abms in life science and mathematical modeling courses is discussed in more detail in sect. 4. in this section, we highlight some of the most used software packages in an educational setting. ecobeaker & simbio. ecobeaker was developed by eli meir and first released in 1996 as software that ran simulated experiments designed to explore ecological principles. in 1998, simbio (meir 1998 ) was founded (then called beakerware) for the release of the second version of ecobeaker and has since grown to include simulated experiments in evolution, cell biology, genetics, and neurobiology. many of the simulated experiments in the simbio virtual labs software are agent-based simulations. in a simbio virtual lab, the user interacts with a graphical interface portraying the agents (individuals in the experiment) and sometimes their environment, which can be manipulated in various ways for different experimental designs. the graphical interface also contains relevant graphs which are updated as the simulated experiment runs. ecobeaker and simbio virtual labs are examples of software where the focus is on experimental design and the use of simulation to understand biological concepts. the user never interfaces with the code and does not need to understand the underlying algorithms which produce the simulation. simbio virtual labs are used in many biology classrooms across the united states. the software is priced per student (currently $6/lab/student or $49/student for unlimited labs). netlogo. the agent-based modeling environment netlogo (wilensky 1999) was developed by uri wilensky and first released in 1999. netlogo is free and continues to be improved and updated (the software is currently on version 6). additionally, a simplified version of netlogo can be run through a web browser at http://netlogoweb. org/. netlogo is a programming platform allowing the implementation of any abm a user might design. as such, its user interface includes a tab where abm code is written in the netlogo programming language, and a tab for the user to view a visualization of the abm and user specified outputs as the abm is simulated. textbooks by railsback & grimm (2012) and wilensky & rand (2015) provide introductions to the netlogo prgramming languages as well as providing a thorough overview of the algorithmic structures of abms. since its initial development, netlogo has built up an extensive model library of abms. additionally, over the years, faculty at various institutions have developed abm modules through netlogo that allow students to explore a variety of biological phenomenon. for example, the virtual biology lab (jones 2016) has created 20 different virtual laboratory modules using netlogo through a web browser for exploring topics in ecology, evolution, and cell biology. the virtual biology labs are similar in scope to the ecobeaker and simbio labs. another example is infections on networks (iontw) which provides an abm framework and teaching modules for examining aspects of disease dynamics on various network structures (just 2015a, b, c) . one of the responses to the bio2010 report (2003) has been a push to create biocalculus courses or to insert more biological application examples and projects within traditional calculus courses. indeed, studies have shown that including applications from the life sciences in classic math courses like calculus leads to students gaining equivalent or better conceptual knowledge than from similar courses without life science applications (comar 2008; . however, many mathematics and biology educators have pointed out that the subset of mathematics applicable to biology extends well beyond calculus, and undergraduates (especially those majoring in biology) should be exposed to a variety of mathematical models and methods of analysis across biology and mathematics courses (bressoud 2004; gross 2004; robeva & laubenbacher 2009 ). the only prerequisites for analyzing the simulations of an abm are a basic understanding of the underlying biology and, in some instances, knowledge of how to perform basic statistical calculations or generate graphical representations of results. many pre-built abms in simbio virtual labs and netlogo generate relevant graphs and/or produce spreadsheets or text files of the relevant data. thus, a student can utilize abms in life science courses without having to learn how to implement (by writing code) an abm. the prerequisites for learning how to implement abms are not as extensive as for other forms of mathematical models. the most essential prerequisite is some exposure to the fundamentals of computer programming such as understanding loop structures and conditional statements, implementing stochastic processes, and understanding how the order of executed operations within a program impacts the program's output. agent-based modeling software like netlogo (wilensky 1999 ) keeps model implementation and analysis relatively simple by providing built-in model visualization tools and automatically randomizing the order in which agents execute programmed operations. due to their appealing visual representation, abms can easily be used in the classroom to demonstrate biological processes ranging from chemical reactions to interactions among species, as if the model were simply an animation. however, using abms in this way is much like using a cell phone to hammer nails (q.v., theobald (2004) ): it may work in the desired fashion, but represents an utter waste of the tool's real potential. adding just one more step, the collection of model-generated data, transforms a passive learning experience into an active one. students can be asked to calculate means and variances, graph relationships between variables, discuss the sample size needed for reliable results, and generate quantitative predictions under different hypotheses, all in the context of a specific biological question. this can be done even in a large classroom with only a single, instructor-controlled computer, bridging the gap between lecture and lab. if students have access to individual computers, much more is possible. either individually or in small groups, students can use abms to collect and analyze their own data. free file-sharing resources such as googledocs make it easy to pool data across many groups, thereby crowd-sourcing problems that would be too large for any one group to handle on their own. in smaller classes and lab sections, individuals or groups can be assigned to model different scenarios (e.g., the interaction effects between different parameters), prompting discussions of the most appropriate parameter values and model settings. such models can even be extended into miniature research projects. for example, in a unit on community ecology, students might be assigned a question about two interacting species, then use online resources to find relevant information and parameter estimates, design and conduct a series of model runs, analyze their data using simple statistical techniques, and present their findings to the class. although abms can be used to simulate almost any biological process, meaningful exploration of a model typically requires a substantial commitment of class time and instructor engagement. as a result, except in modeling courses, it is seldom practical to incorporate abms into every lesson plan. in our experience, their educational value is highest for studying processes involving (a) substantial amounts of stochasticity, (b) nonlinear interactions, and/or (c) a defined spatial structure. the inclusion of abms in math courses generally comes in two different modes: (1) abms are taught as one modeling technique in a course covering multiple modeling techniques, and (2) the construction and analysis of abms are taught in a course where abms are the only type of modeling being used. however, due to the minimal prerequisites for learning agent-based modeling, both types of courses can potentially be offered with one or no prerequisite courses. bodine (2018) provides an example of a course where abms are taught as one type of discrete-time modeling technique in an undergraduate course designed for biology, mathematics, biomathematics, and environmental science majors that has no prerequisites beyond high-school algebra. an example of a course where abms are the only type of modeling being used is given by bodine (2019); this course has only a single prerequisite: either a prior math modeling course which introduces basic computer programming or an introduction to computer science course. the use of biological applications in teaching mathematical modeling (including modeling with abms) is often viewed as having a lower entry point with less new vocabulary overhead than other types of applications (e.g., those from physics, chemistry, or economics). in particular, models of population dynamics, even those involving interactions between multiple subpopulations or different species, usually do not require any new vocabulary for most students, which allows for a more immediate focus on mechanisms of population change and impact of interactions between individuals within the populations. video game vs. scientific process. in our experience, students sometimes react to an abm's many controls and visual output by treating the model as a video game, clicking buttons at random to see what entertaining patterns they can create. other students prefer to complete the exercise as quickly as possible by blindly following the prescribed instructions. neither of these approaches substantially engages students in thinking about the question(s) underlying the model, different strategies for collecting and analyzing data, or the model's limitations. to foster these higher-order cognitive skills, use of the abm should be explicitly framed as an example of the scientific process. this approach begins with a set of initial observations and a specific biological question. for example, what management practices would be most effective in controlling the invasive brown tree snake? after familiarizing themselves with the biological system, students propose hypotheses about the factors contributing to the snake's success on guam, suggest management strategies, and set model parameters that reflect their chosen strategy. finally, they run the model multiple times to collect data that allow them to measure their strategy's effectiveness. under this pedagogical approach, abms become a vehicle for design-ing and conducting a miniature research project, enabling experiments that would not otherwise be practical due to cost, logistics, or ethical considerations. the modeling exercise can also reinforce lessons on how scientific knowledge is constructed and tested (e.g., the three p's of science, namely, problem posing, problem solving, and peer persuasion (watkins 1992) ). as part of this exercise, students should engage in the process of deciding how best to collect, analyze, and present their data. for example, as part of the brown tree snake project, students might be asked to explore the practical steps that other pacific islands could take to prevent invasions or eradicate invaders. one group of students decides to focus on two different control measures: cargo checks of all inbound flights and deployment of poison baits around airstrips. following an overview of different statistical approaches, the students determine that a multiple regression analysis would best allow them to address their question. allowed only a limited 'budget' of 30 model runs, students settle on a factorial design using three treatment levels of cargo checks, three levels of baiting, and three replicates of each combination. the students set up a spreadsheet to record the data from each model run, graph their data in a scatter plot, and use software such as microsoft excel's analysis toolpak to conduct their analysis. a model as a caricature of the real world. students at early stages of their academic career often envision science as a collection of factual information and fixed procedures. students with this mindset may dismiss as useless any model that does not incorporate every detail of a particular biological system. by contrast, scientists recognize that models, whether mathematical, physical, or conceptual, are deliberate simplifications that attempt to capture certain properties of a biological system while ignoring others (dahlquist 2017) . for example, the standard epidemiological sir model (diagrammed in fig. 1a ) divides a population in three subpopulations (susceptible, infectious, and removed) while ignoring any potential heterogeneity within each subpopulation (e.g., age, treatment status, groups at higher risk for transmission). students will need to engage in activities that frame the abm as a hypothesis about the organization and function of a specific biological system (weisstein 2011) . after a description (and possible exploration) of the basic model, students can work in groups to suggest additional processes and variables that seem relevant to understanding the system. they can then choose one or two of the factors that they consider most important to addressing the question being asked. finally, they should consider how to modify the model to incorporate the chosen features. for example, a standard epidemiological model divides the host population into susceptible, infectious, and removed subpopulations, and studies the movement of individuals among these subpopulations (fig. 1a) . a group of students decides to modify this model to track a malarial epidemic. after discussing mortality rates, prevention and treatment options, and genetic and age-related variation in host susceptibility, the students decide to focus on incorporating vector transmission into their model. through guided discussion with the instructor, they realize that transmission now occurs in two directions: from infected vectors to susceptible hosts and from infected hosts to uninfected vectors. they therefore develop a schematic model (fig. 1b) that depicts these revised rules for each agent in the abm. even if the students do not actually build the corresponding computational model, this exercise in extending a model to reflect specific biological assumptions helps students understand the iterative process by which models are developed and the utility of even simple models to clarify key features of the system's behavior. algorithms vs. equations. the concept of an equation is introduced fairly early in mathematics education. in the united states, children can encounter simple algebraic equations in elementary school (common core 2019) and then continue to see increasingly complex equations in math classes through college. because of this long exposure to equations, the use of functions and systems of equations to model systems in the natural world feels "natural" or logical to students when they first encounter differential equation models or matrix models. abms, on the other hand, can seem confusing to students because they lack the ability to be expressed as an equation or set of equations. an abm is constructed as an algorithm describing when and how each agent interacts with their local environment (which may include other agents). often these interactions are governed by stochastic processes, and thus "decisions" by agents are made through the generation of random numbers. when first introducing students to abms, it can be helpful to teach students how to read and construct computer program flowcharts and to create a visual representation of what is occurring within an algorithm or portion of an algorithm (see bodine (2019) for example assignments that utilize program flowchart construction). in life science classes where abms are being analyzed but not constructed and implemented, a flow diagram can be a useful tool for conveying the order processes occur in the model. class discussions can question whether the order of processes make biological sense, and whether there are alternatives. in math modeling classes, the construction of flowcharts, even for simple abms, can help students elucidate where decision points are within the code, and what procedures are repeated through loop structures. the construction of flowcharts as students progress to more complicated abms can help students rectify the order of events in their implemented algorithm against the order in which events should be occurring biologically. whether students are working with abms in life science or math modeling classes, it is helpful for them to learn how to read and understand flow diagrams as they are often included in research publications that use agent-based modeling. describing an abm. much to the alarm of many math students beginning to develop abms, the formal description of an abm requires more than just writing the computer code. the standard method for describing abms in scientific publications, referred to as the overview, design concepts, and details (odd) protocol (grimm 2006 (grimm , 2010 railsback & grimm 2012) , often requires more than one page to fully describe even a simple abm. this can make abms seem overwhelming to students as they first begin to explore abms. in courses which teach the implementation and description of abms, instructors should take care not to introduce the computer code implementation simultaneous to the model description via the odd protocol. note that the introductory text on agent-based modeling by railsback & grimm (2012) does not introduce the concept of the odd protocol until chapter 3, which comes after the introduction and implementation (in netlogo) of a simple abm in chapter 2. in the course materials by bodine (2019) , the concept of the odd protocol is introduced prior to project 2, but the students are not required to write their own odd protocol description until the final project, once they have seen odd descriptions for multiple models. model implementation vs. the modeling cycle. courses that aim to teach methods in mathematical modeling often start with a discussion of the modeling cycle, which is typically presented as a flow diagram showing the loop of taking a real world question, representing it as a mathematical model, analyzing the model to address the question, and then using the results to ask the next question or refine the original real world question. figure 2 shows an example of a modeling cycle diagram. in courses where the mathematical models are encapsulated in one or a small handful of equations, the time spent on representing the real world as a mathematical model (the green box in fig. 2) is relatively short. the construction of abms, however, can be a fairly lengthy process, as abms are designed to simulate interactions between individuals and the local environment. when students are in the middle of constructing their first few abms, they often lose sight of where they are in the modeling cycle because the model implementation becomes a cycle of its own; a cycle of writing bits of code, testing the code to see if it runs and produces reasonable results, and repeating this process to slowly add all the components needed for the full algorithm of the abm. as students are first learning agent-based modeling, they need to be reminded often to pull back and view where they are in the modeling cycle; to see the flock for the boids, as it were. model validation. within the modeling cycle, there is a smaller cycle of model validation (see dashed line in fig. 2) . in a course where students are first introduced to the classic lotka-volterra predator-prey model, the students are usually first introduced to a predator-prey data set (like the 200-year data set of canadian lynx and snowshoe hare pelts purchased by the hudson bay company (maclulich 1937; elton & nicholson 1942) ), which shows the oscillating population densities of the predator and prey populations. when the students then simulate the lotka-volterra model for various parameter sets, they find that they are able to produce the same oscillating behavior of the predator and prey populations. this is a form of model validation, and a model which did not display this distinctive trend seen in the data would be considered invalid for that data set. a similar process must occur for validating abms against observed biological patterns. however, in order to engage in this validation process for abms, students must first understand how decisions of individual agents and interactions between neighboring agents can lead to system-level observable patterns, a phenomenon referred to as "emergence" or the "emergent properties" of an abm. the classic abm example for easily identifying an emergent property is a flocking example (a stock example of flocking exists in the netlogo models library, and is explored in chapter 8 of the abm textbook by railsback & grimm (2012) ). the concept of an emergent property can take a little while for students to fully understand. by definition, it is an observable outcome that the system was not specifically programmed to produced. in particular, it is not a summation of individual-level characteristics, and is typically not easily predicted from the behaviors and characteristics of the agents. for example, a variation of the reynolds (1987) flocking model is included in the netlogo library and is explored in railsback & grimm (2012, chapter 8 ). in the model, each agent moves based on three rules: 1. separate: maintain a minimum distance from nearby agents 2. align: move in the same direction as nearby agents 3. cohere: move closer to nearby agents where all agents move at the same speed and different schemes for determining who is a nearby agent can be used. additionally, there are model parameters for the minimum distance to be maintained between agents, and the degree to which an agent can turn left or right in a single time step in order to align, cohere, and separate. it is not immediately evident from this set of rules that individual agents might be able to form flocks (or swarm together), and indeed that system-level behavior does not emerge for all parameter sets. however, certain parameter sets do lead to the agents forming one or more flocks that move together through the model landscape. students observations: algorithms for pattern recognition. one of the most exciting moments for students when they first begin running simulations of an abm is seeing systemlevel patterns emerge before their eyes. one of the challenges for students (and agentbased modelers, in general) is to develop algorithms to identify and/or quantify the patterns we can easily identify by sight. for example, in the flocking abm discussed above, an observer watching the locations of individual agents change at each time step can easily see the formation of flocks. if however, the observer wanted to systematically explore the parameter space of the flocking abm and determine the regions of the parameter space under which flocking occurs (a process which might involve running hundreds or thousands of simulations), it would be a tedious and time-consuming task to physically watch each simulation and record whether the agents formed a flock or not. instead, the observer must choose the criteria that indicate the formation of a flock, and then determine a measure or an algorithm for determining whether a flock (or multiple flocks) have formed. in a course designed to teach the construction and analysis of abms, this is a point where students should be encouraged to be both creative and methodical about developing such measures and algorithms. the development of these observational measures and algorithms also provides a great opportunity for collaboration between students. it is especially helpful if students with a diversity of academic backgrounds can be brought together to brainstorm ideas; for instance, mixing students with various levels of exposure in mathematics, computer science, and biology can be very beneficial. rapid advances in computing power over the past decades have made agent-based modeling a feasible and appealing tool to study biological systems. in undergraduate mathematical biology education, there are multiple modes by which abms are utilized and taught in the classroom. in biology classrooms, abms can be used to engage students in hypothesis testing and in the experimental design and data collection processes of otherwise infeasible experiments, and to enable students to utilize models as a part of the scientific process. all of this can be done without students having to learn a programming language. by contrast, students who have had some exposure to computer programming can learn the construction, implementation, and analysis of agent-based models in a math or computer science modeling class. biological applications are ideal systems for first attempts at agent-based models as they typically do not necessitate learning extensive new vocabulary and theory to understand the basic components that need to be included in the model. throughout this article, we endeavored to articulate the benefits and challenges of including abms in undergraduate life science and math modeling courses. consistent with the bio2010 report (2003), we recommend that undergraduate biology and life science curricula be structured to ensure that all students have some exposure to mathematical modeling. we additionally recommend that this includes agent-based modeling. while not every student necessarily needs to take a course exclusively focused on agent-based modeling, every undergraduate biology student should have the opportunity to utilize an abm to perform experiments and to collect and analyze data. as we educate the next-generation of life scientists, let us empower them with the ability to utilize abms to simulate and better understand our complex and dynamic world. agent based modelling and simulation tools: a review of the state-of-art software in silico experiments of existing and hypotheticalcytokine-directed clinical trials using agentbased modeling agent-based models in translational systems biology aligning simulation models: a case study and results integration of biology, mathematics and computing in the classroom through the creation and repeated use of transdisciplinary modules. primus, in revision bae jw multi-cell agent-based simulation of the microvasculature to study the dynamics of circulating inflammatory cell trafficking typical pitfalls of simulation modeling-lessons learned from armed forces and business adaptive agents, intelligence, and emergent human organization: capturing complexity through agent-based modeling land operations division dsto, defence science and technology organisation, po box 1500 agent-based modeling: methods and techniques for simulating human systems multi-agent simulations and ecosystem management: a review the changing face of calculus: first-and second-semester calculus as college courses mathematics for the life sciences discrete math modeling with biological applications (course materials) agent-based modeling course materials agentbased modeling in systems pharmacology agent based-stock flow consistent macroeconomics: towards a benchmark model symmetric competition causes population oscillations in an individualbased model of forest dynamics immune system modelling and simulation characterizing emergent properties of immunological systems with multicellular rule-based computational modeling assessing the spatial impact on an agent-based modeling of epidemic control: case of schistosomiasis the integration of biology into calculus courses common core state standards initiatives an invitation to modeling: building a community with shared explicit practices decision-making in agent-based modeling: a current review and future prospectus the case for biocalculus: design, retention, and student performance the ten-year cycle in numbers of the lynx in canada a (2020) how epidemics like covid-19 end (and how to end them faster) mathematical games-the fantastic combinations of john conway's new solitaire game getting at systemic risk via an agent-based model of the housing market review: agent-based modeling of morphogenetic systemsadvantages and challenges agent-based model history and development. in: economics for a fairer society-going back to basics using agent-based models simulation of biological cell sorting using a two-dimensional extended potts model simulating properties of in vitro epithelial cell morphogenesis a standard protocol for describing individual-based and agent-based models the odd protocol: a review and update individual-based modeling and ecology interdisciplinarity and the undergraduate biology curriculum: finding a balance a systematic review of agent-based modelling and simulation applications in the higher education domain agent-based modelling. history, essence, future the history, philosophy, and practive of agent-based modeling and the development of the conceptual model for simulation diagram. doctoral dissertation a survey of agent-based modeling practices agent-based models and microsimulation advancing microbial sciences by individual-based modelling cellular potts modeling of complex multicellular behaviors in tissue morphogenesis virtual biology lab: an inquiry-based learning environment transmission of infectious diseases: data, models, and simulations disease transmission dynamics on networks: network structure versus disease dynamics an interactive computer model for improved student understanding of random particle motion and osmosis idynomics: next-generation individual-based modelling of biofilms hybrid simulation models-when, why, how? modelling and simulation software to support individual-oriented ecological modelling tutorial on agent-based modelling and simulation fluctuations in the numbers of the varying hare (lepus americanus) agent-based land-use models: a review of applications the beginning of the monte carlo method the monte carlo method episims: large-scale agent-based modeling of the spread of disease bio2010: transforming undergraduate education for future research biologists. the national academies press an evolutionary theory of economic change agent-based computing from multi-agent systems to agent-based models: a visual survey artificial economic life: a simple model of a stockmarket retrospectives: the ethology of homo economicus an agent-based approach for modeling dynamics of contagious disease spread an active learning exercise for introducing agent-based modeling agent-based and individual-based modeling: a practical introduction flocks, herds and schools: a distributed behaviour model multiscale computational analysis of xenopus laevis morphogenesis reveals key insights of systems-level behavior mathematical biology education: beyond calculus identifying control mechanisms of granuloma formation during m. tuberculosis infection using an agent-based model dynamic models of segregation an agent-based model of the spatial distribution and density of the santa cruz island fox. integrated population biology and modeling part b undergraduate module on computational modeling: introducing modeling the cane toad invasion why outbreaks like coronavirus spread exponentially, and how to "flatten the curve phenotypic transition maps of 3d breast acini obtained by imaging-guided agent-based modeling 29+ evidences for macroevolution intrinsic and extrinsic noise of gene expression in lineage trees agent-based modeling of multicell morphogenic processes during development agent-based modeling in public health: current applications and future directions learning by modeling: insights from an agent-based model of university-industry relationships intelligent machinery a heretical theory statistical methods in neutron diffusion ulam s (1950) random processes and transformations simulating tissue mechanics with agent-based models: concepts, perspectives and some novel results perspective: dimensions of the scientific method press) von neumann j, morgenstern o (1944) theory of games and economic behaviour multi-scale modeling in morphogenesis: a critical analysis of the cellular potts model teaching students to think like scientists: software enables experiments to be conducted that would be impossible in a laboratory simulating cancer growth with multiscale agent-based modeling building mathematical models and biological insights in an introductory biology course an introduction to agent-based modeling: modeling natural, social, and engineers complex systems with netlogo agent-based modeling: an introduction and primer model reduction for agent-based social simulation: coarse-graining a civil violence model publisher's note springer nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations key: cord-009481-6pm3rpzj authors: parnell, gregory s.; smith, christopher m.; moxley, frederick i. title: intelligent adversary risk analysis: a bioterrorism risk management model date: 2009-12-11 journal: risk anal doi: 10.1111/j.1539-6924.2009.01319.x sha: doc_id: 9481 cord_uid: 6pm3rpzj the tragic events of 9/11 and the concerns about the potential for a terrorist or hostile state attack with weapons of mass destruction have led to an increased emphasis on risk analysis for homeland security. uncertain hazards (natural and engineering) have been successfully analyzed using probabilistic risk analysis (pra). unlike uncertain hazards, terrorists and hostile states are intelligent adversaries who can observe our vulnerabilities and dynamically adapt their plans and actions to achieve their objectives. this article compares uncertain hazard risk analysis with intelligent adversary risk analysis, describes the intelligent adversary risk analysis challenges, and presents a probabilistic defender–attacker–defender model to evaluate the baseline risk and the potential risk reduction provided by defender investments. the model includes defender decisions prior to an attack; attacker decisions during the attack; defender actions after an attack; and the uncertainties of attack implementation, detection, and consequences. the risk management model is demonstrated with an illustrative bioterrorism problem with notional data. toward risk-based regulations, specifically using pra to analyze and demonstrate lower cost regulations without compromising safety. (7, 8) research in the nuclear industry has also supported advances in human reliability analysis, external events analysis, and common cause failure analysis. (9−11) more recently, leaders of public and private organizations have requested risk analyses for problems that involve the threats posed by intelligent adversaries. for example, in 2004, the president directed the department of homeland security (dhs) to assess the risk of bioterrorism. (12) homeland security presidential directive 10 (hspd-10): biodefense for the 21st century, states that " [b] iological weapons in the possession of hostile states or terrorists pose unique and grave threats to the safety and security of the united states and our allies" and charged the dhs with issuing biennial assessments of biological threats to "guide prioritization of our on-going investments in biodefense-related research, development, planning, and preparedness." a subsequent homeland security presidential directive 18 (hspd-18): medical countermeasures against weapons of mass destruction directed an integrated risk assessment of all chemical, biological, radiological, and nuclear (cbrn) threats. (13) the critical risk analysis question addressed in this article is: are the standard pra techniques for uncertain hazards adequate and appropriate for intelligent adversaries? as concluded by the nrc (2008) study on bioterrorism risk analysis, we believe that new techniques are required to provide credible insights for intelligent adversary risk analysis. we will show that treating adversary decisions as uncertain hazards is inappropriate because it can provide a different risk ranking and may underestimate the risk. in the rest of this section, we describe the difference between natural hazards and intelligent adversaries and demonstrate, with a simple example, that standard pra applied to attacker's intent may underestimate the risk of an intelligent adversary attack. in the second section, we describe a canonical model for resource allocation decision making for an intelligent adversary problem using an illustrative bioterrorism example with notional data. in the third section, we describe the illustrative analysis results obtained from the model and discuss the insights they provide for risk assessment, risk communication, and risk management. in the fourth section, we describe the benefits and limitations of the model. finally, we discuss future work and our conclusions. we believe that risk analysis of uncertain hazards is fundamentally different than risk analysis of intelligent adversaries. (14, 15) some of the key differences are summarized in table i . (16) a key difference is historical data. for many uncertain events, both natural and engineered, we have not only historical data of extreme failures or crises, but many times we can replicate events in a laboratory environment for further study (engineered systems) or analyze using complex simulations. intelligent adversary attacks have a long historical background, but the aims, events, and effects we have recorded may not prove a valid estimate of future threat because of changes in adversary intent and capability. both uncertain hazard risks of occurrence and geographical risk can be narrowed down and identified concretely. intelligent adversary targets vary by the goals of the adversary and can be vastly dissimilar between adversary attacks. information sharing between the two events differs dramatically. after hurricanes or earthquakes, engineers typically review the incident, publish results, and improve their simulations. sometimes after intelligent adversary attacks, or near misses, the situation and conduct of the attack may involve critical state vulnerabilities and protected intelligence means. in these cases, intelligence agencies may be reluctant to share complete information even with other government agencies. the ability to influence the event is also different. though we can prepare, we typically have no way of influencing the natural event to occur or not occur. on the other hand, governments may be able to affect the impact of terrorism attacks by a variety of offensive, defensive, and recovery measures. in addition, adversary attacks can take on so many forms that one cannot realistically defend/respond/recover/etc. against all types of attacks. although there have been efforts to use event tree technologies in intelligent adversary risk analysis (e.g., btra), many believe that this approach is not credible. (19) the threat from intelligent adversaries comes from a combination of both intent and capability. we believe that pra still has an important role in intelligent adversary risk analysis for assessment of the capabilities of adversaries, the vulnerabilities of potential targets, and potential consequences of attacks. however, intent is not a some cities may be considered riskier than others (e.g., new york city, washington), but terrorists may attack anywhere, any time. information sharing: asymmetry of information: new scientific knowledge on natural hazards can be shared with all the stakeholders. governments sometimes keep secret new information on terrorism for national security reasons. natural event: intelligent adversary events: to date, no one can influence the occurrence of an extreme natural event (e.g., an earthquake). governments may be able to influence terrorism (e.g., foreign policy; international cooperation; national and homeland security measures). government and insureds can invest in well-known mitigation measures. attack methodologies and weapon types are numerous. local agencies have limited resources to protect potentially numerous targets. federal agencies may be in a better position to develop better offensive, defensive and response strategies. modified from kunreuther. (17, 18) 431-461. (18) factor in natural hazard risk analysis. in intelligent adversary risk analysis, we must consider the intent of the adversary. the adversary will make future decisions based on our preparations, its objectives, and information about its ability to achieve its objectives that is dynamically revealed in a scenario. bier et al. provides an example of addressing an adversary using a defender-attacker game theoretic model. (20) nrc provides three examples of intelligent adversary models. (16) we believe it will be more useful to assess an attacker's objectives (although this is not a trivial task) than assigning probabilities to their decisions prior to the dynamic revelation of scenario information. we believe that modeling adversary objectives will provide greater insight into the possible actions of opponents rather than exhaustively enumerating probabilities on all the possible actions they could take. furthermore, we believe the probabilities of adversary decisions (intent) should be an output of, not an input to, risk analysis models. (16) this is a principal part of game theory as shown in aghassi et al. and jain et al. (21, 22) to make our argument and our proposed alternative more explicit, we use a bioterrorism illustrative example. in response to the 2004 hspd, in october 2006, the dhs released a report called the bioterrorism risk assessment (btra). (19) the risk assessment model contained a 17-step event tree (18 steps with consequences) that could lead to the deliberate exposure of civilian populations for each of the 27 most dangerous pathogens that the center for disease control tracks (emergency.cdc.gov/bioterrorism) plus one engineered pathogen. the model was extremely detailed and contained a number of separate models that fed into the main btra model. the btra resulted in a normalized risk for each of the 28 pathogens, and rank-ordered the pathogens from most risky to least risky. the national research council (nrc) conducted a review of the btra model and provided 11 specific recommendations for improvement to the model. (16) in our example, we will use four of the recommendations: model the decisions of intelligent adversaries, include risk management, simplify the model by not assigning probabilities to the branches of uncertain events, and do not normalize the risk. the intelligent adversary technique we developed builds on the deterministic defenderattacker-defender model and is solved using decision trees. (16) because the model has been simplified to reflect the available data, the model can be developed in a commercial off-the-shelf (cots) software package, such as the one we used for modeling, dpl (www.syncopation.org). other decision analysis software may work as well. 4(23) event trees have been useful for modeling uncertain hazards. (24) however, there is a key difference in the modeling of intelligent adversary decisions that event trees do not capture. as norman c. rasmussen, the director of the 1975 reactor safety study that validated pra for use in nuclear reactor safety, states in a later article, while the basic assumption of randomness holds true for nuclear safety, it is not valid for human action. (25) the attacker makes decisions to achieve his or her objectives. the defender makes resource allocation decisions before and after an attack to try to mitigate vulnerabilities and consequences of the attacker's actions. this dynamic sequence of decisions made by first the defender, then an attacker, then again by the defender should not be modeled solely by assessing probabilities of the attacker's decisions. for example, when the attacker looks at the defender's preparations for their possible bioterror attack, it will not assign probabilities to its decisions; it chooses the agent and the target based on perceived ability to acquire the agent and successfully attack the target that will give it the effects it desires to achieve its objectives. (15) representing an attacker decision as a probability may underestimate the risk. consider the simple bioterrorism event tree given in fig. 1 with notional data. using an event tree, for each agent (a and b) there is a probability that an adversary will attack, a probability of attack success, and an expected consequence for each outcome (at the terminal node of the tree). the probability of success 4 a useful reference for decision analysis software is located on the orms website (http://www.lionhrtpub.com/orms/surveys/das/ das.html). involves many factors, including the probability of obtaining the agent and the probability of detection during attack preparations and execution. the consequences depend on many factors, including agent propagation, agent lethality, time to detection, and risk mitigation; in this example, the consequences range from 0 or no consequences to 100, the maximum consequences (on a normalized scale of consequences). calculating expected values in fig. 1 , we would assess expected consequences of 32. we would be primarily concerned about agent b because it contributes 84% of the expected consequences (30 × 0.9 = 27 for b out of the total of 32). looking at extreme events, we would note that the worst-case consequence of 100 has a probability of 0.05. however, adversaries do not assign probabilities to their decisions; they make decisions to achieve their objectives, which may be to maximize the consequences they can inflict. (26) if we use a decision tree as in fig. 2 , we replace the initial probability node with a decision node because this is an adversary decision. we find that the intelligent adversary would select agent a, and the expected consequences are 50, which is a different result than with the event tree. again, if we look at the extreme events, the worstcase event (100 consequences) probabilities are 0.5 for the agent a decision and 0.6 for the agent b decision. the expected consequences are greater and the primary agent of concern is now a. in this simple example, the event tree approach underestimates the expected risk and provides a different risk ranking. furthermore, the event tree example underestimates the risk of the extreme events. however, while illustrating important differences, this simple decision tree model does not sufficiently model the fundamental structure of intelligent adversary risk. this model has a large number of applications for homeland security. for example, it would be easy to see the use of this canonical model applied to a dirty bomb example laid out in rosoff and von winterfeldt (27) or any other intelligent adversary scenario. in this article, we show a use of a bioterrorism application. we believe that the canonical risk management model must have at least six components: the initial actions of the defender to acquire defensive capabilities, the attacker's uncertain acquisition of the implements of attack (e.g., agents a, b, and c), the attacker's target selection and method of attack(s) given implement of attack acquisition, the defender's risk mitigation actions given attack detection, the uncertain consequences, and the cost of the defender actions. from this model, one could also conduct baseline risk analysis by looking at the status quo. in general, the defender decisions can provide offensive, defensive, or information capabilities. we are not considering offensive decisions such as preemption before an attack; instead, we are considering decisions that will increase our defensive capability (e.g., buy vaccine reserves) (28) or provide earlier or more complete information for warning of an attack (add a bio watch city). (29) in our defenderattacker-defender decision analysis model, we have the two defender decisions (buy vaccine, add a bio watch city), the agent acquisition for the attacker is uncertain, the agent selection and target of attack is another decision, the consequences (fatalities and economic) are uncertain, the defender decision after attack to mitigate the maximum possible casualties, and the costs of defender decisions are known. the defender risk is defined as the probability of adverse consequences and is modeled using a multiobjective additive model similar to multiobjective value models. (30) we have assumed that the defender minimizes the risk and the attacker maximizes the risk. 5 we implemented this model as a decision tree (fig. 3 ) and an influence diagram (fig. 4) using dpl. the mathematical formulation of our model and the notional data are provided in the appendix the illustrative decision tree model (figs. 3 and 4) begins with decisions that the defender (united states) makes to deter the adversary by reducing the vulnerabilities or be better prepared to mitigate a bioterrorism attack of agents a, b, or c. we modeled and named the agents to represent notional bioterror agents using the cdc's agent categories in table ii . for example, agent a represents a notional agent from category a. table iii provides a current listing of the agents by category. there are many decisions that we could model; however, for our simple illustrative example, we chose to model notional decisions about the bio watch program for agents a and b and the bioshield vaccine reserve for agent a. bio watch is a program that installs and monitors a series of passive sensors within a major metropolitan city. (29) the bioshield program is a plan to purchase and store vaccines for some of the more dangerous pathogens. (28) the defender first decides whether or not to add another city to the bio watch program. if that city is attacked, this decision could affect the warning time, which influences the response and, ultimately, the potential consequences of an attack. of course, the bio watch system does not detect every agent, so we modeled agent c to be the most effective agent that the bio watch system does not sense and provide additional warning. adding a city will also incur a cost in dollars for the united states. the second notional defender decision is the amount of vaccine to store for agent a. agent a is the notional agent that we have modeled with the largest probability of acquisition and potential consequences. the defender can store a percentage of what experts think we would need in a largescale biological agent attack. the more vaccine the united states stores, the fewer consequences we will have if the adversaries use agent a and we have sufficient warning and capability to deploy the vaccine reserve. however, as we store more vaccine, the costs for purchasing and storage increase. for (31) category definition a the u.s. public health system and primary healthcare providers must be prepared to address various biological agents, including pathogens that are rarely seen in the united states. high-priority agents include organisms that pose a risk to national security because they: can be easily disseminated or transmitted from person to person; result in high mortality rates and have the potential for major public health impact; might cause public panic and social disruption; and require special action for public health preparedness. b second highest priority agents include those that: are moderately easy to disseminate; result in moderate morbidity rates and low mortality rates; and require specific enhancements of cdc's diagnostic capacity and enhanced disease surveillance. c third highest priority agents include emerging pathogens that could be engineered for mass dissemination in the future because of: availability; ease of production and dissemination; and potential for high morbidity and mortality rates and major health impact. simplicity's sake, each of the defender decisions cost the same amount; therefore, at the first budget level of us$ 10 million, the defender can only choose to one decision. after the defender has made its investment decisions, which we assume are known to the attacker, the attacker makes two decisions: the type of agent and the target. we will assume that the attacker has already made the decision to attack the united states with a bioterror agent. in our model, there are three agents it can choose, although this can be increased to represent the other pathogens listed in table iii . as stated earlier, if we only looked at the attacker decision, agent a would appear to be the best choice. agents b and c are the next two most attractive agents to the attacker, respectively. again, agents a and b can be detected by bio watch whereas agent c cannot. the attacker has some probability of acquiring each agent. if the agent is not acquired, the attacker cannot attack with that agent. in addition, each agent has a lethality associated with it, represented by the agent casualty factor. finally, each agent has a different probability of being detected over time. generally, the longer it takes for the agent to be detected, the more consequences the united states will suffer. the adversary also decides what size of population to target. generally, the larger the population targeted, the more potential consequences could result. the attacker's decisions affect the maximum possible casualties from the scenario. however, regardless of the attacker's decisions, there is some probability of actually attaining a low, medium, or high percentage of the maximum possible casualties. this sets the stage for the next decision by the defender. after receiving warning of an attack, the defender decides whether or not to deploy the agent a vaccine reserve. this decision depends upon how much of the vaccine reserve the united states chose to store, whether the attacker used agent a, and the potential effectiveness of the vaccine given timely attack warning. in addition, there is a cost associated with deploying the vaccine reserve. again, for simplicity's sake, the cost for every defender decision is the same, thus forcing the defender to only choose the best option(s) for each successive us$ 10 million increase in budget up to the maximum of us$ 30 million. in our model (fig. 4) , we have two types of consequences: casualties and economic impact. given the defender-attacker-defender decisions, the potential casualties and the economic impact are assessed. casualties are based on the agent, the population attacked, the maximum potential casualties, the warning time given to the defender, and the effectiveness of vaccine for agent a (if the agent a is the agent and the vaccine is used). economic effects are modeled using a linear model with a fixed economic cost that does not depend on the number of casualties and a variable cost of the number of casualties multiplied by the cost per casualty. of course, the defender would like potential consequences (risk) given an attack to be low, whereas the attacker would like the potential consequences (risk) to be high. our economic consequences model was derived using a constant and upper bound from wulf et al. (34) the constant cost we used is $10 billion, and from the upper bound, also given in wulf et al., we derived the cost per casualty. (34) we believe this fixed cost is reasonable because when looking at the example of the anthrax letters of 2001, experts estimate that although there were only 17 infected and five killed, there was a us$ 6 billion cost to the united states. (35) in this tragic example, there was an extremely high economic impact even when the casualties were low. each u.s. defender decision incurs a budget cost. the amount of money available to homeland security programs is limited by a budget determined by the president and congress. the model will track the costs incurred and only allows spending within the budget (see the appendix). we considered notional budget levels of us$ 0 million, us$ 10 million, us$ 20 million, and us$ 30 million. normally, a decision tree is solved by maximizing or minimizing the expected attribute at the terminal branches of the tree. in our model however, the defender risk depends on the casualty and economic consequences given an attack. we use multiple objective decision analysis with an additive value (risk) model to assign risk to the defender consequences. 6 the defender is minimizing risk and the attacker is maximizing risk. we assign a value of 0.0 to no consequences and a value of 1.0 to the worst-case consequences in our model. we model each consequence with a linear risk function and a weight (see the appendix). the risk functions measure returns to scale on the consequences. of course, additional consequences could be included and different shaped risk curves could be used. some of the key assumptions in our model are listed in table iv (the details are in the appendix) along with some possible alternative assumptions. given different assumptions, the model may produce different results. we model the uncertainty of the attacker's capability to acquire an agent with a probability distribution and we vary detection time by agent. clearly, other indications and warnings exist to detect possible attacks. these programs could be included in the model. we model three defender decisions: add a bio watch city for agents a and b, increase vaccine reserve for agent a, and deploy agent a. we assume limited decision options (i.e., 100% storage of vaccine a, 50% storage, 0% storage), but other decisions could be modeled (e.g., other levels of storage, storing vaccines for other agents, etc). we used one casualty model for all agents. other casualty and economic models could be used. finally, our model makes some assumptions about objectives. in the first of these we assume that the risks important to the defender are the number of casualties and the economic impact, but additional measures could be used. second, we assume defenders and attackers have a diametrically opposed view of all of the objectives. clearly, we could model additional objectives. in addition, we made some budget assumptions, which could be improved or modified. we assumed a fixed budget, but this budget could be modeled with more detailed cost models (e.g., instead of a set amount to add a city to the bio watch program, the budget could reflect different amounts depending upon the city and the robustness of the sensors installed). finally, our model results in a risk of a terrorist attack; the same methodology for a defender-attacker-defender decision tree can be used to determine a utility score instead of a risk; an example of this is in keeney. (15) one thing to consider when altering or adding to the assumptions is the number of strategies the model evaluates. currently, the canonical model has 108 different strategies to evaluate (table v) . with more complexity, the number of strategies that would need to be evaluated could grow significantly. largescale decision trees can be solved with monte carlo simulation. after modeling the canonical problem, we obtained several insights. first, we found that in our model economic impact and casualties are highly correlated. higher casualties result in higher economic impact. other consequences, for example, psychological consequences, could also be correlated with casualties. second, a bioterror attack could have a large economic impact (and psychological impact), even if casualties are low. the major risk analysis results are shown in fig. 5 . risk shifting occurs in our decision analysis model. in the baseline (with no defender spending), agent a is the most effective agent for the attacker to select and, therefore, the agent against which the defender can protect if the budget is increased. as we improve our defense against agent a, at some point the attacker will choose to attack using agent b. the high-risk agent will have shifted from agents a to b. as the budget level continues to increase, the defender adds a city to the bio watch program and the attackers choose to attack with agent c, which bio watch cannot detect. we use notional data in our model, but if more realistic data were used, the defender could determine the cost/benefit ratios of additional risk reduction decisions. this decision model uses cots software to quantitatively evaluate the potential risk reductions associated with different options and make cost-benefit decisions. fig. 5 provides a useful summary of the expected risk. however, it is also important to look at the complementary cumulative distribution (fig. 6) to better understand the likelihood of extreme outcomes. for example, the figure shows that spending us$ 0 or us$ 10 million gives the defender a 10% chance of zero risk, whereas spending us$ 20 or us$ 30 million gives the defender an almost 50% chance of having zero risk. the best risk management result would be that option 4 deterministically or stochastically dominates (sd) option 3, option 3 sd option 2, and option 2 sd option 1. the first observation we note from fig. 6 is that options 2, 3, and 4 stochasically dominate 1 because option 1 has a higher probability for each risk outcome. a second observation is that while option 4 sd option 3, option 4 does not sd option 2 because option 4 has a larger probability of yielding a risk level of 0.4. along the x-axis, one can see the expected risk (er) of each alternative. this expected risk corresponds to the expected value of risk from the budget versus risk rainbow diagram in fig. 5 . this example illustrates a possibly important relationship necessary for understanding and communicating how the budget might affect the defender's risk and choice of options. risk managers can run a value of control or value of correlation diagram to see which nodes most directly affect the outcomes and which are correlated (fig. 7) . because we only have two uncertainty nodes in our canonical model, the results are not surprising. the graphs show that the ability to acquire the agent is positively correlated with the defender risk. as the probability of acquiring the agent increases, so does defender risk. in addition, the value of control shows the amount of risk that could be reduced given perfect control over each probabilistic node, and that it is clear that acquiring the agent would be the most important variable for risk managers to control. admittedly, this is a basic example, but with a more complex model, analysts could determine which nodes are positively or negatively correlated with risk and which uncertainties are most important. using cots software also allows us to easily perform sensitivity analysis on key model assumptions. from the value of correlation and control above, the probability of acquiring the agent was highly and positively correlated with defender risk and had the greatest potential for reducing defender risk. we can generate sensitivity analysis such as rainbow diagrams. the rainbow diagram (fig. 8) shows the decision changes as our assumption about the probability of acquiring agent a increases. the different shaded regions represent different decisions, for both the attacker and the defender. this rainbow diagram was produced using a budget level of us$ 20 million, so in the original model, the defender would choose not to add a city to bio watch, store 100% of vaccine for agent a, but not choose to deploy it because the attacker chose to use agent b. if the probability of acquiring agent a was low enough (in section a from fig. 8) , we see that the attacker would choose to use agent c because we have spent our money on adding another city to bio watch, which is the only thing that affects both agents a and b, but not agent c. as the probability of acquiring agent a increases, both the attacker's and the defender's optimal strategies change. our risk management decision depends on the probability that the adversary acquires agent a. risk analysis of intelligent adversaries is fundamentally different than risk analysis of uncertain hazards. as we demonstrated in section 1.3, assigning probabilities to the decisions of intelligent adversaries can underestimate the potential risk. decision tree models of intelligent adversaries can provide insights into the risk posed by intelligent adversaries. the defender-attacker-defender decision analysis model presented in this article provides four important benefits. first, it provides a risk assessment (the baseline or status quo) based on defender and attacker objectives and probabilistic assessment of threat capabilities, vulnerabilities, and consequences. second, it provides information for risk-informed decision making about potential risk management options. third, using cots software, we can provide a variety of very useful sensitivity analysis. fourth, although the model would be developed by a team, the risk analysis can be conducted by one risk analyst with an understanding of decision trees and optimization and training on the features of the cots software. the application of risk assessment and risk management techniques should be driven by the goals of the analysis. in natural hazard risk analysis, there is value in performing risk assessment without risk management. some useful examples are "unfinished business," a report from the epa and the 2008 u.k. national risk register. (36, 37) in intelligent adversary risk analysis, the defender-attacker-defender decision analysis can provide essential information for risk management decision making. in our example, risk management techniques are important, and this type of model provides insights about resource allocation decisions to reduce or shift risk. in addition, with budget set to us$ 0, the model can be used to assess the baseline risk. as the budget increases, the model clearly shows the best risk management decisions and the associated risk reduction. this model enables the use of cots risk analysis software. in addition, the use of cots software enables the use of standard sensitivity analysis tools to provide insights into areas in which the assumptions are critical or where the model should be improved or expanded. currently, many event tree models including the dhs btra event tree require extensive contractor support to run, compile, and analyze. (16) although one would still need a multidisciplinary team to create the model, once completed the defenderattacker-defender decision analysis model is usable by a single risk analyst who can provide near realtime analysis results to stakeholders and decisionmakers as long as the risk analyst understands the risk management options, decision trees, optimization, and has training in the cots tool. the technique we advocate in this article has limitations. some of the limitations of this model are the same as those of event trees. there are limitations on the number of agents used in the models. we easily modeled 28 bioagents with reasonable run times, but more agents could be modeled. in addition, there are challenges in assessing the probabilities of uncertain events, for example, the probability that the attacker acquires agent a. next, there is a limitation in the modeling of the multiple consequences. another limitation may be that to get more realistic results, we may have to develop "response surface" models of more complex consequence models. these limitations are shared by event trees and decision trees. however, decision trees also have some limitations that are not shared by event trees. first, only a limited number of risk management decisions can realistically be modeled. therefore, care must be used to choose the most appropriate set of potential decisions. (15, 18) in addition, there may be an upper bound on the number of decisions or events that can be modeled in cots software. it is important to note that it may be difficult to determine an objective function for the attacker. as mentioned before, there is a tradeoff in replacing the probabilities assigned to what an attacker might do (event tree approach) with attacker objectives (decision tree approach). we believe it is easier to make informed assessments about the objectives of adversaries than to assess the probabilities of their future actions. however, we need more research on assessing the robustness of risk management decisions to assumptions about adversary objectives. finally, successful model operation and interpretation requires trained analysts who understand decision analysis and defenderattacker-defender optimization. this article has demonstrated the feasibility of modeling intelligent adversary risk using defenderattacker-defender decision analysis. table iv and section 2.7 identified several alternative modeling assumptions that could be considered. we can modify and expand our assumptions to increase the complexity and fidelity of the model. the next step is to use the model with the best data available on the agents of concern and a proposed set of potential risk management options. use of our defender-attacker-defender model does not require a major intelligent adversary research program; it requires only the willingness to change. (16) much of the data used for event tree models can be used in the decision analysis model. assessing probabilities of attacker decisions will not increase our security but defender-attacker-defender decision analysis models can provide a sound assessment of risk and the essential information our nation needs to make risk-informed decisions. g.s.p. is grateful for the many helpful discussions on intelligent adversary risk analysis with his colleagues on the 2008 nrc committee and the defender-attacker-defender research of jerry brown and his colleagues at the naval postgraduate school. the authors are grateful for the dpl modeling advice provided by chris dalton of syncopation. the authors thank roger burk at the united states military academy for his useful reviews and suggestions. finally, the authors thank the area editor and reviewers for very detailed comments and suggestions that have helped us improve our article. this model is a multiobjective decision analysis/game theory model that allows for risk management at the u.s. governmental level in terms of budgeting and certain bioterror risk mitigation decisions. the values for probabilities as well as factors are notional and could easily be changed based on more accurate data. it uses the starting u.s. (defender) decisions of adding a city to the bio watch program (or not) and the percent of storing an agent in the nation's vaccine reserve program to set the conditions for an attacker decision. the attacker can choose which agent to use as well as what size of population to target. there is some unpredictability in the ability to acquire the agent as well as the effects of the agent given the defender and attacker decisions. finally, the defender gets to choose whether to deploy the vaccine reserve to mitigate casualties. the model tracks the cost for each u.s. decision and evaluates them over a specified budget. the decisions cannot violate the budget without incurring a dire penalty. the objectives that the model tracks are u.s. casualties and impact to the u.s. economy. they are joined together using a value function with weights for each objective. we outline our model using a method suggested by brown and rosenthal. (38) probabilistic risk assessment: reliability engineering, design, and analysis risk analysis in engineering and economics risk modeling, assessment, and management reactor safety study: assessment of accident risk in u.s. commercial nuclear plants nuclear regulatory commission (usnrc) fault tree handbook understanding risk management: a review of the literature and industry practice a survey of risk assessment methods from the nuclear, chemical, and aerospace industries for applicability to the privatized vitrification of hanford tank wastes procedural and submittal guidance for the individual plant examination of external events (ipeee) for severe accident vulnerabilities procedure for analysis of common-cause failures in probabilistic safety analysis a technique for human error analysis (atheana) homeland security presidential directive 10 homeland security presidential directive 18 [hspd-18]: medical countermeasures against weapons of mass destruction guiding resource allocations based on terrorism risk modeling values for anti-terrorism analysis department of homeland security's bioterrorism risk assessment: a call for change, committee on methodological improvements to the department of homeland security's biological agent risk analysis insurability of (mega)-terrorism risk: challenges and perspectives. report prepared for the oecd task force on terrorism insurance, organization for economic cooperation and development integrating risk management with homeland security and antiterrorism resource allocation decision-making. chapter 10 in kamien d (ed). the mcgraw-hill handbook of homeland security fort detrick, md: biological threat characterization center of the national biodefense analysis and countermeasures center choosing what to protect: strategic defensive allocation against an unknown attacker robust game theory robust solutions in stackelberg games: addressing boundedly rational human preference models. association for the advancement of artificial intelligence workshop: 55-60 syncopation software. available at probabilistic modeling of terrorist threats: a systems analysis approach to setting priorities among countermeasures probabilistic risk assessment: its possible use in safeguards problems. presented at the institute for nuclear materials management meeting nature plays with dice-terrorists do not: allocating resources to counter strategic versus probabilistic risks a risk and economic analysis of dirty bomb attacks on the ports of los angeles and long beach project bioshield: protecting americans from terrorism the bio watch program: detection of bioterrorism strategic decision making: multiobjective decision analysis with spreadsheets bioterrorist agents/diseases definitions by category center for disease control (cdc) emerging and re-emerging infectious diseases strategic alternative responses to risks of terrorism world at risk: report of the commission on the prevention of wmd proliferation and terrorism unfinished business: a comparative assessment of environmental problems optimization tradecraft: hard-won insights from real-world decision support. interfaces key: cord-265017-byyx2y47 authors: ryan, jeffrey r. title: seeds of destruction date: 2016-03-25 journal: biosecurity and bioterrorism doi: 10.1016/b978-0-12-802029-6.00001-3 sha: doc_id: 265017 cord_uid: byyx2y47 this chapter provides the reader with an understanding and appreciation for the scope and importance of biological threats and the opportunity to see where they may be and have become the desire of terrorist groups and the makings of weapons of mass destruction. the history of biological warfare is covered in depth. these major events are important in helping us understand the issues related to using biological substances against an adversary. the difference between biosecurity and biodefense are explained and then related to homeland security and homeland defense, respectively. this chapter also details how expensive these programs are, with nearly $80 billion having been spent on civilian biodefense since fiscal year 2001 in the united states alone. as discussed herein, there is a significant difference in the reality and the potential of bioterrorism. bioterrorism on a large scale is a low-probability event. bioterrorism on a small scale is a fairly routine occurrence with little potential. biological threats remain very much in the news. recent examples, such as laboratory incidents, the ebola outbreak of 2014, and other emerging threats, are covered in this chapter. the dawning of the 21st century will be characterized as the age of terrorism. terrorism has affected most of us in one way or another. the shocking images of the september 11, 2001, attacks remind us of just how dramatic and devastating terrorism can be. in most developed countries, the concept of bioterrorism and many of the words associated with it are widely recognized. in the united states, bioterrorism became a household word in october 2001, when bacillus anthracis (the causative agent of anthrax) spores were introduced into the us postal service system by several letters dropped into a mailbox in trenton, new jersey (see fig. 1 .1). these letters resulted in 5 deaths from pulmonary anthrax and 17 other cases of inhalation and cutaneous anthrax (thompson, 2003) . in the weeks and months that followed, first responders were called to the scene of thousands of "white powder" incidents that came as a result of numerous hoaxes, mysterious powdery substances, and just plain paranoia (beecher, 2006) . public health laboratories all over the united states were inundated with samples collected from the scene of these incidents. testing of postal facilities, us senate office buildings, and news-gathering organizations' offices occurred. between october and december 2001 the centers for disease control and prevention (cdc) laboratories successfully and accurately tested more than 125,000 samples, which amounted to more than 1 million separate bioanalytical tests (cdc, 2015) . henceforth there has been a national sense of urgency in preparedness and response activities for a potential act of bioterrorism. humankind has been faced with biological threats since we first learned to walk upright. in his thought-provoking book guns, germs and steel, dr. jared diamond points out the epidemiological transitions we have faced since we were hunters and gatherers. more than 10,000 years ago the human experience with biological peril was mostly parasitic diseases that only affected individuals. after that, human societies began to herd and domesticate animals. the development of agriculture allowed for population growth and a shift from small tribal bands to a concentration of people into villages. larger groups of people could stand up to smaller elements, thereby enabling them to successfully compete for resources and better defend the ground that they held. agriculture also brought some deadly gifts: animal diseases that also affected man (zoonotic diseases), outbreaks of disease due to massing of people and lack of innate immunity, and a growing reliance on animal protein (diamond, 1999) . for ages human societies and cultures have been looking for a competitive advantage over their adversaries. advances in weapons of all types and explosives allowed military forces to defeat their enemies overtly on the battlefield and covertly behind the lines. technologies leading to nuclear, biological, and chemical weapons have also been exploited. indeed, each has been used legitimately and illegitimately on different scales to bring about a change in the tactics, the military situation, or the political will to face an enemy in battle. biological agents are no exception to this rule. as such, biowarfare (biological warfare) has a historical aspect to it that must be considered here because advances in the use of biological agents over the last century are one of the main reasons why bioterrorism exists today. when president richard m. nixon said, in november 1969, that "mankind already holds in its hands too many of the seeds of its own destruction," he was signing an executive order putting an end to the united states' offensive capabilities for waging biowarfare. it is arguable that this statement foretold the potential doom we might all face when then state-of-the-art technologies became commonplace techniques in laboratories all over the world today. this chapter accordingly derives its name from the preceding quote and should serve to remind the reader that the seeds we sowed so long ago have now sprouted. the question remains: how shall they be reaped? bioterrorism is the intentional use of microorganisms or toxins derived from living organisms to cause death or disease in humans or the animals and plants on which we depend. biosecurity and biodefense programs exist largely because of the potential devastation that could result from a large-scale act of bioterrorism. civilian biodefense funding (cbf) reached an all-time high after the anthrax attacks of 2001. conversely, the reality of the situation is that these well-intended programs cost taxpayers billions of dollars each year. rapid detection biothreat pathogen tools are available to assist responders with on-site identification of a suspicious substance. in addition, biosecurity and biodefense are "big business" in the private sector. security measures to protect agriculture and certain vulnerable industries from acts of bioterrorism and natural biological threats are also in place. detailed reports published in the journal biosecurity and bioterrorism (schuler, 2005; lam et al., 2006; sell and watson, 2013) show that us government cbf between fiscal year (fy) 2001 and fy2014 amounted to more than $78 billion. comparing fy2001 to fy2005, there was an increase in cbf from $420 million to $7.6 billion. the departments of health and human services and homeland security, which together account for approximately 88% of the fy2006 request, have remained relatively constant in their funding. other agencies, most notably the department of agriculture and the environmental protection agency, have been more variable. these two agencies saw increased budget requests in fy2006, focusing on programs that protect the nation's food and water supplies. civilian biodefense spending, not including special allocations for project bioshield, reached a consistent level of approximately $6 billion from fy2003 to fy2013 (sell and watson, 2013) . refer to table 1 .1 for a summary of the cbf budget for fy2010-14. bioshield is a program that was designed to give the united states new medical interventions (eg, vaccines, treatments) for diseases caused by several biothreat pathogens. when bioshield was conceived, it cost us taxpayers a total of $5.6 billion, which was metered out to the department of health and human services over a 10-year period. reports surfaced that suggest bioshield funds were being squandered and that few useful products were realized (fonda, 2006) . however, biothreat pathogen research and product development for unusual or rare diseases is fraught with numerous hurdles. this program will be addressed in chapter biosecurity programs and assets. the us postal service spent more than $800 million developing and deploying its biohazard detection system (bds). at the peak of its utilization, the us postal service was spending more than $70 million each year to operate and maintain the system. the bds is used only to provide early warning for the presence of a single biothreat pathogen, anthrax. furthermore, the system screens letter mail that comes from sources such as mailboxes and drops, which accounts for approximately 17% of all letter mail volume (schmid, 2006) . this model program and the technology it uses will be covered extensively in chapter consequence management and a model program. all of this seems rather incredible when comparing the level of funding given to one of the greatest biological threats of our time, the human immunodeficiency virus (hiv), which causes aids. an estimated 1.8 million people are currently living with hiv in the united states, with approximately 50,000 new infections occurring each (nih, 2015) compared with the $1.6 billion level of funding it receives for biodefense (sell and watson, 2013) . before delving into the subtleties of biosecurity and biodefense, one should explore the historical aspects of the use of biological agents in warfare and terrorism. the history presented here is not all inclusive. rather, it is a fair assessment of key events and characterizations that can be examined in other more comprehensive documents. pathogens and biological toxins have been used as weapons throughout history. some would argue that biological warfare began when medieval armies used festering corpses to contaminate water supplies. over several centuries this evolved into the development of sophisticated biological munitions for battlefield and covert use. these developments parallel advances in microbiology and include the identification of virulent pathogens suitable for aerosol delivery and large-scale fermentation processes to produce large quantities of pathogens and toxins. however, the history of biological warfare is shrouded by several confounding factors. first, it is difficult to verify alleged or attempted biological attacks. these allegations might have been part of a propaganda campaign, or they may have been due to rumor. regardless, some of the examples we have been given cannot be supported by microbiological or epidemiologic data. in addition, the incidence of naturally occurring endemic or epidemic diseases during that time complicates the picture so that attribution is impossible (christopher et al., 1997) . more important, our awareness that infectious diseases are caused by microbes does not go back very far in human history. germ theory, or the fact that infectious diseases are related to and caused by microorganisms, emerged after 1860 through the independent works of pasteur, lister, and koch (tortora et al., 1995) . therefore how could the attacking or defending commander know that the festering corpses might cause disease when people at that time thought that epidemics were related to "miasmas," the smell of decomposition, or heavenly "influences"? one need only consider the origin of certain disease names to appreciate this confusion. for instance, malaria gets its name from malaria, or "bad air" (ie, swamp gases; desowitz, 1991) . it was not until 1880 that we learned that the etiologic agents of malaria are protozoans in the genus plasmodium. the name influenza refers to the ancient belief that the disease was caused by a misalignment of the stars because of some unknown supernatural or cosmic influence (latin influentia). it was not until 1933 that we learned the flu was caused by the influenza virus (potter, 2001) . regardless of the lack of awareness of germs at the time, a few of the historic reports about the use of biological weapons in battle are worth noting here: • in the 6th century bc, assyrians poisoned enemy wells with rye ergot, a fungus. • in the 4th century bc, scythian archers tipped their arrows with blood, manure, and tissues from decomposing bodies. • in ad 1340, attackers hurled dead horses and other animals by catapult at the castle of thun l'eveque in hainault (northern france). castle defenders reported that "the stink and the air were so abominable…they could not long endure" and negotiated a truce. • in ad 1422 at karlstein in bohemia, attacking forces launched the decaying cadavers of men killed in battle over the castle walls. they also stockpiled animal manure in the hope of spreading illness. however, the defense held fast, and the siege was abandoned after 5 months. russian troops may have used the same tactic using the corpses of plague victims against the swedes in 1710. • in ad 1495 the spanish contaminated french wine with the blood of lepers. • in the mid-1600s a polish military general reportedly put saliva from rabid dogs into hollow artillery spheres for use against his enemies. • francisco pizarro reportedly gave smallpox virus-contaminated clothing to south american natives in the 15th century. • in a letter dated july 16, 1763, general jeffrey amherst, a british officer, approved the plan to spread smallpox to delaware indians (robertson, 2001) . amherst suggested the deliberate use of smallpox to "reduce" native american tribes hostile to the british (parkman, 1901 ). an outbreak of smallpox at fort pitt resulted in the generation of smallpox-contaminated materials and an opportunity to carry out amherst's plan. on june 24, 1763, one of amherst's subordinates gave blankets and a handkerchief from the smallpox hospital to the native americans and recorded in his journal, "i hope it will have the desired effect" (sipe, 1929 ). • the same tactic was used during the civil war by dr. luke blackburn, the future governor of kentucky. dr. blackburn infected clothing with smallpox and yellow fever virus, which he then sold to union troops. one union officer's obituary stated that he died of smallpox contracted from his infected clothing (guillemin, 2006) . as previously mentioned, scientists discovered microorganisms and made advances toward understanding that a specific agent causes a specific disease, that some are foodborne or waterborne, that an agent can cycle through more than one species, and that insects and ticks are the vectors of disease. furthermore, medical professionals established that wars, famines, and poverty opened populations to the risk of epidemics. once these links were established, we learned that we could apply control and intervention methods. scientific knowledge about disease transmission coupled with social stability and active public health campaigns aided human survival. it subsequently became possible for advanced populations to protect their citizens from the burden of some of the most insidious infectious diseases, such as plague, cholera, diphtheria, smallpox, influenza, and malaria. these epidemics swept across nations in previous centuries, hitting hardest in crowded urban centers and affecting mostly the poor (guillemin, 2006) . at the opening of the industrial revolution, public health in cities had improved, water and food sources were monitored by the state, and vaccines and drug therapies were being invented as further protection. with many childhood diseases conquered, more people were living longer, and they were now dying of more "civilized" diseases such as cancer, heart disease, and stroke (diamond, 1999) . in underdeveloped nations, public health did not develop; hence, epidemics were prevalent and continued to be devastating. the dichotomy between developed and developing nations remains marked by generally good health versus widespread, preventable epidemics (guillemin, 2006) . as western nations were taking advantage of innovations in public health and medicine to mitigate epidemics, their governments invented biological weapons as a means of achieving advantage in warfare (diamond, 1999) . the german military has the dubious honor of being the first example of using biological weapons following a state-sponsored program. however, during world war i, they used disease-causing organisms against animals, not people. the goal of their program was to interrupt the flow of supplies to the allied frontlines. to do this they targeted the packhorses and mules shipped from norway, spain, romania, and the united states. in 1915, dr. anton dilger, a german-american physician, developed a microbiology facility in washington, dc. dilger produced large quantities of anthrax and glanders bacteria using seed cultures provided by the imperial german government. at the loading docks, german agents inoculated more than 3000 animals that were destined for the allied forces in europe (wheelis, 1999) . from the german perspective, these attacks violated no international law. in addition, these activities were dwarfed by the atrocities of chemical warfare that was being waged on both sides of the line. to counter the german threat and explore the potential of air warfare the french sought to improve their integration of aerosols and bombs. at the same time as the french were signing the 1925 geneva protocol, they were developing a biological warfare program to complement the one they had established for chemical weapons during world war i (rosebury and kabat, 1947) . after world war i the japanese formed a "special weapons" section within their army. the section was designated unit 731. the unit's leaders set out to exploit chemical and biological agents. in 1936 they expanded their territory into manchuria, which made available "an endless supply of human experiment materials" (prisoners of war) for unit 731. biological weapon experiments in harbin, manchuria, directed by japanese general shiro ishii, continued until 1945. a post-world war ii autopsy investigation of 1000 victims revealed that most were exposed to aerosolized anthrax. more than 3000 prisoners and chinese nationals may have died in unit 731 facilities. in 1939 the japanese military poisoned soviet water sources with intestinal typhoid bacteria at the former mongolian border. during an infamous biowarfare attack in 1941, the japanese military released millions of plague-infected fleas from airplanes over villages in china and manchuria, resulting in several plague outbreaks in those villages. the japanese program had stockpiled 400 kg of anthrax to be used in specially designed fragmentation bombs. in 1942, shortly before the battle of stalingrad, on the german-soviet front, a large outbreak of tularemia occurred. several thousand soviets and germans contracted the illness. some estimate that more than 70% of the victims had inhalation tularemia, which is rare and considered to be evidence of an intentional release. it was determined later that the soviets had developed a tularemia weapon the prior year (alibek and handelman, 2000) . during world war ii the allies had great fear of german and japanese biological weapons programs. their fears were sparked by sketchy reports that the japanese had an ongoing effort, and british intelligence suggested that germany might soon target britain with a bomb packed with biological agents. on the basis of these fears, great britain began its own bioweapons program and urged officials in the united states to create a large-scale biological warfare program. on december 9, 1942, the us government convened a secret meeting at the national academy of sciences in washington, dc. the meeting was called to respond to great britain's request. army officers had urgent questions for an elite group of scientists. only a few months before, the president of the united states had grappled with the issue of biological weapons. president franklin d. roosevelt stated that "i have been loath to believe that any nation, even our present enemies, would be willing to loose upon mankind such terrible and inhumane weapons." secretary of war, general henry stimson, thought differently: "biological warfare is…dirty business," he wrote to roosevelt, "but…i think we must be prepared." president roosevelt approved the launch of the united states' biological warfare program. for the first time us researchers would be trying to make weapons from the deadliest germs known to science. in spring 1943 the united states initiated its bioweapons program at camp detrick (now fort detrick), maryland. the program focused primarily on the use of the agents that cause anthrax, botulism, plague, tularemia, q fever, venezuelan equine encephalitis, and brucellosis. production of these agents occurred at camp detrick, maryland, and other sites in arkansas, colorado, and indiana. the british had made two primary requests of us: (1) to mass produce anthrax spores so that they could be placed in bomblets and stored for later deployment against the germans in retaliation for any future strike and (2) the british supplied us with the recipe to make botulinum toxin and wanted to see if we could mass produce it. naturally the entire program was wrapped in a cloak of secrecy. fig. 1.2 is a collage of some important facilities built at camp detrick to produce and test bioweapons formulations. the british program focused on the use of b. anthracis (anthrax) spores and their viability and dissemination when delivered with a conventional bomb. gruinard island, off of the coast of scotland, was used as the testing site for formulations. at the time british scientists believed that the testing site was far enough from the coast to not cause any contamination of the mainland. however, in 1943 there was an outbreak of anthrax in sheep and cattle on the coast of scotland that faced gruinard. as a result, the british decided to stop the anthrax testing and close down the island site. despite the cessation of experiments, the island remained contaminated for decades until a deliberate and extensive decontamination program rendered the island inhabitable again. the us bioweapons program continued to grow in scope and sophistication. much of this was prompted by fear of a new enemy: the threat of communism, the soviet union, and its allies. experiments to test bioweapons formulations were routinely performed on a small scale with research animals. however, more comprehensive field and laboratory studies were performed with human research volunteers exposed to actual live agents and some situational scenarios using surrogate nonpathogenic bacteria to simulate the release of actual pathogens inside of buildings or aimed at cities. in 1949 researchers from detrick visited the pentagon on a secret mission. disguised as maintenance workers, they released noninfectious bacteria into the duct work of the building to assess the vulnerability of people inside large buildings to a bioweapons attack. the pentagon trial was considered to be a success because it revealed that germs could be formulated and released effectively for a small-scale act of sabotage. however, there was considerable doubt that biological weapons could be effective against a target the size of a city. accordingly, several tests were conducted on american cities (miller et al., 2001) . in 1977 the us army admitted that there were 239 intentional releases of noninfectious bacteria in bioweapons experiments (cole, 1988) . one such trial took place in san francisco in september 1950, when a us navy ship sailed a course adjacent to the golden gate bridge to release a plume of seemingly nonpathogenic bacteria (serratia marcescens). this trial was intended to simulate the dispersion of anthrax spores on a large city. on the basis of results from monitoring equipment at 43 locations around the city, the army determined that san francisco had received enough of a dose for nearly all of the city's 800,000 residents to inhale at least 5000 of the particles. although the researchers believed that what they were releasing was harmless, one report shows that 11 people reported to area hospitals with severe infections because of the release of this agent, 1 of which was fatal (cole, 1988) . three years later, bioweapons experts took their secret exercises to st. louis and minneapolis, two cities that resembled potential soviet targets, where sprayers hidden in cars dispersed invisible clouds of harmless bacillus spores. in 1966 nonpathogenic bacillus globigii spores were released into the new york subway system using a broken light bulb to demonstrate the ability of a specific formulation to make its way from a central point source to both ends of the system in less than an hour. revelations of these experiments became known in 1977 when a senate subcommittee panel heard testimony from pentagon officials (us department of the army, dtic b193427 l, 1977) . until that point, neither us citizens nor their representatives in washington knew anything about the american germ program. after nearly 3 decades of secret research aimed at producing the ultimate biological weapons and stockpiling them for use against our enemies, president richard nixon surprised the world by signing an executive order that stopped all offensive biological agent and toxin weapon research and ordered all stockpiles of biological agents and munitions from the us program be destroyed. accordingly, on november 25, 1969, he uttered these historic words in a speech to the nation on …biological warfare-which is commonly called "germ warfare. " this has massive unpredictable and potentially uncontrollable consequences. it may produce global epidemics and profoundly affect the health of future generations. therefore, i have decided that the united states of america will renounce the use of any form of deadly biological weapons that either kill or incapacitate. mankind already carries in its own hands too many of the seeds of its own destruction. subsequently, in 1972 the united states and many other countries were signatories to the convention on the prohibition of the development, production and stockpiling of bacteriological (biological) and toxin weapons and on their destruction, commonly called the biological weapons convention. this treaty prohibits the stockpiling of biological agents for offensive military purposes and forbids research into offensive use of biological agents. although the former soviet union was a signatory to the biological weapons convention, its development of biological weapons intensified dramatically after the accord and continued well into the 1990s. in late april 1979, an outbreak of pulmonary anthrax occurred in sverdlovsk (now yekaterinburg) in the former soviet union. soviet officials explained that the outbreak was due to ingestion of infected meat. however, it was later discovered that the cause was from an accidental release of anthrax in aerosol form from the soviet military compound 19, a soviet bioweapons facility. (this event is examined thoroughly in chapter case studies as a case study to demonstrate the potential of weaponized anthrax.) the robust bioweapons program of the soviet union employed more than 60,000 people. building 15 at koltsovo was capable of manufacturing tons of smallpox virus each year. in kirov, the soviets maintained an arsenal of 20 tons of weaponized plague bacterium. by 1987 soviet anthrax production capacity reached nearly 5000 tons a year. in the later part of the 1990s the russians disassembled their awesome bioweapons production capacity and reportedly destroyed their stocks. as the soviet union dissolved, it appeared that the threat of biowarfare would diminish. however, the age of bioterrorism emerged with the anthrax attacks of 2001. in addition, the us department of state published a report in 2004 that affirmed that six countries had active bioweapons programs. table 1 .2 summarizes some of these events. biodefense programs and initiatives come out of a sense of vulnerability to biowarfare potentials. bioterrorism is deeply founded in what has been gained from active biowarfare programs (miller et al., 2001) . in the early 1970s the leftist terrorist group, the weather underground, reportedly attempted to blackmail an army officer at fort detrick working in the research institute of infectious diseases (usamriid). the group's goal was to get him to supply organisms that would be used to contaminate municipal water supplies in the united states. the plot was discovered when the officer attempted to acquire several items that were "unrelated to his work." several other attempts are worth mentioning here: • in 1972 members of the right-wing group order of the rising sun were found in possession of 30-40 kg of typhoid bacteria cultures that were allegedly to be used to contaminate the water supplies of several midwestern cities. 1915 anton dilger produces anthrax and glanders bacterium to infect horses intended for the warfront. notable and documented use of bacteria against animals. june 17, 1925 delegates in switzerland create a geneva protocol banning the use of chemical and bacteriological methods of warfare. first international effort to limit use of biologicals in warfare. the japanese army gives general ishii control of three biological research centers, including one in manchuria. most despicable character in bioweapons history gets his start. 1934 great britain begins taking steps toward establishing its own biological weapons research project. allies start to develop a program. july 15 nixon announces that the united states will renounce the use of any form of deadly biological weapons that either kill or incapacitate. the end of an era in us offensive biological weapons research, production, and storage. april 10, 1972 the biological weapons convention, which bans all bioweapons, is completed and opened for signature. seventy-nine nations signed the treaty, including the soviet union. march 26, 1975 the biological weapons convention officially goes into force; the us senate also finally ratifies the 1925 geneva protocol. political will to ban biological weapons on the international front. april 1979 nearly 70 people die from an accidental release of anthrax spores in the soviet city of sverdlovsk. the united states suspects that anthrax bacterial spores were accidentally released from a soviet military biological facility. 1984 the rajneeshees contaminate food with salmonella bacterium in a small town in oregon to influence local elections. these small-scale incidents keep us mindful that some biological agents are easy to acquire and utilize in crimes and small-scale acts of terrorism. salmonella bacteria in a small town in oregon. it was the largest scale act of bioterrorism in us history. more than 750 cases of salmonellosis resulted from the salad bar contamination. it was later discovered that the rajneeshees wanted to influence the local county elections. cult members obtained the salmonella strain through the mail from american type culture collection and propagated the liquid cultures in their compound's medical clinic. • in 1989 a home laboratory producing botulinum toxin was discovered in paris. this laboratory was linked to a cell of the german-based bäder meinhof gang. • in minnesota, four members of the patriots council, an antigovernment extremist group, were arrested in 1991 for plotting to kill a us marshal with ricin. the group planned to mix the homemade ricin with a chemical that speeds absorption (dimethylsulfoxide) and then smear it on the door handles of the marshal's car. the plan was discovered and all four men were arrested and the first to be prosecuted under the us biological weapons anti-terrorism act of 1989. • in 1995 aum shinrikyo, a japanese doomsday cult, became infamous for an act of chemical terrorism when members released sarin gas into the tokyo subway. what many people do not know about the group is that it developed and attempted to use biological agents (anthrax, q fever, ebola virus, and botulinum toxin) on at least 10 other occasions. despite several releases, it was unsuccessful in its use of biologicals. this program is examined more thoroughly in chapter case studies. • several small-scale incidents involving the biological poison ricin (refer to fig. 1.3 mississippi man was convicted of crimes related to these incidents and sentenced to 25 years in prison. • in 2014 a philadelphia man sent a romantic rival a scratch-and-sniff birthday card laced with ricin. in 2015 he was convicted on several charges related to the incident and subsequently received a sentence of 20-40 years in prison. on june 12, 2002, president george w. bush uttered these remarks from the white house at the signing of hr 3448, the public health security and bioterrorism response act of 2002: bioterrorism is a real threat to our country. it' s a threat to every nation that loves freedom. terrorist groups seek biological weapons; we know some rogue states already have them…it' s important that we confront these real threats to our country and prepare for future emergencies. it is clear that september 11 and the anthrax attacks of 2001 sent the country to war and sparked several initiatives against all forms of terrorism. biological agents have some unique characteristics that make weaponizing them attractive to the would-be terrorist. most biological weapons are made up of living microorganisms, which means that they can replicate once disseminated. this possibility amplifies the problem and the effect of the weapon in several ways. first, some agents are capable of surviving in various different hosts. the target might be humans, but the disease may manifest in other animal hosts, such as companion animals (pets). in doing so, the problem may be more difficult to control. second, when people become infected with a disease-causing organism, there is an incubation period before signs of illness are apparent. during this incubation period and the periods of illness and recovery, the pathogen may be shed from the victim, causing the contagion to spread (a possibility only with diseases that are transmitted from person to person). there is no rule of thumb for how many people might be infected from a single patient. however, the nature of contagion clearly compounds the problem well beyond the initial release of the agent. in this instance the initial victims from the intentional outbreak become more weapons for the perpetrator, spreading the problem with every step they take. as grigg et al. (2006) stated so precisely in their paper, "when the threat comes from the infected population, selfdefense becomes self-mutilation." the would-be terrorist could surely derive great pleasure from watching government officials and responders tread on the civil liberties of such victims as they attempt to limit the problem from spreading among the population. making an effective biological weapon is no easy undertaking. the process and complexity depends largely on the pathogen selected to be "weaponized." if the pathogen is a spore-forming bacteria, such as b. anthracis (the causative agent of anthrax), there are five essential steps: germination, vegetation, sporulation, separation, and weaponization. the first three steps are designed to get small quantities of seed stock to propagate into a starter culture, grow them to a significant stage of growth in the proper volume, and turn those active cells into spores. the goal of the last two steps is to separate the spores from the dead vegetative cells and spent media. all five steps have dozens of secondary steps. in addition, each of the five steps requires a fairly sophisticated and well-equipped laboratory if the goal is to develop a sizable quantity of refined materials. weaponization is a term that applies to the processes necessary to purify, properly size, stabilize, and make biological agents ideally suited for dissemination. stabilization and dissemination are important issues because of the susceptibility of the biological agents to environmental degradation, not only in storage but also in application. these issues are problems whether the end use is for biological weapons, pharmaceuticals, cosmetics, pesticides, or food-related purposes. the susceptibility of the organisms to inactivation by the environment varies with the agent. as an example, anthrax spores released into the environment may remain viable for decades, whereas plague bacterium may survive for only a few hours. loss of viability or bioactivity is likely to result from exposure to physical and chemical stressors, such as exposure to ultraviolet radiation (sunlight), high surface area at air-water interfaces (frothing), extreme temperature or pressure, high salt concentration, dilution, or exposure to specific inactivating agents. this requirement of stabilization also extends to the methods of delivery because the organisms are very susceptible to degradation in the environments associated with delivery systems. the primary means of stabilization for storage or packaging are concentration; freeze drying (lyophilization); spray drying; formulation into a stabilizing solid, liquid, or gaseous solution; and deep freezing. methods of concentration include vacuum filtration, ultrafiltration, precipitation, and centrifugation. freeze drying is the preferred method for long-term storage of bacterial cultures because freeze-dried cultures can be easily dehydrated and cultured via conventional means. freeze-dried cultures may remain viable for more than 30 years. deep freezing of biological products is another long-term storage technique for species and materials not amenable to freeze drying. the method involves storage of the contained products in liquid nitrogen freezers (−196°c/−325°f) or ultralow-temperature mechanical freezers (−70°c/−94°f). culturing viruses is a more costly and tenuous process because host cells are required for viral propagation. this means that cultures of host cells must be kept alive, often in an oxygen-deficient and temperature-stable atmosphere. in some cases, viruses may be more fragile when deployed as weapons, some becoming inactive on drying. biological toxins can be difficult to produce and purify, each requiring its own special set of circumstances. two specific examples are covered in subsequent chapters when those agents are discussed in detail. however, past bioweapons programs have determined that these agents are most effective when prepared as a freeze-dried powder and encapsulated. biological attacks by a terrorist group are apparently not easy to conduct or a practical option. if they were easy or practical, then many terrorist groups and hostile states would have done so long ago and frequently. our experience today with acts of biological terrorism has to do mainly with small-scale, limited attacks. however, if one were to acquire the means to produce the weapons, as described here, or purchase viable, sophisticated materials on the black market, a small group of persons could bring about the infection of a large percentage of targeted persons. clinical illness could develop within a day of dispersal and last for as long as 2-3 weeks. in a civil situation, major subway systems in a densely populated urban area could be targeted for a biological agent strike, resulting in massive political and social disorganization. it would take little weaponized material to bring about the desired effect. looking at this potential comparatively on a weight-toweight basis, approximately 10 g of b. anthracis (anthrax) spores could kill as many people as a ton of the nerve agent sarin. with bioweapons in hand, small countries or terrorist groups might develop the capability to deliver small quantities of agents to a specific target. under appropriate weather conditions and with an aerosol generator delivering 1-to 10-μm particle-sized droplets, a single aircraft could disperse 100 kg (220 lb) of anthrax over a 300-km 2 area (74,000 acres) and theoretically cause 3 million deaths in a population density of 10,000 people/km 2 (us dod, ada 330102, 1998). much has been made of the potential of aerosolized powders and respiratory droplets in factual and fictitious biothreat scenarios. the largest infectious disease outbreak in the history of the united states occurred in april 1993. the event was caused by an accidental waterborne contamination. the outbreak of cryptosporidiosis, which occurred in the greater milwaukee area, was estimated to have caused more than 430,000 people to become ill with gastroenteritis among a population of 1.6 million (mackenzie et al., 1994) . approximately 4400 people were hospitalized and about 100 people died as a result of the outbreak. the milwaukee outbreak was attributed to failure of filtration processes at one of the two water treatment plants that served the city. several deficiencies were found at the plant, including problems relating to a change in the type of water treatment chemicals used for coagulation of contaminants before the filtration step. weather conditions at the time were unusual, with a heavy spring snowmelt leading to high source water turbidity and wind patterns that may have changed normal flow patterns in lake michigan, the raw water source for the city. describe the fundamental difference between biodefense and biosecurity. the secrecy of bioweapons programs of the previous century has been uncloaked. some of the most insidious disease agents ever to afflict humans, animals, and plants have been mass produced and perfected for maximum effectiveness. terrorist groups and rogue states may be seeking to develop bioweapons capabilities. these significant developments in bioweapons gave military leaders and politicians cause for great concern over the past few decades. the military necessity to protect the force and defend the homeland is the goal of a good biodefense program. simply put, biodefense is the need for improved national defenses against biological attacks. these are national programs, mostly planned and carried out by military forces and other government agencies. initially, biodefense programs require an intelligence-gathering capability that strives to determine what may be in the biological weapons arsenal of an aggressor. intelligence is needed to guide biodefense research and development efforts aimed at producing and testing effective countermeasures (ie, vaccines, therapeutic drugs, and detection methods). in addition, a real-time reporting system should be developed so that officials can be informed about an emerging threat before an agent has a chance to affect armed forces and millions of people in the homeland. the development of integrated systems for detecting and monitoring biological agents is instrumental to this goal. although most biodefense initiatives rest with the military, civilian government agencies contribute greatly to the biodefense posture. this is evident by the increases in cbf over the past few years and will be discussed in great detail in part iv of this book. on the other hand, biosecurity refers to the policies and measures taken for protecting a nation's food supply and agricultural resources from accidental contamination and deliberate attacks of bioterrorism. as i sit here today writing the second edition of this book, i am reflecting on the most recent concerns that we have for biological threats in modern society. for what it is worth, we seem to be much less concerned about acts of bioterrorism and/or biowarfare than we were 10-50 years ago. instead i see a great deal of concern, and rightfully so, for emerging infectious diseases and reemerging biological threats. we are also keenly aware of the accidental release of biological agents from research and reference laboratories. to illustrate these points we will briefly discuss four items of international interest that have been emphasized in the media: accidental shipment of live anthrax-positive controls samples, the 2014/2015 ebola outbreak in west africa, cases of middle east respiratory syndrome coronavirus (mers-cov) in south korea and saudi arabia, and a massive outbreak of highly pathogenic avian influenza (hpai). as previously mentioned, concerns for biological threat led to a wellspring of funding (nearly $80 billion in 15 years) for civilian biodefense programs in the united states. with all of this money the united states was able to build tremendous capabilities to detect and diagnose the agents and the diseases, respectively. with this money a few medical countermeasures (vaccines and treatments) were developed and produced. centers of excellence were funded and highly secure containment (biosafety level 4) laboratories were built. with these new programs, testing modalities, and laboratories came the need to provide a ready supply of positive control agents and contracting opportunities for private biotechnology firms. as one very recent example, the us army laboratory in dugway proving grounds, utah, provided positive control samples of anthrax (b. anthracis) spores to public and private laboratories. before shipment, the spores had been propagated in the army laboratory and were exposed to gamma radiation to ensure no living spores were in the vials being provided. upon receipt of the samples, one laboratory in maryland questioned the integrity of the contents of the vial they received because there was no "death certificate" accompanying the samples. out of an abundance of caution they removed a small portion of the vial and streaked it onto sheep blood agar plates. to their amazement, several days later the plates showed growth and tested positive for anthrax. they immediately notified the cdc and the army. the cdc initiated an investigation and notified the media of the incident. the investigation showed that the living anthrax samples had been shipped to 69 laboratories in 19 us states and 5 other countries (usa today, 2015) . once again the seeds of our destruction are sprouting, and some are of the opinion that we are our own worst enemy. more than 1100 laboratory incidents involving potential bioterror germs were reported to federal regulators during 2008 through 2012. ebola virus was first discovered in 1976 in the sudan and zaire. ebola virus exists naturally in fruit bats, with sylvatic transmission to other mammals and sometimes humans when they consume raw or undercooked meat from an infected animal. infection with ebola virus in humans leads to severe viral hemorrhagic fever (vhf), which is often fatal (cdc, 2015) . in march 2014 an outbreak of ebola virus disease (evd) began in guinea, a western african nation. public health agencies at all levels failed to react quickly to the outbreak and it quickly spread to urban areas in liberia and sierra leone. subsequently, evd spread to nigeria and senegal. international air travel brought evd to the united states and europe, although the number of cases was very small and the threat was stamped out with ample infection control procedures in health-care facilities and aggressive public health measures for those exposed to actual case patients (cdc, 2015) . this is the largest outbreak of evd in history. at the time of this writing, the outbreak has been quelled by a "better late than never" effort. volunteers and medical relief groups from the united states and other countries received special training and deployed to west africa to help identify cases and treat the victims (see fig. 1.4) . however, new cases continue to be reported from guinea and sierra leone. as of june 26, 2015, there have been 16,801 evd cases (suspect, probable, and confirmed) worldwide with approximately 6411 deaths; this equates to a 38% mortality rate (who, 2015a) . to most the threat of ebola virus remains distant and out of mind. however, the stark reality is that international travel can interject evd into any populace on any continent within a matter of days. no country, person, or organization is immune to this threat. what makes evd such a great concern? first, ebola virus is a us health and human services category a agent. it meets all of the criteria for such a designation. evd results in high morbidity and mortality. evd requires special preparedness measures for public health and health care. evd is spread from person to person. evd can lead to panic and social disruption (cdc, 2015) . with this outbreak in particular, we are seeing all four criteria fulfilled. to make things worse, there is no food and drug administration (fda)-approved vaccine for humans and no fda-approved drug for treating vhf case patients. in a health-care setting, evd patients receive supportive care (hydration therapy) and rarely experimental drugs (cdc, 2015) . perhaps the only good thing to come from this outbreak is the development of a vaccine for ebola virus. there are currently three vaccine candidates undergoing phase iii clinical trials in west africa (who, 2015b) . a case study on this outbreak is offered in chapter case studies of this book. how have international and national attitudes toward the biological threat changed since the early post-9/11 era? include some discussion about the reality of versus the potential for biological threats. mers-cov (see fig. 1 .5) was recognized by the world health organization (who) as a newly emerging pathogen in 2012 (berry et al., 2015) . the initial case where virus isolation and characterization came from occurred in jeddah, saudi arabia. subsequent infections were reported in middle eastern countries (jordan, qatar, and the united arab emirates), with a few cases also identified in europe, north africa, and the united states. mers-cov leads to severe respiratory illness in susceptible patients and is spread through person-to-person contact. (2014). the program had been designed to educate participants who would be deployed as members of the west african ebola response team as to the proper protocols to be followed when treating evd patients. the two participants were displaying the personal protective equipment worn by treatment specialists who would be interacting with evd patients. courtesy of the cdc/nahid bhadelia, md. south korea has recently been the epicenter of the largest outbreak of mers-cov outside of the middle east, reporting 180 cases and 29 deaths (who, 2015c). the outbreak in south korea was traced to a single infected traveler. once again, this demonstrates the vulnerability to unexpected outbreaks of unusual diseases that all countries share in this highly mobile world. a report from a joint who-south korean investigation of this outbreak identified several reasons for the severity of the outbreak in south korea. these include a lack of awareness among health-care workers and the general public about mers-cov, the practice of "doctor shopping" (seeking care at multiple hospitals), people visiting infected patients in multibed hospital rooms, substandard infection control and prevention measures in health-care facilities, and contact of infected mers-cov patients in crowded emergency rooms. nearly all of the country's confirmed mers-cov patients were infected while seeking care or visiting hospital patients (boston globe, 2015) . more about mers-cov and other emerging pathogens is in chapter category c diseases and agents. hpai has been very much in the news since 1997 when the novel strain h5n1 jumped from domestic bird populations (poultry) to humans in south east asia (ryan, 2008) . h5n1 was very much feared by public health and government officials for its pandemic potential. since 2003 there have only been approximately 650 cases of h5n1 infection in humans, with a mortality rate of approximately 60% (hhs, 2015) . since that time, numerous other novel strains have emerged. in fact, a novel h1n1 arose out of swine in 2009 and was the cause of a mild pandemic in humans. more recently, the novel strains h5n2 and h5n8 have been found to be the cause of major morbidity and mortality in poultry operations (chicken and turkey) in the united states, with 223 detections affecting more than 48 million birds (usda, 2015) . refer to fig. 1.6 for a graphic representation of this outbreak. the financial impact on the poultry growers and the egg and meat industry has been enormous. more can be found on this topic in chapter recent animal disease outbreaks and lessons learned. from this first chapter we can now understand and appreciate the scope and importance of biological threats and see where they may be and have become the desire of terrorist groups and the makings of weapons of mass destruction. biowarfare has a history. the major events are important in helping us understand the issues related to using biological substances against an adversary. we now know the difference between biosecurity and biodefense and can relate them to homeland security and homeland defense, respectively. we also know how expensive these programs are because nearly $80 billion has been spent on civilian biodefense since fy2001 in the united states alone. as discussed herein, there is a significant difference in the reality and the potential of bioterrorism. bioterrorism on a large scale is a low-probability event. bioterrorism and biocrimes on a small scale (eg, small amount of ricin directed at one or a few individuals) are fairly routine occurrences with little potential. biological threats remain very much in the news. recent examples, such as laboratory incidents, the ebola outbreak of 2014/2015, the outbreak of mers-cov in south korea, and the hpai outbreak affecting poultry in the united states, make us aware that we must remain vigilant and utilize the biosecurity and biodefense programs to help us identify and respond to these accidental exposures and emerging threats. • biodefense. the collective efforts of a nation aimed at improving defenses against biological attacks. within these efforts are programs and agencies working toward increasing data collection, analysis, and intelligence gathering. the intelligence is applied to programs aimed at mitigating the effects of bioweapons by developing vaccines, therapeutics, and detection methods to increase the defensive posture. ultimately, biodefense initiatives protect the military forces and the citizens from the effects of biological attack. • biosecurity. the policies and measures taken for protecting a nation's food supply and agricultural resources from accidental contamination and deliberate attacks of bioterrorism. • bioterrorism. the intentional use of microorganisms or toxins derived from living organisms to cause death or disease in humans or the animals and plants on which we depend. bioterrorism might include such deliberate acts as introducing pests intended to kill us food crops; spreading a virulent disease among animal production facilities; and poisoning water, food, and blood supplies. • biowarfare, also known as germ warfare. the use of any organism (bacteria, virus, or other disease-causing organism) or toxin found in nature as a weapon of war. it is meant to incapacitate or kill an adversary. • pathogen. a specific causative agent of disease, mostly thought of as being an infectious organism (eg, bacteria, virus, rickettsia, protozoa). • weaponization. when applied to biologicals, the term implies a process of taking something natural and making it harmful through enhancing the negative characteristics of it. with biological agents, one might weaponize the agent by making more lethal, more stable, and more easily delivered or disseminated against an intended target. there is considerable debate about the use of this term. • zoonotic disease. an animal disease that may be transmitted to humans. • how was the decision made to begin the us biological weapons program? • what are the significant events in the history of biowarfare? what makes them significant? • when president nixon said that "mankind already holds in its hands too many of the seeds of its own destruction" in november 1969, what did he mean by that? • weaponizing a biological agent is easy to do, right? • no one knows exactly who perpetrated the anthrax attacks of 2001, and there has been no repeat of them since. why do you think we have seen no repeat of the anthrax attacks since 2001? the center for arms control and nonproliferation has an online course in biosecurity. type the url that follows into your internet browser and click on view course and select unit 2: "the history of biological weapons." the six sections in this unit provide an excellent overview and reinforce the material presented in the subheading about the history of biowarfare: www.armscontrolcenter.org/resources/biosecurity_course. the cdc's emergency and preparedness website offers a segmented video short lesson on the history of bioterrorism. the seven sections give a general overview on bioterrorism and separate vignettes on anthrax, plague, tularemia, vhfs, smallpox, and botulism: www.bt.cdc.gov/training/historyofbt. biohazard: the chilling true story of the largest covert biological weapons program in the world-told from the inside by the man who ran it. random house forensic application of microbiological culture analysis to identify mail intentionally contaminated with bacillus anthracis spores identification of new respiratory viruses in the new millennium. viruses south korea faulted on mers response biological warfare: a historical perspective clouds of secrecy. the army's germ warfare tests over populated areas the malaria capers: more tales of parasites and people, research, and reality guns, germs and steel: the fates of human societies medical aspects of chemical and biological warfare: a textbook in military medicine inside the spore wars. controversial contracts, bureaucratic bungling-the fed's biodefense drug program is a mess. how did it go so wrong? time the biological disaster challenge: why we are least prepared for the most devastating threat and what we need to do about it scientists and the history of biological weapons. a brief historical overview of the development of biological weapons in the twentieth century billions for biodefense: federal agency biodefense funding, fy2006-fy2007 a massive outbreak in milwaukee of cryptosporidium infection transmitted through the public water supply germs: biological weapons and america's secret war the conspiracy of pontiac and the indian war after the conquest of canada a history of influenza rotting face: smallpox and the american indian pandemic influenza: emergency planning and community preparedness postal testing increasing five years after anthrax deaths billions for biodefense: federal agency biodefense budgeting, fy2005-fy2006 federal agency biodefense funding, fy2013-fy2014 the indian wars of pennsylvania killer strain: anthrax and a government exposed microbiology. an introduction ebola virus disease update on avian influenza findings the militarily critical technologies list part ii: weapons of mass destruction technologies h5n1 avian flu (h5n1 bird flu). available at hundreds of bioterror lab mishaps cloaked in secrecy army lab lacked effective anthrax-killing procedures for 10 years biological and toxin weapons: research, development and use from the middle ages to 1945 world health organization ebola vaccines, therapies and diagnostics middle east respiratory syndrome coronavirus (mers-cov) -republic of korea key: cord-016361-upjhmfca authors: tshilenge mfumu, jean-claude; mercier, annabelle; occello, michel; verdier, christine title: a multiagent-based model for epidemic disease monitoring in dr congo date: 2019-07-16 journal: biomedical engineering systems and technologies doi: 10.1007/978-3-030-29196-9_17 sha: doc_id: 16361 cord_uid: upjhmfca any infectious diseases have been reported in sub-saharan countries over the past decade due to the inefficiency of health structures to anticipate outbreaks. in a poorly-infrastructure country such as the democratic republic of congo (drc), with inadequate health staff and laboratories, it is difficult to respond rapidly to an epidemic, especially in rural areas. as the drc’s health system has three levels (peripheral, regional and national), from the production of health data at the peripheral level to the national level that makes the decision, meantime the disease can spread to many people. lack of communication between health centres of the same health zone and health zones of the same health provincial division does not contribute to the regional response. this article, an extended version of [1], proposes a well elaborated solution track to deal with this problem by using an agent-centric approach to study by simulation how to improve the process. a new experiment is described by arranging twenty-eight health zones of kinshasa to show how their collaboration can provide unique health data source for all stakeholders and help reducing disease propagation. it concerns also 47 health centres, 1 medical laboratory, 1 provincial health division and 4 rapid riposte teams. the simulation data, provided by provincial health division of kinshasa, concerned cholera outbreak from january to december 2017. the interaction between these agents demonstrated that health zone agent can automatically alert his neighbours whenever he encountered a confirmed case of an outbreak. this action can reduce disease propagation as population will be provided with prevention measures. these interactions between agents have provided models to propose to the current system in order to find out the best that can help reducing decision time. 1 introduction experiencing serious difficulties in improving the living conditions of its population, particularly in the field of basic health care. life expectancy at birth is 50 and 53 respectively for men and women [3] . the country is currently divided into city-province of kinshasa and 25 other provinces. the provinces are subdivided into territories which are divided into sectors. to facilitate the supervision of health structures, drc health system is divided into three levels [4] : central, intermediate and peripheral. the nearest level to population is the peripheral area composed of 518 health zones (hz) that coordinate the actions of health facilities. a health zone is divided into health area (ha) containing one or more health centers (hc) and a general referral hospital (grh). the central (national) level defines the policies, strategies and resources of the sector. it enforces strategies and policies at the peripheral level through the intermediate level called the provincial health division (phd), which coordinates primary health care and technical support activities for health zones in a province. provincial health minister (phm) is the political authority of the province. each health zone has a health information bureau (hib) which retrieves aggregated data from all its supervised health area to send to national level for decision measures. hib has a health zone executive team (hzet) that organizes weekly meetings to discuss about suspicious cases to report to hierarchy if needed. health zone executive team manages health facilities (hf) that includes health center and grh. as who member country, drc benefits from the technical and financial support of the partners to respond to epidemics under the conditions stipulated in the international health regulations (ihr) [5] . all cases of these four diseases must be automatically notified to who: smallpox, poliomyelitis due to wild-type poliovirus, severe acute respiratory syndrome (sars) and cases of human influenza caused by a new subtype. on the ever-changing list of diseases provided by ihr, each country is free to add other diseases, with epidemic potential or not, which constitute a public health problem. access to basic care is difficult for a large part of population. people visit the health facilities in case of extreme emergency. this is more evident in rural areas where the diminishing resources of farmers do not allow them to consult medical services regularly. most of time health care is provided during free medical workers' campaigns. the most usual ways to collect data about cases that must be sent to the hierarchy for decision-making are described below [6] : -pharmacies must report when same medicines are increasingly purchased by population or when they detect some recurrent treatments; -schools reporting unusual rates of pupil absences due to strange signs and symptoms; figure 2 presents an exhaustive list of actors used in providing health data. despite great efforts to improve disease surveillance and response, drc faces big challenges in identifying, diagnosing and reporting infectious diseases properly due to the remoteness of communities, the inadequate transport and communication infrastructures, and the lack of qualified health staff and laboratory facilities to ensure accurate diagnosis. the challenge, in this paper, is to find new solutions based on real population life and situation to improve health services organization and data sharing in order to detect infectious disease very quickly, organize the response and prevent the spread of disease. in this paper, we present a part of this challenge. we propose a multi-agent system to simulate the interactions between actors working together to organize an optimal response to epidemic detection. when a new case of infectious disease is suspicious in health center, actors will collaborate to report it to provincial health division through health zone executive team. the approach will be based on the current drc heath system processes to extract relevant actors' tasks. the identification of these actors and their tasks will provide the opportunity to simulate a new system that distributes the entire competences of the old heath system to those actors in order to improve their collaboration and eventually shorten the making-decision time response. work-sharing protocols will be proposed to simplify the complexity of the data sources. this paper focuses precisely on improving the process of reporting health data from the peripheral level to the hierarchy for rapid decision-making and anticipate as much as possible the medical response using multi-agent systems (mas). hierarchical dependency between three levels forbids peripheral to directly transmit health data to national level for quick decision. as information must pass through intermediate level (provincial health division), with defective means of communications, it drastically hampers the fight against a propagation of a disease. since decision-making is pushed back to the central level, it can intervene belatedly at the risk of witnessing an alarming spread of an epidemic with high epidemic potential. the next section shows the related work in healthcare and multiagent domains. section 3 describes the healthcare system and problematic in drc. the methodological approach and agent's models are explained in sect. 4. the model is validated by a simulation presented in sect. 5. future research directions and conclusion are developed in sect. 6. information and communication technology is a powerful solution for health care in developing countries [7] . it made possible the improvement of remote patient follow-up [8] , controlling the progression of malaria [9] , improving the uptake of information from health systems [10] . two main ways of research can be studied in this paper: the use of mobile phone as a relevant medium to rapidly transfer medical data and the multi-agent system that is powerful to simulate organizational skills to anticipate diseases spreading. mobile phone coverage in africa grew from 10% in 1999, 65% in 2008 to more than 70% in 2012 [11] . this technology is used to cover numeric fraction. to improve drug adherence and suppression of plasma hiv-1 rna load in kenyan, mobile phone communication between health-care workers and patients starting antiretroviral therapy was set up [12] text-message reminders sent to health workers' mobile phones improved and maintained their adherence to treatment guidelines for outpatient pediatric malaria [13] . phone traces are powerful tools to estimate population migration while investigating an outbreak. these techniques were used to demonstrate the feasibility of rapid estimates and to identify areas at potentially increased risk of outbreaks in haiti. they produced reports on sim card movements from a cholera outbreak area at its immediate onset and within 12 h of receiving data. results suggest that estimates of population movements during disasters and outbreaks can be delivered rapidly and with potentially high validity in areas with high mobile phone. a trial of mobile phone text messaging for diabetes management in an eight-month period to transmit data such as blood glucose levels and body weight to a server that automatically answered with a monthly calculated glycosylated hemoglobin result. the trial results suggest that sms may provide a simple, fast and efficient adjunct to the management of diabetes. in developed countries sms messages have been widely used to remind patients of scheduled appointments car and all, 2008). similarly, more complex mobile phone applications have shown significant improvement in the follow-up of malaria patients in thailand. the same approaches have been tested in africa as part of the sms reminder package to improve patients' adherence to antimalarial treatment schedules in .six sub saharan countries [9] . even if text-messaging is the simplest and the most widely easy to use technology function in reporting periodic data from the health system peripheral to control managers, it however needs to be experienced in interventions targeting individual patients, for whom a high facility workload or illiteracy may present a barrier. a multi-agent system is a set of agents situated in a common environment, which interact and try to reach a set of goals. through these interactions, a global behavior, more intelligent than the sum of the local intelligence of multiagent system components, can emerge. by 'agent' we mean a software entity evolving in an environment that it can perceive and within which it reacts. there are several kinds of agents. a reactive agent reacts to the stimuli of the environment. cognitive agents have a more developed intelligence: they manage knowledge and make decisions according to their internal states and according to their perception of the environment. it is provided with autonomous behaviors and some objectives. autonomy is the main concept in the agent issue: it is the ability of agents to control their actions and their internal states. the agents' autonomy implies no centralized control [14] . when a mas integrates agents on the fly dynamically during the execution, the system is qualified as an opened one. otherwise, it is a closed one. one of the advantages of mas is to model systems where a global description is not possible at any given moment. multi-agent conception is well suitable to model actors described in fig. 1 . simulation based on mas approach has wide potential applications in healthcare. first case, mas approaches are suitable to applications where a complete control is unreachable, the high number of entities or the complexity of entities' behavior make hard to represent the overall system. the second case is related to the monitoring of the epidemics. in a general way, the model of each agents going into action in the system is designed in microscopic level. the environment, the agents' interactions and the social organizations are defined at the macroscopic level. in mas domain, there are numerous methods to approach the analysis and the design of the software application. the methods come from various domains such as object oriented design, knowledge engineering or reproduction of behavior or natural phenomena. the first methods were aaii [15] which uses an external view (roles services, organization) and internal view (agent design bases on belief desire intention architecture). cassiopée [15] is a bottom-up approach based on natural behavior and desire [16] is based on knowledge engineering. the method vowels [17] allows obtaining modularity at the level of the multi-agents models by decomposing the problem into four elementary facets. other method like gaia [18] extends the concepts used in classical object engineering and provides a microscopic and a macroscopic view of the software system. there are also complete approaches for developing multi-agents systems from the analysis to the deployment by using mase [19] or prometheus [20] . moreover some models al-low the designer to develop the agent like agr [19] based on agent/group/role or bdi [21] based on belief/desir/intention. the main concept of all those approaches is the agent and the communication between agents to lead to the main objective of the software. the general classification is clinical, operational, managerial and educational simulation. managerial and operational simulations are closely interrelated. both are the core components for healthcare process management. some challenges and trends of simulation models in healthcare in the past two decades have been developed the design of a web-based clinical decision support system that guides patients with low back pain in making suitable choices on self-referral has been experienced in netherlands. mas is used to describe an approach to the analysis and development of telemedicine systems [22] , to manage communications in wireless sensor networks [23] , the epidemiological decision support system [24] , the care of seniors at home [25] , decision-making for monitoring and prevention of epidemics [26] , evaluation of disaster response system [27] , medical sensor modules in conjunction with wireless communication technology supporting a wide range of services including mobile telemedicine, patient monitoring, emergency management and information sharing between patients and doctors or among the healthcare workers [28] . mas can be considered as a suitable technology for the realization of applications providing healthcare and social services where the use of data and remote collaborations among users are often the most requirements [29] . cooperation in agent technology can provide better healthcare than the traditional medical system [30] . real programs built on the multiagent paradigm are still evolving towards a complete maturity. the variety and complexity of the e-health scenario make it one of the most interesting application fields, able to check the advantages of their use to condition their evolution [31] . mas was used to monitor a generic medical contact center for chronic disease environment, detect important cases, and inform the healthcare and administrative personnel via alert messages, recommendations, and reports, prompting them to action [32] . developed mas applications in healthcare can provide a reasonable way to mitigate the cost due to increased demand for services [33] . an agent-based model (abm) with geospatial and medical details was used to evaluate the efficiency of disaster responders to rescue victims in a mass casualty incident situation in south korea [27] . abm can cooperate to share tasks between sensors observing a phenomena [34] , to manage diabetes treatment between caregivers and patients. the usability evaluation of a collaborative information system for dementia assessment built using a usercentered design approach was experienced in norway [35] . but from several research papers we have reviewed we didn't find a paper addressing abm in sharing tasks from multisource health information to organize a rapid response to a high epidemic potential disease. the patient health care and the reporting of suspicious cases are managed at the peripheral level by health facilities (hf): health centers plus the general referral hospitals. health data collected by the health facilities are transmitted to health zone executive team for consolidation, analyzes and decision-making. aggregated data from entire health zone are also transmitted to provincial health division. each week this intermediate level structure convenes meetings to analyze data from each health zone, decide on actions to take. provincial health division produces consolidated data for its province. provincial health division must transmit the health data from its province to the central level for a second analysis and national consolidation. if suspicious cases reported by health zone require deeper investigation, laboratory tests or kits intervention, provincial health division will ask for technical and financial supports from central level. disease control direction (dcd) is a central respondent at central level. it also organizes weekly meetings to analyze health data from all provinces. it often provides advice and recommendations to provincial health division to monitor suspicious cases in accordance with the national policy of the sector. whenever provincial health division asks for a help, dcd can approach government authorities, special programs, partners and even the international community, to fill up needs. difficulties encountered by health facilities to better report information and structures dependency are well expressed at next section. the first challenge in managing epidemics begins with the multi-sources data processing at the health zone level. national policy has provided list of groups of individuals who can retrieve information from suspicious cases. this information transmitting by phone calls or narrative is not exhaustive. hence, the interest in diversifying the mode of communication by adding text and voice sms, tweets and phone calls on green lines can enhance data completeness. a second difficulty in an accurate identification of suspicious cases is the insufficient number of qualified health staff [36] . despite training courses organized by health zone executive team for community relays and staff of health facilities, there are gaps in the implementation of the information brought to their attention. for example, the provincial health division can conduct a thorough investigation with qualified staff as soon as the number of suspicious cases reaches the threshold for the pathology. lack of information on the list of the nearest laboratories delays response time to confirm cases and ensure accuracy of diagnosis. hierarchical dependences do not favor communication between structures of the same level such as health areas of a same health zone or of contiguous health areas but belonging to different provincial health division. this lack of dialogue can lead to the non-detection of an epidemic for the simple reason that the number of suspicious cases to organize investigation is not reached in an health zone. however, by combining number of suspicious case found in contiguous health areas, we could detect the pathology at the intersection of the provinces. structures authorized to report information relating to suspicious cases to health zone executive team are health facilities. but, community relays can also report their observations that need to be considered by health facilities. reports concern pathologies described at international sanitary regulations (smallpox, poliomyelitis due to wild polio virus, human influenza and severe acute respiratory syndrome (sars)) and local list of diseases with epidemic eradication measures or elimination and other chronic diseases provided by authorities. as soon as it appears, suspicious cases must be transmitted to health zone executive team by all data providers indicated on fig. 2 . when number of suspicious cases in health zone equals to the threshold of observed pathology, a rapid riposte team (rrt) has to investigate some health center and the population of the concerned health area to make sure the allegation was correct. the investigation of rapid ripost team could result to laboratory tests of some samples. in case of riposte many hierarchical structures such as provincial health division or national level would intervene to provide technical and financial supports. the process used to organize riposte simulation is heavily based upon computer science, mathematics, probability theory and statistics: yet the process of simulation modeling and experimentation remains very much an intuitive art. simulation is a very general and somewhat ill-defined subject. for the purpose of this paper, we will define simulation as «the process of designing a computerized model of a system and conducting experiments for the purpose of understanding the behavior of the system and/or of evaluating various strategies for the operation of the system. thus we will understand the process of simulation to include both the construction of the model and the analytical use of the model for studying a problem shortly described in fig. 3 . health zone executive team analyze the report of surveillance to determine if the number of suspicious case has reached the threshold to order an investigation. rapid ripost team will research new cases at health area according to the clinical definition of case. it will find out new determinants of the outbreak to report to provincial health division in other to realize the response. a final evaluation of outbreak response presented as a report of the process can be shared with other health zone and health facilities. this type of system is well suited to mas using an aeio representation. real system is analyzed with four elements: agent, environment, organizations and interactions between agents. this model will be detailed in the next part. the hierarchical organization of the healthcare system in drc is a good candidate for a multiagent model because there are several kinds of agents with personal goals sharing the same global achievement. in the process described in fig. 3 , the agents use some knowledge and tasks to perform a main goal together: collecting data in order to respond with efficiency to epidemic. at the stage of this research, the simulation's objective is to understand how the drc's healthcare system reacts in epidemic diseases. it is a preliminary work before (i) determining metrics to analyze process simulations and (ii) developing modules in embedded systems -phones or tablets-to assist the end-user in the data collection, coupled with the multiagent system. multiagent-based-simulation (mabs) allows explicitly modeling the behavior of each individual and viewing the emergent system from the interactions between the individuals. morvan in [37] proposes a survey on mabs and presents several multiagents platforms. in these existing platforms, we have not found solutions which can act as both a simulation system and a tool to end-users on embedded systems. diamond method and its associated simulator mash -multiagent software hardware simulation -developed in lcis laboratory provided the capacity of testing a method in a tool of simulation [23, 34] . diamond -decentralized iterative multiagent open network design -is a method that guides the designer from the requirements to the implementation. it is based on a-e-i-o decomposition for agent, environment, interactions and organization of the system. the agent dimension concerns the model of the reactive agent represented by a simple automaton, the cognitive agent manipulating information or a more complex agent based on a knowledge system. the environment is the context in which the agent reacts, its geographic location. the interaction dimension specifies the way the agents communicate; it mainly consists of protocols of interaction or the type of communication. the organization dimension reflects the structural relations between the agents (group, hierarchy, market). these four key concepts are considered under a global angle (the society) and a local one (agent view). the diamond method proposes an analysis phase of the problem in four steps: the situation phase helps to find the society's circumference to be represented by defining the limits of the system, agents and environment; the individual phase concerns the internal functioning of the agents (behavior and the knowledge); the social phase defines the relations between agents, particularly by integrating into its knowledge communication protocols and information structures to understand the society organization and -the phase of socialization consists of integrating the individual agents into the society by adding the social influences into its behavior (possible answers in the requests from the outside, the launch of interaction protocols and the choice to be made according to its position in the organization). we decide to use this method because it allows explicitly designing the behavior of each individual. however, coupled with the mash simulator platform, it is useful to view the emergent system from the interactions between the individuals. the process in fig. 3 can be modelized by agents able to be simulated in the mash simulator. the mash simulation platform was used to simulate systems with embedded and software agents. it is suitable to our problem because we plan to provide a tool for collecting data with phones or tablets applications. to have an individual-centered vision of the process is one of the advantages of this simulation. afterwards, we will be able to contribute to the improvement of the process with an exterior view by proposing changes and ideas to reduce the response time for example. this section shows the steps to break down multiagent system's elements. to start the analysis, each individual agent's behavior is studied. it is a way of seeing things at a micro level. the phenomenon at macro level does not change and the process remains the same even if observer's level changes. the objective is to be able to adjust the behavior of each individual agent and probably to add some skills to certain nodes or node types. the first step of this approach is to find out what to model as agents from information of the process. in our study context, agents are: health centers, general referral hospital, health areas, provincial health division, any national health entity related to the administrative structures, health zone executive team and rapid ripost team for human team working group. figure 4 shows for instance the internal behavior of one agent (rapid ripost team). for each agent, we have to list all its skills, what information will be required to store or to handle and how agent acquires this information. this information should be acquired directly by perception (e.g. the user grasps something) or on demand by asking other nodes (higher hierarchy or same level nodes). in this step, we obtain for each agent a vision of the relevant knowledge to perform its individual tasks. that information is required by agents working in the same environment. the result is a set of tasks that an agent can perform. these tasks correspond to the skills of each node. some skills are executed by one agent without the need of other agents. but to achieve a goal, an agent should have partial information and should ask to other agents to complete their goal. however, this will result to a cooperative behavior in place of an individual one which is entirely internal to agent. this kind of social behavior reflects an interaction between several agents: either to gain information or to share tasks. social behavior. in this second step, we will have to create interactions between nodes for example to back up information (health center to health zone executive team) or to receive orders (health zone executive team from provincial health division). these interactions should intervene between different partner groups such as health areas. in the implementation, we define very simple interaction protocols for data exchange such as receiving information, answers/queries or order to perform a task. for some tasks, such as health alert surrounding areas, it is no longer just a request for information but cooperative behavior brings into play several kinds of agents. to communicate with others, agent uses interaction protocols. we will there-fore express how this behavior will be realized by defining a more sophisticated interaction protocol than query/response. various protocols are available for negotiating, giving orders, waiting for answers. the interaction patterns that will govern this cooperative behavior will be organized between the agents. interaction protocols. the protocol is a part of the agents' knowledge. in our simulation, agents have a list of protocols they should initiate to answer others. for the moment, we use a simple protocol with two states as represented in fig. 6 : "inform" and "inform back". for example, agent a1 launches an instance of protocol p1, with the state s1. the agent a2 receives a call from a1 with the performative "information" in the state s1. a2 knows the protocol and searches the next transition; it slips into state s2 and sends an acknowledgment to agent a1. a1 treats the message and the conversation finishes. the acl fipa compliant performatives are used: • rdcmessage.acl_query_ref for queries/answers, • rdcmessage.acl_request for an order to perform a task and, • rdcmessage.acl_inform for inform/acknowledgment. position in the organization. the last step is to consider an agent's position in organization when he initiates interactions with others. an organization should be a hierarchy or a simple group. for example, to alert neighboring health zone, health center agents must know the surrounding areas of health zone, which is a group or an organization in the multiagent system, and make a decision based on its position in this group. the agent's position in the organization is integrated in the decision loop. agent's internal decision making loop. the previous steps showed agent's skills, agent's complex behavior (internal and social) and the knowledge of interaction protocols. on the hypothesis that each agent gets this information, we can now build the agent decision loop. on one hand purely individual behavior runs only with an agent's context information and does not need other agents to complete the agent goal. on other hand, social behavior involves relationships between agents. an individualcentered approach defines agent at micro level so that interactions with other agents have to be merged with the internal behavior in the agent's decision loop. the individual and cooperative behaviors are both integrated into the decision loop. in the individual behavior, a set of tasks is launched in the internal decision loop. in its decision loop, the agent should have to respond to the message from others, which are part of interaction protocols initiated by other agents or parts of the agent interaction patterns. these tasks have to be synchronized with the messages received from others agents. as an external view, huge decision loops, which are decentralized in the several kinds of agents, seem to be synchronized at the system level. but in fact, each agent decides in what state to pass according to its knowledge and the state of its interactions. agents would collaborate to achieve some objectives. to investigate on health area, rapid ripost team must wait for an order from health zone executive team. the later receives health data every week from health center and checks if the threshold of the followed pathology has been reached. the same collaboration is needed between rapid ripost team and health area, rapid ripost team and laboratory. the sequence diagram (fig. 5) gives a snapshot of the kind of collaboration found in agents concerned with an outbreak investigation. message format for interaction. the messages exchanged between agents contain sender and receiver agents, protocol information and data to manage like [sender; receiver; conversation; perform; protocol; inst_prot; state_prot; data]. the data follow a format according to the performative. to interact through a message sent by another agent, a simple protocol is established. for instance when rapid ripost team asks a laboratory to perform exams, he has to first check its state to be convinced that it can answer his request. a simple protocol with acknowledgment is used. the agent changes state when he asks for information and when he receives answers to his request (fig. 6) . in a future simulation, a negotiation protocol with a call for proposal to several laboratories will be tested. however, the agent launching the conversation should negotiate among laboratories which one is available, near or powerful. the analyze of the cycle of outbreak response presented on 4.4 isolated some individuals playing different roles to achieve the objectives assigned to the riposte process. a simple class diagram (fig. 7) gives a quick overview of main actors. health center as agent can exchange message with agent 'health area' to get information about rivers crossing the area of its location. since this knowledge is crucial to anticipate cholera epidemic in raining season, this collaboration is very important. rapid riposte team (rrtagent) investigate on suspicious case in a health area which means visiting population and health centers found in it if the disease requires laboratory analyzes to confirm the case rrtagent must take samples and transmit it to laboratoryagent available. reported worldwide due to fear of economic consequences, often insufficient surveillance systems and a lack of standardized terminology to define a case of cholera. 2 the disease results from the absorption by the mouth of water or food contaminated by the cholera vibrio. after a few hours to a few days of incubation, violent diarrhea and vomiting occur, without fever. in the absence of treatment, death occurs in 1 to 3 days, by cardiovascular collapse (fall in blood pressure) in 25 to 50% of cases. mortality is higher among children, the elderly and vulnerable individuals. the treatment consists essentially of compensating the digestive losses of water and electrolytes. rehydration is given orally or intravenously, depending on the degree of dehydration. the improvement is noticeable after a few hours and healing, without sequelae, is obtained in a few days. antibiotic therapy is recommended by who only for severe dehydration. in dlm is a disease control direction located in kinshasa which collects the national data for disease monitoring. it provided us with ten years data of cholera outbreak from 2008 to november 2017. we extracted the data for kinshasa grouped on its 35 health zones to compare with provincial health division's data. provincial health division's data contained more details about health centers that reported the suspicious cases and theirs health area. we analyze the data from first january 2017 to december 30th 2017. we reported cumulated data of each health zone to a map, as shown in fig. 8 , to find their neighboring and try to suggest the best way to establish collaboration between them in order to stop disease propagation. we focus on two groups of health zones. in the first group we have binza-météo, mont-ngafula, kokolo and kintambo. the second contains limete and kingabwa. our hypothesis was: if actors from each heath zone could exchange disease information with their neighbors as soon as an outbreak happens, it would be possible to reduce the propagation. for example, epidemic began at kintambo on january 2017. as the communication and sensitization weren't establish with its neighbors, some weeks after disease was reported from kokolo and binza-météo. to test and evaluate our approach, we adapted mash simulator developed by lcis lab for a wireless sensor multiagent system [34] . we focus our simulation to these health zones: binza-météo, mont-ngafula, kokolo, kintambo, limete and kingabwa. the simulation concerns precisely kintambo and limete that register more suspicious case of cholera outbreak during 2017 and from them other neighboring health zones were affected. the main idea is to see how the future system would react if each actor of health system could perform its own task with autonomy. these experiences could result to many scenarios and the best of them will be proposed to drc's health system to reduce decision time as each actor can execute his talks according to the knowledge of the environment and the outbreak determinants he will be provided with. we chose two health-zones of kinshasa provincial health division for the simulation. kintambo and limete are health zones that register respectively 57 and 64 suspicious cases with death in 2017. the epidemic began from them and the propagation of disease affected their neighbors with an important amount of suspicious cases. kokolo counted 208 cases while binza-météo registered 153 suspicious cases at the same period. to respond to an outbreak noticed at a health center, health staff of the concerning health center must refer to health zone executive team. in their turn, health zone executive team must refer to provincial health division and provincial health division to central level. this chain of hierarchical contacts can enlarge time decision. in our simulation, we worked with these hypotheses: (i) each health center actor must contact immediately its health executive team as soon as it encounted a suspicious case; (ii) health zone executive team cumulate suspicious case and create an rrt when the disease's threshold is reached; (iii) rrt can contact the nearest laboagent able to answer to his request or to use his information to make decision. we considered twenty-eight health zone executive team in yellow or red, forty seven health center in green or blue, rrt in grey one provincial health division and one medical test laboratories (laboagent). the first suspicious case was detected in health center #49 in green. figure 9 illustrates those actors working as agents. the below map represents health zones in north of kinshasa. the simulation trace file (fig. 10) shows the communication between agents. as they are autonomous they perform their own tasks like "cumulate new cases", "conduct the investigation on health zone # 16" and manage messages like "realize analyzes from health center", "receive sample to_ analyze" or "report suspicious case detected". hzet agents (#16 and #13) are waiting for a suspicious case message from any hc. whenever health zone executive team agent receives such a message he computes cases and compare with the threshold of disease monitoring to decide the necessity of building a rrt agent able to investigate in health centers around suspicious case in health area. the visited health center will turn from green to blue color. rrt agent could send samples to laboagent while searching for new cases in health areas. the organization of outbreak riposte could depend on the results from laboagent and investigation report from rrt agent. the agents operate independently: hcagent transmits data to health zone executive teamagent, rrtagent completes a full investigation, laboagent conducts medical testing and transmits results to hzetagent and phdagent which manages all information from health zone executive team under its supervision. message synchronization between kinds of agent is done in the agent's decision loop. a protocol with two states is used and implements kqml-like performatives. the four numbers in the message are "1" for inform (give information), "2" for query information (ask for an information) and "3" for request (ask for a task to be done). the agents communicate and achieve their goal by reacting to messages from others or executing their inner task as response to a query. with positive results from laboagent, health zone executive team sends warning and preventive measures to his all health centers. phdagent can also alert the surrounding health zone executive teams. in this paper, we have showed how multi-agent system can improve the organizations' tasks in order to decrease the time to response when an epidemic disease is suspected or detected. a main research result can be highlighted: the use of a multi-agent method previously dedicated to wireless sensors and applied to human organizations. we have proved that the method was generic enough and gave good results with real data and a hierarchical and complex health eco-system. mas is often used in health domain and our paper's result complete the panel of applications with real data and under eco-system constraints. some limits must be underlined: (i) in order to adapt the method, we have defined hypothesis that strongly constraint the models; (ii) the stakeholders have been introduced in the method only with their job characteristics and adding their experience in the simulation can probably enhance the results. nevertheless the multiagent method can cover only one part of the global problem. we discussed about an organizational method useful to enlarge all aspects of the problem laid down. we propose in the future works to widen the approach to take into account different and complementary aspects: health data collection and transmission, health data quality, improvement of the complete riposte process, improvement of the health system organization. in order to, we propose a three-phases innovation method named chicken useful to supervise and improve the epidemic disease riposte: -phase 1: define the life cycle of the epidemic disease to watch over. -phase 2: health data monitoring -phase 3: riposte and feed-back each phase of this innovation method is build with models, methods' fragments, tools coming from different sciences domains and proposes a road map to improve the riposte and the health data quality. then the multiagent method becomes a method's fragment in the wider method chicken. towards an agent-based model to monitor epidemics and chronic diseases in dr congo programme des nations unies pour le développement: rapport sur le développement humain organisation mondiale de la santé rdc: guide technique pour la surveillance intégrée de la maladie et riposte icts for poverty alleviation: basic tool and enabling sector supporting home based health care in south african rural communities using ussd technology mobile phone text messaging: tool for malaria control in africa improving health information systems for decision making across five sub-saharan african countries: implementation strategies from the african health initiative mobile phones and economic development in africa effects of a mobile phone short message service on antiretroviral treatment adherence in kenya (weltel kenya1): a randomised trial the effect of mobile phone text-message reminders on kenyan health workers' adherence to malaria treatment guidelines: a cluster randomised trial a methodology for agent-oriented analysis and design a methodology and modelling technique for systems of bdi agents desire: modelling multiagent systems in a compositional formal framework voyelles the gaia methodology for agent-oriented analysis and design analysis and design using mase and agenttool prometheus: a methodology for developing intelligent agents bdi agents: from theory to practice agents acting and moving in healthcare scenario -a paradigm for telemedical collaboration designing embedded collective systems: the diamond multiagent method modélisation et gestion de flux par systèmes multiagents: application à un système d'aide à la décision en épidémiologie solutions multi-agents pour la prise en charge à domicile des séniors mise en place d'un système d'information décisionnel pour le suivi et la prévention des épidémies evaluation of disaster response system using agent-based model with geospatial and medical details multi-agent system based efficient healthcare service using multi-agent systems to support e-health services a multi agent system for hospital organization multi-agent systems for e-health and telemedicine a multiagent system enhancing home-care health services for chronic disease management multi-agent system applications in healthcare: current technology and future roadmap a multiagent tool to simulate hybrid real/virtual embedded agent societies usability evaluation of a collaborative health information system -lessons from a user-centred design process multi-level agent-based modeling -a literature survey key: cord-004935-z86x3hnu authors: baykasoglu, adil; durmusoglu, zeynep d. u. title: a classification scheme for agent based approaches to dynamic optimization date: 2012-01-03 journal: artif intell rev doi: 10.1007/s10462-011-9307-x sha: doc_id: 4935 cord_uid: z86x3hnu several papers in the literature employ agent-based modeling approach for providing reasonable solutions to dynamic optimization problems (dops). however, these studies employ a variety of agent-based modeling approaches with different strategies and features for different dops. on the other hand, there is an absence in the literature of a formal representation of the existing agent-based solution strategies. this paper proposes a representation scheme indicating how the solution strategies with agent-based approach can be summarized in a concise manner. we present these in a tabular form called “agent based dynamic optimization problem solution strategy” (abdopss). abdopss distinguishes different classes of agent based algorithms (via communication type, cooperation type, dynamism domain and etc.) by specifying the fundamental ingredients of each of these approaches with respect to problem domain (problems with dynamic objective functions, constraints and etc.). this paper also analyzes 18 generic studies in the literature employing agent-based modeling based on abdopss. on the other hand, some of the researchers studying operations research (or) have attempted to solve these problems by using the well-known traditional optimization techniques like linear programming and integer programming . however, solutions obtained for a certain time that is desirable or optimal for that time slot, often have not be preferable for another time slot. furthermore, obtaining exact solutions using mathematical methods for each time slot has been unaffordable or even sometimes has not ended-up with feasible solutions. therefore, notwithstanding its widespread use, the static optimization approach has increasingly approached its limits (borst et al. 2005) . this fact certainly inspired many researchers to employ agent-based modeling for the optimization of dynamic systems. as indicated by jung et al. (2011) , the agent paradigm has been shown to be a promising approach to develop intelligent, heterogeneous, and open systems, because of its features, such as autonomy, flexibility, and cooperative problem solving behavior (jung et al. 2011) . in this regard, reusable agents have provided many opportunities and they saved researchers from moving beyond reinventing, representing, and re-implementing the problem and thereby the cost of providing solutions has been decreased (neches et al. 1991) . agent-based models addressing dynamic optimization problems (dops) have also allowed several optimization mechanisms to track the moving optima in the search space. this fact certainly inspired many researchers to employ agent-based modeling for the optimization of dynamic systems. even though there have been several papers in the literature employing agent-based modeling approach to provide significant solutions for dops, each of these existing studies employs different agent-based modeling approaches having different strategies and agent features. therefore, currently, we don't have a common successful methodology for the design of agents (corchado et al. 2008) . with dependability and safety in mind, it is vital that the mechanisms for representing and implementing agents are clear and consistent (fisher et al. 2007) . in this regard, it is certain that there is a vacancy in the standard representation and classification of agent-based solution approaches employed for dops. a previously developed scheme (calégari et al. 1999 ) for evolutionary algorithms for combinatorial optimization has also inspired us to prepare a standard representation format. in this perspective, one of the objectives of this paper is to present the fundamental features/ingredients of dops solved by agent-based modeling. the features positioned in a standard scheme provide an opportunity to present dops in a standard way. it can succeed further and in future studies these representation forms can be used to study the role of these features and their importance for solution quality. thereby, this study provides a basis to analyze the best ways of implementing agent-based approaches to the dops. in this perspective, this paper presents an attempt to develop a notion for dops solved by agent-based approach. the paper shows how solution strategies with agent-based approach for dops can be summarized in a concise manner by informing about the agents employed and the key elements of dop's. a classification scheme is introduced and presented in a tabular form called agent based dynamic optimization problem solution strategy (abdopss). abdopss distinguishes between different classes of agent based algorithms (via communication type, cooperation type, dynamism domain and etc.) by enumerating the fundamental ingredients of each of these agent-based approaches. at the end, possible uses of the abdopss are illustrated and exemplified for some agent based approaches. the paper is organized as follows. sections 2 and 3 introduces the features listed as the ingredients of abdopss. section 2 specifically focuses on the features of typical dops. subsequently, sect. 3 introduces the generic features of agent-based approaches. in sect. 4, some typical agent-based solution approaches that are implemented for dops are described by using the abdopss. section 5 illustrates the application of abdopss for 18 relevant articles collected from the literature. finally, sect. 6 presents concluding remarks with the possible benefits of the proposed classification scheme. a dynamic problem is considered to be a problem that changes some or all of its characteristics in time (lung and dumitrescu 2009) . thereby, optimization problems that are subject to changes, belongs to the class of dynamic optimization problems (dops) (allmendinger and knowles 2010) . in other words; dops are "optimization problems whose optimal solution changes over time during the optimization, which could result from change of environmental parameters, change of constraints, change of objectives and change of problem settings (representations)" (jin 2004) . all of the changes that are mentioned above take place in the dynamic environment and these changes create unpredictability even chaos in some situations. these changes can be opportunities improving the solutions or sometimes they can be threats decreasing level of desirability of a solution. there are different approaches in the literature that is classifying these changes in the dynamic environments. yang (2007) defines environmental change as: cyclic and random. in cyclic change of environments, with the time going, an old environment reappears exactly (yang 2007) . a cyclic change covers the same fixed length sequence of contraction and/or growth is repeated over time (abbas et al. 2004 ). the pattern repeats itself indefinitely (abbas et al. 2004 ). however, in random change, there is no information about the structure of change. random change is also called as acyclic and in random change there are no repeated patterns over the time frame; that is, there is an indefinite random sequence of contractions, growths, and random changes (abbas et al. 2004) . all of these changes in dynamic environment cause significant changes in model's structure, objective function, restrictions/constraints, parameters, and in the decision variables. bui et al. (2011) describe three of these changes through the examples as presented below. time-varying objective functions: for example enemy units arrive at a location, making some parts of the objective more difficult. the objective function is not constant over time. therefore, the objective value of a solution can be different at different times. this usually causes the occurrence of new optima (bui et al. 2011) . time-varying variables: an example problem of this category is dynamic machine scheduling where unexpected new jobs arrive. a time-varying variable can be used for determination of the objective value, but can be a late addition or change (bui et al. 2011) . time-varying constraints: for example the precedence relationship of tasks. the objective function and variables do not change for this category; however, the constraints may change over time. this category of change does not change the fitness landscape driven by the objective function, but it will affect the areas of feasibility. it is interesting to note that this category of change has not been analyzed in detail in comparison to the other two categories (bui et al. 2011) . cruz et al. (2010) provides a valuable survey of research done on optimization in dynamic environments over the past decade. in addition to the content of their paper, they provide a web site (cruz et al. 2011 ) and present characteristics of the dops with respect to objective function or the restrictions change with time. we have considered the study of abbas et al. (2004) to illustrate correspondence between the change types and the changes in the modeling components. abbas et al. (2004) defines change in the sample bias parameter change since processing times for machines are estimated using sampling, an unexpected sampling error can end up with the change in actual processing times six common kinds of changes for a typical system having certain data flow structure. these changes are: change in the model, change in the number of concepts, change in the number of features, change in the level of noise, change in the class distribution, and change in the sample bias. table 1 both summarizes the change states defined by abbas et al. (2004) and their corresponding change in a typical mathematical model which is similar to the one defined by cruz et al. (2010) . examples of each change categories have also been provided through the possible changes in a dynamic production facility. according to the example case, a typical production facility that is operating in a dynamic environment faces with several expected and unexpected changes with the above given change states. change 1: change in the model: change in objectives of a system may lead a significant "change in the model". think of the production facility trying to minimize the cycle time and adjusting all of its production strategy with respect to cycle time minimization. if the cost of tardy job increases significantly and the managers will be obligated to change their objective as the minimization of the number of tardy jobs. in this case, the objective function, constraints and the number of decision variables along with the type of parameters may change. change 2: change in the number of concepts: sometimes, concept classes may not be as clear-cut as initially thought so that additional classes may arise or old classes may vanish over time (abbas et al. 2004) . in case of change in the number of concepts, resources used by the system can increase or decrease. new machines can arrive to the system or some machines can be removed from the system because of the oldness or in functionality. this change can affect the model by changing the variables used in objective function and constraints. this type of change is not expected to create structural change. it is a kind of reconsideration of the model with added/removed variables that is significant for decisions. change 3: change in the number of features: in some problems, the number of available features characterizing a problem instance may vary over time. additional information may become available or new tools may be developed that allow more accurate classification (abbas et al. 2004 ). orders received for new product types are a typical example of this change. in the face of such a change, a production facility may face with new operations required and this can lead changes in objective function, constraints and the number of decision variables. change 4: change in the level of noise: the noisiness of the data may change as well (abbas et al. 2004 ). this type of change is particularly common when dealing with data coming from sensors. for example, acoustic data collected in an open area can have different noise levels based on the state of the environment (abbas et al. 2004) . in this respect, in the production facility example, if some customers broke the deal and cancel their orders for a production facility, the plans unfortunately changes and it may cause change in objective function and constraints. change 5: change in the class distribution: abbas et al. (2004) exemplifies the change in the class distribution with sars virus. when the sars outbreak took place, the class distribution changed almost every day with more healthy people becoming infected effectively increasing the proportion of positive cases. similar to sars example, in case of an economic boom, arrival of some certain jobs to the system can be more frequent and this leads change in parameters. change 6: change in the sample bias: since system parameters like processing times for machines are estimated by the sampling of several actual processes, an unexpected sampling error can end up with the change in actual processing times and/or system arrival times. all those changes defined above also directly effects the unpredictability of the problem domain. level of unpredictability is vital to determine a proper agent-based solution strategy. although there are several efforts to match the corresponding changes with the unpredictability, they are still in their infancy. therefore, there is a serious need for measuring the level of unpredictability of the dops. the relation between the corresponding changes in constraints, objective function and the both with uncertainty and dynamism has been previously presented by (cruz et al. 2010) in the web site (cruz et al. 2011 ) as shown in fig. 1 . while their framework as shown in fig. 1 does not cover the variability in parameters and variables; it can still be a good reference for future studies. considering parameter variability and variability in the number of decision variables will certainly increase the complexity of such an analysis. it is also remarkable that dynamism is one of the components of environmental uncertainty. indeed, as presented in fig. 2 , environmental uncertainty has three dimensions: dynamism, heterogeneity, and hostility (newkirk and lederer 2006) . dynamism, as being the one dimension of the environmental change, (teo and king 1997) found to have two dimensions. the dimensions are referred to here as changeability (i.e., concerned with the rate of obsolescence and of technology change) and unpredictability (i.e., concerned with competitors' moves and demand changes). the rigor of that study and relative recency of the findings motivated the use of those two dimensions in the current research (newkirk and lederer 2006) . in this regard the above defined six different changes cover different level of changeability and unpredictability. categorizing changes into these two dimensions is a topic of future research. objective function objective function different dops tried to be solved in the literature for different scopes by using various approaches. some of the existing studies on dops via agent-based modeling propose only frameworks some other studies also present applications. it is remarkable to state that a framework is not a detailed hypothesis or set of hypotheses; rather, it is a suggested point of view for an attack on a scientific problem, often suggesting testable hypotheses (aggestam and söderström 2005) . therefore, studies covering solely the frameworks do not provide comparable outputs. some of the applications cover real-life instances and some have hypothetical data obtained via simulation. some of the studies attempting solve dops by using agent-based modeling use test problems that has been previously presented in the literature. thereby, they compare their findings with the best known results in the literature. some other studies compare the results with themselves by providing different methodologies to solve the problems. examples of these cases will be discussed in the applications section. although some problems like scheduling, production planning and travelling salesman have been previously solved as they are static problems, indeed these problems are not totally static problems due to the nature of the real life. consider an example of scheduling and production planning problems where new machines with advanced capacities are required to be included to the systems or some machines are required to be removed from the system and/or new products are can also be introduced to the system etc. another example is for a more real life oriented travelling salesman problem (tsp) has been defined by (homayounfar et al. 2003) . they defined three different types of changes for tsp: * changing the distances (time) between the cities (due traffic congestion, road repair etc.) * changing the number of the cities (i.e. adding or deleting some cities) * swapping the cities a distance (e.g. the distance between the first and second cities) is changed after a specific time or specific generation, determined by dynamic frequency. the amount of change is referred to as the dynamic rate. these two parameters are set prior to the test. after each period of time, which is after a generation specified by dynamic frequency, dynamic rate is added to the specified cost value (i.e. distance between city 1 and 2). this shifts the peak (optimum cost) to another location, so a new global solution can be found (homayounfar et al. 2003) . numerous examples can be considered for real-life oriented versions of these well-known problems. the distinction between these conventional problems and real-life oriented ones can be performed as follows. if during the solving of an optimization problem, parameters of the system (i.e. objective and constraint functions) do not change due to nature of the problem then optimization is static, otherwise it is considered to be dynamic (homayounfar et al. 2003) . in such environments, the optimization process is non-terminating. in the real world, a function to be optimized may vary from time to time and the optima have to be found in time (guan et al. 2005) . therefore, it is very difficult to optimize such type of dynamic optimization problems with conventional methods which are developed to deal with non-stationary environments. moreover, it is also quite hard to define dependencies or correlations between the solutions obtained at a time slot and in another time slot. as stated by o'hare et al. (2007) , the real world is both unpredictable and unforgiving. decisions often need to be made where the contributory evidence is uncertain, incomplete, contradictory and highly dynamic (o'hare et al. 2007 ). therefore, performing design of experiment may not end up with significant findings. since finding a feasible solution to the static problem is np-hard, we must make certain assumptions about the input. intuitively, if the input is such that even finding a static solution is hard we cannot expect to find a good solution with respect to the dynamic objective function. thus the problem instance must be "easy enough" that a relatively straightforward agent can find a feasible solution at each time step. in practical terms, this means there must be enough resources to easily satisfy the demand if we ignore the quality of the solution in the sense of the dynamic objective function. in the worst case, we can fall back on the heuristic to find a feasible solution. in fact, agent-based algorithm will degrade gradually to this extreme, but should perform much better in the common case. this challenging dynamism was apparently a significant problem for researchers and it could only be effectively handled by using more advanced approaches like agent-based modeling. although there is no universal agreement in the literature on the precise definition of agent, their property of being autonomous has been common to all definitions (kulkarni and tai 2010 ). an agent is an encapsulated computer system that is situated in some environment and that is capable of flexible, autonomous action in that environment in order to meet its design objectives (jennings et al. 2001) . agents that act in an autonomous fashion to perform delegated tasks have aroused much attention over the last decade (liu et al. 2006) which are named as intelligent agents. an intelligent agent performs interactive tasks tailored to a user's needs without humans or other agents telling it what to do (máhr et al. 2010) . the intelligent agent has been proposed as a software design paradigm, which is different from the sequential, the structural and the object-oriented approaches by its autonomy in deciding when to invoke its action (parunak 1997) . the fundamental approach of agent based system is to simulate real-world systems with a group of interacting autonomous agents modeled as computer programs (zhou et al. 2009 ). since agent based modeling is a consideration that is developed according to system requirement, it is expected that they have several common and distinguishing characteristics. in fact, in spite of the growing interest in multi-agent systems, there is no agreement on what actually constitutes agenthood, that is, what are the fundamental characteristics of agents (garcia et al. 2004 ). on the other hand, jou and kao (2002) defined the characteristics of an agent based on agent's behaviors: fundamental, auxiliary, implicit and application oriented (role based). the first one is fundamental characteristics. as a goal is state of affairs to be achieved in the environment, an agent must perceive and process the environment changes (jou and kao 2002) . the second one is auxiliary characteristics. an agent also needs information from users and other agents through communication mechanisms such as user interfaces and agent communication language to complete the delegated tasks (jou and kao 2002) . in addition to fundamental and auxiliary characteristics, some implicit characteristics with presumptive feature exist within an agent (jou and kao 2002) . trustworthiness (veracity and benevolence), perceptibility, rationality, persistence, flexibility, and competence are some examples (jou and kao 2002) . the final characteristics are related to specific applications, e.g., the surfing behaviors of cyberspace, the facial expressions attached to entertainment, and so forth (jou and kao 2002) . all these features make agent technology an interesting approach for a wide set of applications (garcia-montoro et al. 2007 ). an agent based model has a set of agents defined by the model creator. number of agents can be variable or constant. their types or functions may also vary or they can be assigned for repetitive tasks. in meta-heuristics based agents, number of agents can directly be proportional (or the same) with the population/colony size. both "number of agent types and number of agents" should be considered in the design of agent-based modeling for dops since they can affect the solution space directly. in his study, tan (1993) , claims that, as the number of agents is increased, state space exponentially increases in terms of the number of agents. a large state space means more state ex-ploration for the model that tan (1993) considers, and this yields slower learning. multi-agent system, also called 'self organized system' is a computational system in which multiple interacting intelligent agents work together to solve difficult problems which may be impossible for an individual agent (yan et al. 2010) . multi-agent based modeling allows complex natural behavior of various interacting entities to emerge from a set of simple individual rules (razavi et al. 2011) . communication is the most common means of interaction among intelligent agents. any observable behavior and its consequences can be interpreted as a form of communication (mataric 1995) . two different approaches have been defined for communication (genesereth and ketchel 1994) . direct communication covers the agents handling their own coordination. designer of an agent-based system can consider direct communication with respect to problem domain. the cost of communication and availability of obtaining qualified solutions are the factors to be considered for employing direct communication mechanisms in the models. on the other hand, assisted/indirect coordination covers the systems in which agents rely on special system programs to achieve coordination (genesereth and ketchel 1994) . one of extensively employed indirect communication method is stigmergy. stigmergy is a class of mechanisms that mediate animal-animal interactions. it consists of indirect communication that is taking place between individuals of an insect society by local modifications induced by these insects on their environment (hadeli et al. 2004) . although communication is essential for agent-based modeling, it may not be perfect as expected. in most of the current research on multi-agent systems, people assume that communication of agents is guaranteed (satoh et al. 2000) . however communication can be broken, suspended or delayed depending on the flow of data or information. in this regard, research of problem solving under incomplete communication is very important (satoh et al. 2000) . since dops may have varying objectives, constraints/restrictions and parameters; agents cannot be informed on the global state of the system by themselves. therefore, in order to achieve global objectives, coordination is essential to allow the agents to adjust their local states or conditions. communication provides channels for coordination. coordination is a property of a system of agents performing some activity in a shared environment (huhns and stephens 1999) . the degree of coordination is the extent to which they avoid extraneous activity by reducing resource contention, avoiding livelock and deadlock, and maintaining applicable safety conditions (huhns and stephens 1999) . figure 3 illustrates the taxonomy for the types of coordination defined by (huhns and stephens 1999) . panzarasa et al. (2001) prefer to use the generic term "interaction" instead of "coordination". they (panzarasa et al. 2001 ) define cooperation as working together to achieve a common objective and they also define negotiation as coming to a mutually acceptable agreement on some matter (panzarasa et al. 2001) . in other words, cooperation is coordination among nonantagonistic agents, while negotiation is coordination among competitive or simply self-interested agents (huhns and stephens 1999) . however, perhaps the most fundamental and powerful mechanism for managing interagent dependencies at run time is negotiation-the process by which a group of agents come to a mutually acceptable agreement on some matter (jennings et al. 2001) . negotiation underpins attempts to cooperate and coordinate (both between artificial and human agents) and is required both when the agents are self interested and when they are cooperative (jennings et al. 2001 ). planning distributed planning negotiation fig. 3 taxonomy of coordination for agent-based modeling (huhns and stephens 1999) there are different approaches used as optimization mechanism in the agent-based systems for dops. these methods can be classified as: heuristic based approaches, exact mathematical methods, and market based approaches. heuristic based approaches cover, evolutionary based approaches and other techniques like: particle swarm optimization, immune-based algorithms ant colony optimization etc. heuristics: for large size problems it is hard to obtain the optimal solution due to the large size of the problem files. in order to reduce the computational time, heuristic approaches can be used for obtaining solutions . heuristics includes some set of procedures for obtaining acceptable solutions to the problems by decreasing the time requirements. there is a natural correspondence between autonomous entities and meta-heuristics, and problem solving with an optimization problem (pelta et al. 2009a,b) . in the literature, agents are sometimes defined as individuals, particles etc. (wang and shixin 2010) which is the main actors for heuristics. if one takes a look at natural process like bird flocks or fish schools a strong similarity to multi-agent systems can be found very quickly (wagner et al. 2003) . in this regard, there are many studies adapting heuristic approaches to agent-based systems. mathematical methods: although it is too complex to deal with dops using classical mathematical methods, increasing power of computer technologies let some researchers to use mathematical methods. change frequency and the response time of the mathematical methods directly affect the solution's availability at the desired time. in this regard, the preferability of mathematical methods appears to be less when compared with the other methodologies. market based approaches: for an optimization approach to be considered as market oriented, it must at least fulfill the following basic requirements (karlsson et al. 2007 ): -there must be a well-defined market mechanism which includes some notion of prices (which often are expressed in terms of some monetary unit). the market mechanism regulates how negotiations and trade are performed among the participating agents and hence determines how certain commodities can be traded for certain other commodities (karlsson et al. 2007 ). -there must be some arguments for why the agent strategies are reasonably realistic, given the market model. that is, assume we have some model of information available for the different agents and a well-defined model for how they interact (the market mechanism). then the strategies must be consistent with the agents' attempt to maximize utility, given bounded rationality (karlsson et al. 2007 ). the first generic step of developing a solution strategy for any optimization problem is to define specifying class of the problem. class of the problem is an important factor that is affecting the required solution approach. therefore, the ingredients that characterize an agent-based modeling approach for a dop must somehow posses a presentation scheme to provide a solution strategy and a basis for comparison of one with another. therefore, this paper presents how solution strategies with agent-based approach can be summarized in a concise manner. in this regard, a classification scheme is designed and presented in a tabular form called abdopss (agent based dynamic optimization problem solution strategy). abdopss distinguishes different classes of agent based algorithms applied to dops (via agent-related features like: communication type, cooperation type and problem related features like: dynamism of objective functions, parameters etc.) by enumerating the fundamental ingredients of each of these algorithms with respect to problem domain. the ingredients that have been considered for abdopss have been discussed in the previous sections. the main structure of the table consists two rows (row 1, row 2 as illustrated in table 2 ), indicating the problem related features (row 1) and agent related features (row 2), respectively. at each corresponding column of a row, an entry, gives the necessary indication for the corresponding criteria. table 2 shows such an empty table and corresponding values for the table. since it takes a considerable time to determine features listed by abdopss, 18 articles are exemplified within the contents of this paper. it is remarkable to state here that, this paper does not seek to find all publications on dops and represent those using abdopss, but it is rather sought for exemplification of the proposed abdopss. in this regard typical versions of different approaches have been tried to be illustrated in table 3 . two rows are merged into one for easiness of readability. the studies covered here are roughly classified and some unclear parts made it relatively difficult to prepare abdopss. it is certain that those 18 papers could be better presented by their developers. pelta et al. (2009a,b) presented a multi-agent decentralized cooperative strategy (mad-cos) for dynamic optimization problems. in madcos, a population of cooperative agents has moved over a grid containing solutions. in order to test and analyze both different communication schemes for the agents and real need of an explicit diversity preserving mechanism in madcos, researchers focused on cooperation and diversity mechanisms. moreover, two types of diversity mechanisms were investigated. different configurations of the moving peaks benchmark problem were employed as a test bed. a set of computational experiments have been set to evaluate how different communication strategies affect the search; and the usefulness of having explicit and implicit diversity mechanisms. last offline error (oe) ("accuracy") and relative error between the best found solutions ("error") have been recorded to measure the performance of each configuration. results showed that having an researchers indicated that regarding diversity, the implicit mechanism provided by their multi-agent model seems to be enough for the scenarios tested. however, it was stated that the explicit mechanism proposed, based on randomization, did not perform as expected. wang and liu (2010) presented an agent-based evolutionary search algorithm (aes) for solving dynamic travelling salesman problem (dtsp). in their study, the term agent indicated a candidate position in the search space. agents resided in a lattice-like environment and a collection of such agents were termed as population. they applied a recombination and local updating procedure to the fittest agent referring to a predefined neighborhood. they also combined the perturbation learning strategy to further reinforce the performance of the local updating rule and hope to seek the global optimum rapidly under changing environments. dynamic version of kroa100 tsp was selected as the case of implementation. offline performance which was the best of generation fitness averaged over 100 runs and then averaged over the data gathering period. experimental result and relevant t test result indicated that the performance of aes was excellent. researchers indicated that the superiority of aes lie in its faster convergence and optimum tracking ability in dynamic environments. it is concluded that aes algorithm had a more quickly and robust convergence capability than standard genetic algorithm (sga) on dynamic problem. billiau and ghose (2008) proposed a new algorithm for solving distributed constrained optimization problems (dcops). the proposed algorithm, which has been called as support based distributed optimization (sbdo), has used agent level objectives instead of weighted or soft constraints that have been used by other dcop algorithms. in dcops, constraints/ objectives of the problem change over time by adding or removing constraints/objectives. researchers compared the performance of this algorithm with asynchronous distributed optimization (adopt) and distributed pseudotree optimization procedure (dpop) by using 120 meeting scheduling problem. result showed that, although sbdo algorithm has not guaranteed to find the optimal solution, it reliably found solutions within two percent of the optimal solution. results also revealed that sbdo was able to find near optimal results significantly faster than adopt and in a comparable time to dpop. máhr et al. (2010) compared two structurally different planning approaches, which were agent based solution and on-line optimization approach, for drayage operations in an uncertain environment. the structurally differentiating feature of the two solution approaches was the level of control: centralized (on-line optimization) and decentralized (agent based solution). dynamic vehicle routing problem (dvrp) with two types of uncertainty, arrival time and service time, was employed in order to compare the performance of agent-based solution and on-line optimization approach. moreover, scenarios were developed under different arrival time and service time. in their study, containers and truckers were defined as agent. each container agent sold itself on an auction to the truck agents. researchers concluded that when the online optimization approach performed better, it was by capitalizing the optimal (or near optimal) balance between routing and rejection cost. when agents performed better, their flexibility provided by their distributed nature was the competitive advantage. voos (2009) focused on dynamic resource allocation problems especially in continuous systems. in this study, resource allocation was expressed as an optimization problem, which could be decomposed into single optimization problems, under certain constraints. researchers proposed to solve this optimization problem in a distributed fashion by using multiple agents. these agents have acted as local optimizer and coordinated their local solutions to an overall consistent solution. in this study, market based interaction mechanism was employed and the agents have calculated and negotiated complete supply and demand trajectories using model based predictions. in order to test the performance of the proposed approach, researcher employed this approach to three technical examples. results revealed that this approach could be used to cope with resource allocation in dynamic environments. li and li (2009) proposed a multi-agent coordination and integration method which has been called intercommunication job-sharing hybridization, for solving complex problems. these complex problems could be decomposed into smaller separate sub-problems, with communications/exchange between sub-problems identified, human decision-makers' roles clearly defined and managerial judgment incorporated. with the proposed method, the overall problem was divided into distinct jobs which were then assigned to relatively independent and distributed agents. these agents have shared data, information knowledge and carried out different tasks synchronously or asynchronously in the context of internet/intranets/extranets to produce solutions. in addition, these agents have linked human participants' judgments and intuition together for joint problem-solving. the architecture of the internet-enabled multi-agent-based hybrid support system for international marketing planning was exemplified. researchers constructed a prototype, which has been called agent international, of the proposed multi-agent hybrid framework. this prototype covered some key features of the conceptual framework and its architecture. the potential and value of the prototype was evaluated by a questionnaire which was prepared and delivered to the corresponding firms in london. according to the responses, researchers concluded that the prototype system was rated somewhat moderately in helping understand pertinent marketing decision making factors, exploring various alternatives, performing analysis, coping with uncertainty, incorporating managers' judgment an improving the confidence and quality of international marketing decision-making. tang et al. (2004) presented an auction based dynamic optimization decision support system. the presented decision support system was applied in solving an automobile load makeup planning problem. the proposed system included three types of agents. these were the yard agents, representing the shipping yard, the truck scheduler agent, representing the transportation company who sold trucks and executed transportation and load agent representing a truck during the planning state. each type of agent had its own behaviors. in order to solve the static load make up problem, researchers implemented 3 heuristics which were empirical (em), minimum spanning tree (mst), vehicle routing optimization (vro). in addition to these heuristics, a virtual market enabled by auction mechanism was employed to solve dynamic optimization problem. researchers developed the mixture of static optimization with mst algorithm and dynamic optimization with the auction mechanism, mst dyn, at the beginning of the day, they apply static optimization on the guaranteed information, and then they apply the dynamic optimization until the loads are fixed at the beginning of lining up. in order to test the performance of the algorithms, same scenario was used and each algorithm has run for 120 days. transportation cost and dwell time were selected as the performance measures. results showed that, em algorithm produced the worst performance and mst dyn the best. the mst algorithm produced the worst performance and em algorithm the best based on the transportation cost. researchers indicated that mst dyn algorithm produced a little bit more transportation cost than em algorithm. therefore, it was concluded that mst dyn algorithm produced the best comprehensive performance. berro and duthen (2001) presented an optimization method in dynamic environment. the proposed method, using software agents, tried to provide accurate solutions and react quickly to changes in the state of the problem. in this method, software agents were randomly created and they try to colonize an optimum of the function. the system composed of two parts namely control system and agent. an agent has not communicated but it perceived the presence of other agents. when the functions to optimize and the dimensions of the search space have been defined, the user must specify two parameters namely force which would influence the searching speed and epsilon which calculated precision of the optimal points. in order to test the proposed method, 2 cases of optimization problems, multimodal function and multi-objective function, were used. the test results were compared with the genetic algorithm based approaches. researchers concluded that the first tests result of the proposed method were satisfactory in particular for the computing speed and precision. jiang and han (2008) presented a simulated annealing (sa) based algorithm to solve real time multi-agent decision making problem. in this study, the sa algorithm was implemented in a centralized version and performed by the agents in parallel, without assuming the availability of communications. since there was no standard benchmark to evaluate multi-agent decision algorithm, a random generator was used to generate all test sets. the sa algorithm was tested by comparing it, with other algorithms, especially with variable elimination (ve) algorithm with respect to the scalability and relative payoff. results revealed that, the proposed method was almost optimal with a small fraction of the time that ve took to compute the policy of the same coordination problem. in addition, researchers indicated two main benefits of this approach as follows: the time taken by the algorithm has grown polynomial with the number of agents and the algorithm could report a near-optimal answer at any time. researchers concluded that, sa was a feasible approach for action selection in large complex cooperative autonomous systems. zhou et al. (2008) proposed a model combining discrete event systems and multi agent systems (mas) to simulate a real time job shop. all entities of the generic job shop were modeled as autonomous agents namely job agent, machine agent, work-center agent, shop floor agent, controller agent and job releaser agent. all agents pursued their own interest with unique functions. all communications in the mas were realized through message passing. proportion machines busy (work-center), average number in queue (work-center), maximum number in queue (work-center), average daily throughput (shop floor), average time in system (shop floor), average total time in queues (shop floor), maximal size of work in process (shop floor) were selected as performance indicators. results of the case study demonstrated the advantage of distributed data collection and analysis. it was indicated that, the case study also validated the proposed system by statistical analysis and comparison to existing simulation results in similar test case. gonzález et al. (2010) presented a new centralized cooperative strategy based on trajectory methods (tabu/taboo search) for solving dops. the proposed strategy was compared with two methods namely a multi-qpso and agents. the multi-qpso is a particle swarm optimization (pso) variant with multiple swarms and different types of particles where there exists an implicit cooperation within each swarm and competition among different swarms. on the other hand, agents are an explicit decentralized cooperation scheme where multiple agents cooperate to improve a grid of solutions. researchers tried to assess the possibilities of trajectory methods in the context of dops and to draw attention on explicitly including cooperation schemes in methods. a set of solver hreads were consisted in the proposed strategy. each solver could implement the same of a different resolution algorithm for the problem at hand. the coordinator was responsible for processing the information received from the solver and producing subsequent adjustments of their behavior by sending "orders". exchange of data was achieved by using a blackboard model, with two blackboards. one of them was written by the solvers that wrote the reports of their performances and read by the coordinator; and another was written by the coordinator that wrote the orders and read by the solvers. the information flow in the proposed strategy was achieved by using the following three steps: firstly, performance information was sent from the solvers to the coordinator, and then this information was processed and stored by the coordinator and finally, the coordinator sent directives to the solvers. a set of rule, based on reactive search ideas, was employed to control the solvers. in order to test compare the performance of the proposed strategy, moving peaks benchmark problem and three commonly used multimodal real test functions were selected. to emphasize the importance of dynamism and optimal tracking, and to reduce the number of variables to control in the experiments, only the position of the peaks was altered. proposed strategy and the other two methods were compared according to offline error. just before a change, offline error of each algorithm was recorded and this value was averaged over all changes, for all independent runs. therefore, mean fitness error was calculated. researchers indicated that the proposed strategy could consistently outperform the results of the two other methods. they concluded that the cooperation included in agents provided some benefits over multi-qpso on the easier problems, but since the optimization done in agents relied only on using simple random perturbations of solutions it may be enough to cope with more difficult problems, even with the help of the cooperation. lepagnot et al. (2010) presented a new method called multi agent dynamic optimization (mado) to solve dynamic optimization problems. in mado, a population of agent has explored the search space. three modules namely, memory module, agent manager and coordinator were employed. the number of agents in the system may have varied temporarily, but the number of agents along the whole search process tended to be equal to the predetermined value. this could be achieved by the coordination who would send a delete instruction if the number of agents was higher than a predetermined value. to test the performance of the mado, moving peaks benchmark problem was employed. offline error and standard deviation (sd) were selected as performance measure. different variants of mado were compared with the proposed one. results indicated that mado was better than all the simpler variants. researchers concluded that the proposed mado was a promising method for solving dynamic optimization problems. yan et al. (2010) proposed an agent based evolutionary search (aes) search method. in aes, a population of agents has represented potential solutions. similar to eas, aes gradually converged in the search space during the run, especially when the environment has been stationary for some time. in order to improve the performance of aes for dops, two diversity maintaining mechanisms, random immigrants and adaptive dual mapping were employed. dynamic 0-1 optimization problems which were generated from static optimization problems by using xor generator were investigated. nine different dynamic test problems were constructed. the environment was periodically changed every predetermined number of generations. different change severities were employed to test the performance of aes. the performance of aes was compared with traditional sga, the primal dual ga and the ga with random immigrant, where the worst 10% individuals were replaced with random individuals every generation. mean best of generation fitness was selected as performance measurements. according to the mean best of generation fitness, researchers indicated that aes could always outperform other eas on almost all problems. researchers stated that the competitive and learning behaviors of agents could always help aes to obtain a better performance than the peer gas could do. they concluded that some dynamic characteristics of the environments may affect the performance of the algorithm. hanna and cagan (2009) presented an evolutionary multi-agent system (emas) for adaptive optimization. emas which has employed the evolution of design strategies within a cooperative virtual team has evolved as conditions change and as new solution states were discovered during the optimization process. a set of strategies for creating solutions were represented by population of agents. in emas, the strategies for generating solutions have been recombined, altered, and removed by applying genetic operators that are in typical genetic and evolutionary algorithms. however, researchers have made emas different from genetic and evolutionary programming techniques by adding cooperation dimension. cooperation was employed to combat the problem of not knowing which strategy to use in an unknown but static design space. it has been provided by embodying each strategy in an autonomous agent and allowing the population of agents to communicate. researchers used the well known combinatorial optimization problem of travelling salesman. researchers tried to illustrate that the proposed framework was capable of increasing the effectiveness of individual solution strategies by evolving and coordination them a decentralized manner. three simple heuristic construction algorithms called nearest insertion, farthest insertion and arbitrary insertion were employed. beside these, three heuristics based algorithms, 2-opt, 3-opt and simple mutation were also used. two basic reduction algorithms, random reduction and best partial reduction were employed. different from the typical genetic algorithms, gene strings represented agent strategies for generating solutions, not the solutions themselves. the design architecture used in this study was similar to asynchronous team architecture. in this study fitness was based on the ability of the agent to make positive changes in its destination memory. first of all researchers selected a 48-city problem to compare the resulting tours generated by emas with those generated by both individual algorithms on their own as well as priori determined hybrid algorithms. when the results of emas were compared with the results from running construction algorithms from each starting city and then running improvement algorithms on the resultant tour, researchers found that trials that correlated more strongly with the average patterns had better final values in terms of distance from the optimal value of the average solution quality. in addition to the 532-city problem and 48-city problem, researchers applied these patterns to a team solving a euclidean 237-city problem modeled on a very large scale integration layout problem. researchers has seen that the cooperative teams of individual strategies evolved to generate better solutions than both individual strategies alone and a priori set hybrid strategies. it was indicated that the strength of the emas algorithm lies in its ability to evolve the best team of agent dynamically. it was also stated that, one of the strength of the emas algorithm was as a predictive or learning guide for which set of algorithms or strategies should or should not employed and when. researchers concluded that, utilizing emas in the proposed way has been shown to lower computation time while maintaining or even improving solution quality. pelta et al. (2009a,b) investigated if the type of rules employed previously in cooperative systems for static optimization problems have had sense when applied to dops. the proposed strategy was based on the joint use of a population of solutions and optimizers (agents). researchers analyzed the roles that diversity and decentralized cooperation mechanisms in the performance of the methods. researchers aimed to propose and compare two kinds of control rules to update the solution's set. these rules were a simple frequency based re-sampling (probabilistic) rule and a fuzzy-set based rule. researchers also tried to understand the behavior of both rules in order to develop more efficient cooperative strategies for dops. in this study, cooperation was understood as a mechanism for information sharing. the population of solutions had two purposes which were to serve as an implicit memory that has evolved by means of the action of agents and to be a communication channel for them. an explicit cooperation mechanism was proposed. this mechanism, which was not always triggered, was based on a simple idea. in order to test the performance of the proposed strategy, a set of experiments were developed by using moving peaks benchmark problem. the main aim of the experiments was to evaluate the behavior of both rules which may trigger the explicit cooperation mechanisms. offline error was used as performance measurement. the performance of the cooperative system for different setting of each rule and the system's dynamic behavior were also investigated. the results of the proposed strategy were compared with some published results. researchers stated that the fuzzy-based rule has been better than the frequency rule. as a conclusion, researchers indicated that both rules have been competitive when compared with a state-of-the art of the algorithm. xiang and lee (2008) proposed an agent-based dynamic scheduling approach which employs ant colony intelligence (aci) with local agent coordination. the goal of their study was to represent a dynamic manufacturing system through an mas. they also used aci to improve the global performance of the system. in the proposed system, entities were modeled as intelligent agents with related knowledge of their own functions and goals. mas was used to provide parallel execution of commands. beside this, mas had the intelligence of negotiation to enhance system performance. the agent coordination mechanism used in the study was inspired by both foraging and division of labor of ant colony in mas. there were five types of agents in the proposed mas. they were order agents, job agents, machine agents, work center agents and shop floor agents. researchers indicated that, different from the previous studies, a more generic manufacturing model with less unrealistic assumptions was considered. furthermore, aci was integrated with both machine agents and job agents to solve both task allocation and task scheduling problems. two types of disturbances were introduced. one of them was resource related disturbance including machine break down and machine recovery. the impact of integrating aci in agent coordination was investigated by simulating a realistic shop as a multi-agent manufacturing system. in this disturbance, unreality was expressed in terms of mean time between failure and the mean time to repair. another was source related disturbance including new order\job arrival and existing order\job cancellation. to types of agent coordination, namely coordination based on aci (mas + aci) and coordination using fifo dispatching rule (mas + fifo) were compared. mean flow time, mean tardiness, throughput, buffer size and machine utilization was employed to measure the performance of the proposed approach. results showed that a mas + aci reduced buffer size, max queue number, mean flow time and tardiness. researchers concluded that a mas + aci were outperformed mas + fifo. wang and usher (2002) presented an agent-based job shop model which employed the contract-net protocol as an agent's negotiation mechanism. in this study, two types of agents, named as job agents and machine cell agents, were used with pure hierarchical control structure. researchers considered routing flexibility to provide more options for job agents. average flow time and average queue time were selected as performance measurement. in order to measure the impact of the collaborative factor that was incorporated into the contract-net protocol on the performance, a job shop with five different loading levels were simulated. according to the findings, the collaborative factor did not have much effect on mean flow time when the system load is light, but a significant decrease was observed when the system was heavily loaded. when they examined the average queue time for jobs at each machine cell, they observed that the negotiation scheme with the proposed factor reduced the wip levels of the bottleneck machine cells when the system was under heavy load. based on the simulation results, researchers concluded that the collaborative factor could improve the performance of the contract net-based negotiation scheme in agent-based scheduling problems. wang et al. (2008) proposed a multi-agent approach to study the dynamic scheduling problem in a flexible manufacturing system (fms) shop floor. the proposed approach was combined with a filtered beam search (fbs). researchers aimed to show the feasibility of the proposed approach and to evaluate the approach via computational experiments. dynamisms in this system were provided by new job arrivals. minimization of a weighted quadratic tardiness function was selected as the objective function. there were two types of agents, a system optimal agent (soa) and several cell coordinated agents (ccas). cooperation and coordination among distributed ccas and soa were employed to realize the scheduling goal. five modules which were called as communication, cooperation and coordination, fbs-based algorithm for decision making, execution and monitoring, human interface, were used. in addition to these modules, one local knowledge base and one capacity database were also included. fipa cnip-based negotiation protocol was selected. fbs was performed by filtering phase and beam selection phase. a prototype system was built to show the practicability of the proposed approach. the performance of the proposed scheduling scheme on the prototype system was compared to two dispatching rule combinations. two dispatching rules were used for cell selection, named as dispatching by objective function value and dispatching by make-span, one dispatching rule for routing assignment, named as modified shortest processing time in the selected cell and one dispatching rule for determination of starting time of an operation on the selected machine. when the results on the performance of the average number of jobs tardy were investigated, researchers indicated that the quality of the global schedules generated with the proposed scheduling scheme was better than those of two dispatching-rue combinations. in addition, researchers sought the time required for a complete process of scheduling negotiation. they concluded that the proposed scheduling scheme was promising for real world implementation in multiple manufacturing cells of size. this paper proposes a classification scheme (abdopps: agent based dynamic optimization problem solution strategy) for agent-based approaches which are employed for solving dynamic optimization problems (dops). in this paper 18 typical articles providing agent based solutions to the dops are scanned through the literature and represented using the abdopps. the abdopps is expected to be beneficial to researchers in many ways. similarities of the features located in abdopps can be used to define classes of solution strategies by their descriptions. in this regard, classes of the problems may orient researchers to focus on certain strategies. using the dynamism related features of the corresponding dops presented in abdopps, unpredictability levels of certain problems can be determined and be used to reclassify problems. these representation forms can also be used to discover the role of presented features and their importance for solution quality. online adaptation in learning classifier systems: stream data mining. illinois genetic algorithms laboratory managing critical success factors in a b2b setting evolutionary optimization on problems subject to changes of variables solving vehicle deployment planning problem by using agent based simulation modeling dynamic optimization in a dynamic and unpredictable world search for optimum in dynamic environment: an efficient agent-based method robust, flexible multi-agent optimization using sbdo dynamic optimization in future cellular networks adaptation in dynamic environments: a case study in mission planning a taxonomy of evolutionary algorithms in combinatorial optimization replanning mechanism for deliberative agents in dynamic changing environments optimization in dynamic environments: a survey on problems, methods and measures models of decision and optimization (modo) research group web computational logics and agents: a road map of current technologies and future trends agents in object-oriented software engineering a software architecture-based taxonomy of agent-oriented programming languages software agents communication of the acm a cooperative strategy for solving dynamic optimization problems evolving dynamic multi-objective optimization problems with objective replacement multi-agent coordination and control using stigmergy evolutionary multi-agent systems: an adaptive and dynamic approach to optimization an advanced island based ga for optimization problems automated negotiation: prospects, methods and challenges real time multi-agent decision making by simulated annealing a tutorial on evolutionary computation in dynamic and uncertain environments agent-based infrastructure and an application to internet information gathering a survey of security issue in multi-agent systems market-based approaches to optimization probability collectives: a multi-agent approach for solving combinatorial optimization problems a new multi-agent algorithm for dynamic continuous optimization a multi-agent-based hybrid framework for international marketing planning under uncertainty a multi-agent particle swarm optimization framework with applications evolutionary swarm cooperative optimization in dynamic environments issues and approaches in the design of collective autonomous agents can agents measure up? a comparative study of an agentbased and on-line optimization approach for a drayage problem with uncertainty enabling technology for knowledge sharing incremental and comprehensive strategic information systems planning in an uncertain environment embedding intelligent decision making within complex dynamic environments social mental shaping: modeling the impact of sociality on the mental states of autonomous agents go to the ant": engineering principles from natural multi-agent systems a study on diversity and cooperation in a multi-agent strategy for dynamic optimization problems simple control rules in a cooperative system for dynamic optimization problems multi-agent based simulations using fast multipole method: application to large scale simulations of flocking dynamical systems speculative computation by abduction under incomplete communication environments multi-agent reinforcement learning: independent vs. cooperative agents wireless-based dynamic optimization for load makeup using auction mechanism integration between business planning and information systems planning: an evolutionary-contingency perspective agent-based distributed resource allocation in continuous dynamic systems, intechopen. multi-agent systems agent-based problem solving: the ant colonies metaphor an agent-based evolutionary search for dynamic travelling salesman problem an agent-based evolutionary search for dynamic travelling salesman problem fbs-enhanced agent-based dynamic scheduling in fms an agent-based approach for flexible routing in dynamic job shop scheduling ant colony intelligence in multi-agent dynamic manufacturing scheduling agent based evolutionary dynamic optimization explicit memory schemes for evolutionary algorithms in dynamic environments simulating the generic job shop as a multi-agent system agent-based simulation of electricity markets: a survey of tools key: cord-007367-e31zhty6 authors: tassier, troy; polgreen, philip; segre, alberto title: network position and health care worker infections date: 2015-09-07 journal: j econ interact coord doi: 10.1007/s11403-015-0166-4 sha: doc_id: 7367 cord_uid: e31zhty6 we use a newly collected data set coupled with an agent-based model to study the spread of infectious disease in hospitals. we estimate the average and marginal infections created by various worker groups in a hospital as a function of their network position in order to identify groups most crucial in a hospital-based epidemic. surprisingly, we find that many groups with primary patient care responsibilities play a small role in spreading an infectious disease within our hospital data set. we also demonstrate that the effect of different network positions can be as important as the effect of different transmission rates for some categories of workers. vaccines are a primary way to stop or slow the spread of many infectious diseases, perhaps most notably, influenza. the lack of appropriate vaccination levels is a major health problem. for instance, influenza is a major cause of morbidity and mortality throughout the world despite the availability of a highly effective and inexpensive vaccine. in the us alone, influenza causes an estimated 36,000 deaths and 120,000 hospitalizations annually yet only around 1/3 of healthcare workers are vaccinated each year (thompson et al. 2003) . efficient provision of vaccinations poses a difficult problem in that the positive externality associated with a vaccination is the product of the probability of infection, the cost of the infection, and the marginal infections generated by an agent if infected (all of which may vary across agents). there is great concern over the spread of infectious diseases in hospitals, but little knowledge is available to identify healthcare workers who are most likely to acquire and transmit infectious diseases in hospitals. the problem is especially difficult because the transmission of many infectious diseases is not observable. for instance if someone in your household acquires influenza, you likely do not know which of the potentially hundreds of people you come in contact with each day that may have caused the infection. further if a vaccine is available for an infectious disease and the vaccine is in short supply or is expensive, it is imperative to know which individuals should have the highest priority in vaccine campaigns. in this paper we use a newly collected data set on hospital worker contacts in order to identify hospital worker groups that have the potential to create the largest number of infections based on their location in a hospital contact network. to achieve this goal, we have collected person-to-person contact information on 140 individuals belonging to one of 15 types of healthcare workers at the university of iowa hospitals and clinics (uihc). the data contain information on the contacts between healthcare workers and patients and between healthcare workers and other healthcare workers at the hospital. with this information we develop a network model describing the spread of an infectious disease in a hospital. we estimate, using an agent-based model, the effect of network position of different hospital worker groups on the spread of infectious diseases in a hospital. through this model we are able to identify the hospital worker groups that create the largest externality if removed from the network (perhaps through a vaccination or a quarantine). we argue that methods such as those used in this paper can help hospitals, health care professionals, and epidemiologists to design efficient programs for healthcare worker vaccinations. of importance, we note that we only study the externality in terms of network position within the hospital. in this paper we do not consider other potential heterogeneity among agents such as differences in transmission rates across workers, or differences in behavior outside the hospital that may play a role. 1 while these effects are also important, the large differences across classes of workers shown below are worthy of independent study. this is the first paper to use specific micro-level contact data within a hospital that can be used to help guide policy makers and public health officials in the problem of efficiently allocating vaccines within hospitals. the data used in this paper is unique and detailed in comparison to other studies. the data consists of shadow data (where a research assistant follows a specific, randomly chosen hospital worker for an entire shift) for the 15 major groups of healthcare workers at the uihc, a 700-bed major medical center. this results in over 600 h of direct hospital worker observations and the notation of over 6500 specific worker to worker or worker to patient contacts throughout the hospital. to the best of our knowledge, the data that we have collected comprises the most detailed micro-level healthcare worker contact data set in existence. as a comparison, ueno and masuda (2009) collect data on contacts between doctors, nurses, and patients. their data is based on two calendar days from a small, 129 room, community hospital in tokyo. they model contacts between nurses, physicians and patients. their data does not consider contacts with and between other healthcare worker groups (other than nurses and doctors). based on our data at the uihc, these assumptions would ignore over 60 % of hospital staff, including most of the groups we identify as most crucial to the spread of an infection disease. we begin by discussing the background and motivation for our study and then move to develop a simple model of infectious disease transmission. in the model, we initially assume homogeneous contacts as in traditional epidemiological models. we then discuss a similar model with heterogeneous contacts and discuss the difficulties of achieving efficient vaccination levels. following the theoretical discussion, we use our newly collected data on healthcare worker and patient contacts to model the spread of an infectious disease in a hospital setting. the model allows us to identify the healthcare worker groups that would be expected to play the largest role in the spread of infectious diseases, in terms of network position, in this hospital setting. traditionally, epidemiology research has focused on well-mixed (randomly mixed) populations where agent contacts are homogeneous. in these models, every agent in a population may "bump into" any other agent with equal probability, much as a gas molecule may bump into any other gas molecule with an equal probability over a fixed time period. only recently have epidemiologists and other researchers begun to study the heterogeneous contact structures between people over which infectious diseases spread (early studies include comins et al. 1992; grenfell and harwood 1997; wallinga et al. 1999) . we focus this study on healthcare workers and a particular class of infectious disease, of which influenza is an example. healthcare workers are at especially high risk of contracting influenza. one study of healthcare workers with a low rate of influenza vaccination demonstrated that 23 % of healthcare workers had evidence of influenza infection during a single influenza season (elder et al. 1996) . two features of influenza make its spread difficult to control in hospitals. first, people with influenza are infectious 1-4 days before the onset of symptoms. thus, they can spread the virus when they are still feeling well and are unaware of their own infectious state. second, only about 50 % of infected persons develop classic influenza symptoms (cdc 2002 (cdc , 2003 consequently, restricting healthcare workers with influenza-like symptoms from the workplace will not completely prevent transmission of influenza because healthcare workers with atypical symptoms could continue working and spreading the virus. furthermore, studies show that healthcare workers are more likely than other workers to return to work early or to keep working when they have influenza-related symptoms (weingarten et al. 1989) . because of the ease with which influenza may be contracted and spread by healthcare workers, the centers for disease control and prevention (cdc) have, for the past two decades, recommended influenza vaccination for all healthcare workers. yet, in the us, only 36 % of workers who have direct patient contact are immunized against influenza annually (smith and bresee 2006) . outside of concerns about traditional influenza, there are additional reasons to study the spread of infectious diseases in hospitals. first, healthcare-associated infections affect about 2 million patients in us hospitals each year (jarvis 1996). second, there is a growing fear that hospitals could become a breeding ground for new strains of influenza such as the recent h1n1 influenza outbreak, the potential emergence of person-to-person transmission of avian flu, or other "new viruses." much as sars spread widely in hospitals to begin the sars epidemic in toronto (chowell et al. 2004) , person-to-person transmission of avian flu may start in hospitals as well, and, if a more lethal version of h1n1 were to develop, hospitals again could be a breeding ground for new infections. this last point is of particular salience. with the recent h1n1 outbreak and the subsequent work to develop a vaccine, controversies arose concerning which individuals to vaccinate first. healthcare workers were high on the list. but, as we show below, not all healthcare workers are equal in terms of their importance in spreading infectious diseases. thus, one primary focus of this paper is identifying the individual hospital workers who are most important to vaccinate should a similar outbreak occur again. there is a growing literature in economics on the vaccination choices of individuals and of the externalities associated with vaccinations. but scant attention is paid to the network effects determined by heterogeneous contacts that we focus on in this paper. for example, francis (2004) solves for the optimal tax/subsidy policy for influenza in an sir model with a constant contact rate and random mixing among the population. geoffard and philipson (1997) examine how the individual incentives for vaccination decrease as disease incidence decreases and thereby argue that relying exclusively on private markets is unlikely to lead to disease eradication. boulier et al. (2007) , the most similar paper to ours, investigate the magnitude of the externality associated with a vaccination as a function of the number of vaccinations in the population, the transmission rate of the disease, and the efficacy of the vaccination. they find nonmontonic relationships between each of these items and the magnitude of the vaccine externality. more specifically, the externality and the number of infections prevented per vaccination initially increases before eventually decreasing. however, like francis, they do not consider the case of heterogenous contact number or heterogenous sorting among the population. finally, much of the recent literature on the economics of infectious disease is summarized in philipson et al. (2000) . we begin by describing a simple model where agents in a population have contacts with each other with a uniform probability. this is the traditional random mixing model in epidemiology pioneered by kermack and mckenrick (1927) . the important results in this paper describe exceptions to this homogeneous contact assumption, but we use the simplified model to develop intuition before describing a richer model with heterogeneous contacts. in this simple model, we assume that all agents are homogeneous in that all agents have the same number of contacts with other agents and that these contacts are randomly drawn with uniform probability from the population at large. suppose that agents are assigned to one of three states: susceptible (s), infected (i), or recovered (r). a susceptible agent can transition to being infected with probability α if she is in contact with an infected agent. once infected, an agent transitions to the recovered state according to a parameter κ. once recovered, the agent is immune to the possibility of future infection. this is a classic susceptible-infected-recovered (sir) model for infectious diseases such as influenza (kermack and mckenrick 1927) . the description and parameters yield the following differential equations describing the flows of agents among the various states, assuming a constant population of size n and contact rate of γ . each equation describes the rate of growth for one of the three populations in the sir model. equation 1 describes how susceptible agents contact γ other agents in the population, of which i t /n are infected, and how, of these contacts with infected agents, a percentage α cause the susceptible agent to transition to being infected. equation 2 describes the previously mentioned flows from susceptible into infected and that each infected agent moves to being recovered at rate κ. finally, eq. 3 describes the flows from infected to recovered. we can write these equations in terms of population shares by dividing each of the above equations by the population size n and using lower case letters to denote these population shares, s t = s t /n , i t = i t /n , and r t = r t /n , yielding the following population share equations: the number of infected agents will increase in the population if the flows into the infected state from the susceptible state exceed the flows out of the infected state into the recovered state, di t dt > 0. this condition is equivalent to αγ s t > κ or s t > κ αγ . if this inequality holds then we say that the population is above the epidemic threshold. note that we cannot remain above the epidemic threshold forever without an introduction of new susceptible agents since ds t dt < 0: eventually the population will run out of susceptible agents to infect unless the susceptible population is replenished at a sufficient rate. one goal of healthcare policy is to attempt to place a population below the epidemic threshold so that the number of infectious agents in a population does not grow subject to some cost constraint. a population is most vulnerable to being above the epidemic threshold when the infectious disease first enters a population because s t ≈ 1. this implies that each infected agent infects approximately αγ κ new agents in the population. this fraction is sometimes referred to as the initial reproduction number in the population and is commonly denoted as r 0 ≡ αγ κ . without new individuals entering a population in the susceptible state this reproduction number can only decline as the infectious disease spreads. a successful vaccination moves an agent from state s to state r without incurring the costs of infection. if we reduce the initial population of susceptible individuals, s 0 , by enough we can push the population below the epidemic threshold. specifically, if s 0 < κ αγ then the infectious disease dies out of its own accord without further action. thus an epidemic is prevented whenever s 0 < κ αγ which occurs when (1 − κ αγ )n agents are successfully vaccinated. vaccinating enough agents to produce this effect is called herd immunity (smith 1970) ; once enough people are vaccinated, the entire population (herd) is effectively protected without everyone being vaccinated. the question then becomes, given a cost of vaccination, c(v), what is the efficient level of vaccinations to provide in a population and how do we obtain this efficient level? we begin to approach this question by introducing standard value function notation. in this initial model, once an agent enters state r she remains there forever. thus the value of being in state r is simply the lifetime discounted utility received in state r. we also introduce the possibility of having heterogenous contacts at this stage by indexing agent j's contacts (γ j ) and other terms that we allow to vary across agents. where u j ( ) = utility of agent j from the specified state and β is the discount rate. if an agent is in state i, she will remain in state i for 1/κ periods, on average, until recovered and then enter state r. if an agent is in state s, she receives the same utility as she would if she was recovered, unless she becomes infected. the value to an agent of being in state s is the value of being in state r less the product of the probability that the agent becomes infected and the cost of the infected period. where π(γ j , α, i) is the probability of becoming infected over the course of the epidemic as a function of the contacts of the agent and the transmission rate of the infectious disease. the cost of being infected is the difference in utility between states s and i during the time spent in state i , with the value functions specified we can now specify the vaccination problem for the individual and the social planner. to simplify the vaccination choice of individuals we assume that an agent can only be vaccinated at time period 0. at this time the agent will choose to be vaccinated if the value of being in the recovered state less the cost of the vaccination is greater than the value of being in the susceptible state. thus the agent will choose to be vaccinated if which implies the agent will choose to be vaccinated if c(v) < π(γ j , α, i)c j . the social planner's vaccination problem is more difficult than the individual vaccination problem. essentially, the social planner's problem is to vaccinate agent j if the cost of the vaccination is less than the expected dis-utility of the increase in infections created by agent j if agent j is infected weighted by the probability that agent j is infected. we define the term "marginal infections" of agent j to be the additional infections that occur if j is infected that would not occur if j was not infected. note that this is not simply the number of infections that agent j creates. as an example, agent k maybe infected by agent j. but, if k would have been infected by another agent had she not been infected by j then this would not be a "marginal infection" of agent j. k will be infected regardless of the vaccination choice of agent j. along these lines of thinking, measuring marginal infections is a difficult problem for epidemiologists for at least three reasons: 1. as mentioned earlier, many infectious disease transmissions are not observable. thus it is not easily known how many infections a given agent causes. 2. even if transmissions are observable, the marginal infections created by agent j are not simply the number of other agents that j infects. this is because some agents that j infects may get infected by agents other than j even if j does not infect them herself. further one needs to know how many additional agents are infected by those that j infects and any additional infections created by these agents and so forth. thus one needs to know information on the dynamics of the entire epidemic to measure the true marginal infections of a given agent. 3. the marginal value of vaccinating an agent depends on the behavior and vaccination choices of other agents. it eventually must be decreasing in the number of other vaccinations that are performed. in the extreme, if there are enough vaccinations in the population to produce herd immunity the marginal value of vaccinating an additional agent only involves the probability that the agent is infected from outside the agent population. in effect, the only value is preventing a single agent from infection because she will not, on average, infect anyone else. because of these difficulties we use a simulation approach to help us measure the average and marginal effects of individuals belonging to different worker groups in our hospital contact data. with simulations one can monitor the various infections that occur and also perform controlled experiments to sort out the effects of various groups on potential hospital epidemics. define m j (γ j , α, κ, i, v) as the true marginal infections created by agent j if infected, where v is the number of agents vaccinated in the population. for the majority of the paper we will suppress the notation that does not differ across agents and simply refer to marginal infections as m j (γ j ) since the primary focus of the paper will be on the effect of heterogeneous contacts on the spread of infectious diseases. as shown in boulier et al. marginal infections may be increasing in v for sufficiently small v. 2 but, marginal infections must eventually decrease in v; at the extreme, marginal infections are 0 for any level of v above the point at which herd immunity is reached. thus marginal infections may be increasing or decreasing in the number of vaccinations depending on the specifics of disease transmission, contact patterns, and the number of vaccinations performed. we can now state the social planner's vaccination problem. vaccinate agent j if: note that the individual and social planner's vaccination problems differ by the term π(γ j , α, i)m j (γ j )c j (i). this is the positive externality created by a vaccination when a vaccinated agent j protects other agents which he contacts from acquiring the disease from him. the key terms to investigate in this externality are the probability agent j gets infected, and the marginal infections created by agent j if infected. note that if these marginal infections, m j (γ j ), go to 0, the social planner's problem and the individual problem converge, and the externality is removed. similarly the externality is removed if enough vaccinations are performed to reach herd immunity, again because m j (γ j ) = 0 when herd immunity is achieved because the epidemic never occurs. the main focus of this paper concerns the network position of hospital workers. as such, we assume throughout the paper that the cost of an infection is equal across all agents, c j (i) = c k (i) for all j and k. further, without loss of generality, we can normalize c j (i) = 1. thus the externality above is the product of the probability of infection and the marginal infections produced by the agent if infected. one question then emerges: how are π(γ j ) and m(γ j ) related? at a simple level, if contacts only vary in degree, that is if the only difference in contacts between two agents is the number of contacts and not other, qualitative, aspects of the contacts, then you would expect π(γ j ) and m(γ j ) to be highly correlated. if an agent has a high likelihood of being infected because he has many contacts then he also has many contacts to pass on the infection. suppose that each agent pair is connected with a fixed probability. then, by chance, some agents will have a higher than average number of contacts and some a lower than average number of contacts. in this case, any agent who has a large number of contacts will also generate a large number of secondary infections since there is a lack of structure within the network population. thus any agent with a high probability of infection, π(γ j ) will also be expected to generate a high level of marginal infections m(γ j ). example one is fairly direct. however, various relationships are possible as we show below. in this case each agent in a population is directly connected to every other agent in the population. if this is the case, then any agent that becomes infected is directly tied to all other agents and can infect anyone in the population. thus once someone is infected each agent has a high probability of becoming infected (either from the original agent or from secondary infections). but, since the first agent contacts everyone in the population, and many agents will be infected from him, the other agents in the population may have a low m(γ j ). thus it is possible to have a high probability of infection π(γ j ) and low marginal infections generated m(γ j ) from the same agent. imagine that there are three groups of agents in a population. two of these groups, call them a and b, are separate fully connected graphs containing equal numbers of agents, who do not have any connections to the other group. in other words an agent a ∈ a is connected to every agent a ∈ a but no agent a ∈ a is connected to any agent b ∈ b. suppose that group b is formed in a similar manner. the third group is composed of one agent, j. agent j has only two contacts: one to agent x ∈ a and one to agent y ∈ b. in this example it may be unlikely that agent j gets infected, especially if there is a low transmission rate, because he only has two contacts in the populations. but, if agent j is infected, he may be integral to spreading the disease to the second fully-connected group. suppose an agent in a becomes infected and subsequently infects agent x (or any of the other agents in a) as well as several other agents in a. agents in group b are safe from infection as long as agent j is not infected. but, if j becomes infected, then it is possible that a large fraction of agents in b may become infected as well. thus agent j may have a low probability of being infected, π(γ j ), but create a large number of marginal infections, m(γ j ), if he does become infected. note that each of these three examples offer different implications for public policy approaches to encouraging vaccinations. in the first example, each agent has a probability of being infected that is in line with the number of marginal infections generated. in the other two examples, the infection rate and the number of marginal infections generated may not have a simple relationship with each other. this is an important observation if one considers using subsidies or other means to encourage increased vaccination rates. in the remaining portion of the paper we examine a data set on contacts of and between healthcare workers and patients in a large hospital on the university of iowa campus. we discuss the data and use agent-based models to identify the healthcare workers whose position in the hospital contact network has the potential to create large numbers of infections in the hospital. observational data on contacts between healthcare workers and patients was collected during the winter and early spring of 2006-2007 (the 2006-2007 "flu season") at the university of iowa hospitals and clinics (uihc). the uihc is a 700-bed comprehensive academic medical center and regional referral center in iowa city. data were collected by randomly selecting uihc employees from each of 15 job classifications (specified below) and then using research assistants to "shadow" the 140 selected employees. the research assistants manually recorded every human contact of the subjects (within approximately three feet) over a work shift. note that these observed contacts include anyone contacted within the hospital, not just with the other workers in the shadow sample. this resulted in a total of 6654 recorded contacts over 640 h of observation. additionally, the ra recorded the worker or patient group category for each observed contact (patient or category of healthcare worker) in our data set, and the location in the hospital where the contact occurred. 3 the job categories and number of observed subjects in the data set are as follows: floor nurse (8), food service (11), housekeeper (8), intensive care nurse (8), nurse assistant (10), pharmacist (8), phlebotomist (10), physical/occupational therapist (9), resident/fellow/ medical student (8), respiratory therapist (11), social worker (8), staff physician (11), transporter (7), unit clerk (9), and x-ray technician (14). the data for each group contain approximately 40 h of shadowing. the data were summarized into tallies of contacts over 30-min intervals and then aggregated into contacts per 8 h shift by the authors. table 1 lists the average number of non-repeated contacts 4 per 8 h that occur between the various worker (and patient) categories. note that we were not allowed to choose patients as subjects in our shadow data directly because of privacy concerns. we were only able to observe patient contacts as a result of shadowing other members of the hospital. thus they do not appear as a row in the table. we use this contact data to model the spread of an infectious disease across the uihc hospital. with this data we create a probabilistic contact network for the hospital worker groups. the network is constructed to match the distribution of worker groups at the university of iowa hospitals and clinics. this totals 5232 employees. the distribution of workers across the categories is given in table 2 . we create a contact network among these agents. in the model, each worker in a given group connects to other workers according to the rates observed in our shadowed subjects given in table 1 . as an example, all floor nurses in the model create 11 contacts to other randomly selected floor nurses on average, 0 contacts to food service workers, etc. we assume that the contacts are symmetric in our model, that is, a contact from a given floor nurse to a given housekeeper is also a contact from the housekeeper to the floor nurse. there are at least two reasons for this assumption. first, if our subject is in close enough proximity to pass on the influenza virus to a second agent, the second agent is also within close enough proximity to pass on the influenza virus to our subject. thus the ability to acquire or to pass on virus is a symmetric relationship. second, the reader may note in the table that the matrix of observed contacts is not symmetric because of randomness in the observation of subjects. for instance one notices that the subject floor nurses were not observed to contact food service workers, but a small number of food service worker to floor nurse contacts were observed. thus by assuming that all contacts that occur in the matrix are undirected, we create a symmetric contact matrix where the total number of contacts from a member of group x to group y (and from group y to group x) is one-half the sum of the observed average contacts from group x to group y and from group y to group x. we create the contacts in a uniform random manner within groups. let ρ i j be the ratio of the average contacts between a member of group i and j (taken from table 1 ) to the total number of group j employees (taken from table 2 ). we then take each pair of employees across each group and create a contact with probability ρ i j . specifically, let agent a be a member of group i and agent b be a member of group j. then the probability that a and b are connected is average number of contacts observed between members of groups i and j and n j is the number of employees belonging to group j. 5 before moving to the computational model, we mention two limitations of our data set. human contact networks frequently have a small number of individuals with a much larger than average number of contacts, perhaps differing by orders of magnitude. these individuals are often called hubs. these hubs have the potential to significantly influence disease transmission because they are highly likely to be infected, and if so, to pass on infections to a large number of individuals. because our sample includes approximately 2.5 % of the hospital worker population, we may be missing hubs in our sample if they exist in this setting. however, we note two things in relation to this: first, the relatively homogeneous workday responsibilities of workers within categories, likely limit the variation of contacts within a category. for instance two physicians or two floor nurses are likely constrained to see a relatively similar number of patients each day. this is unlike many other social network data sets, (such as general friendship or online networks) where there are not these work responsibility constraints on individual contacts. thus the possibility of hubs with orders of magnitude differences in numbers of contacts is more limited in our data context. second, our data set is more comprehensive than any other within hospital contact data set in existence in terms of the worker categories included. recall from earlier in the paper that the ueno and masuda data set only includes physicians, nurses, and patients in a hospital much smaller than is studied here. our results below suggest that many of the most important groups in the hospital are not included in their study. thus, at a minimum, our study highlights the importance of funding for studies that aim to collect even more comprehensive data sets that include individuals with nonpatient care responsibilities and more comprehensively cover a larger share of hospital workers. in addition, we note that we consider all contacts in our data set to be equivalent, or "equally weighted." there may be some concern that not all contacts are created equal in our context. for instance, a contact between a physician and a patient that occurs during a physical examination, may be more likely to spread an infectious disease than other contacts in our data set. we attempt to control for this possibility later in the paper when we consider heterogeneous transmission rates and repeated contacts. as mentioned above, transmissions of infectious disease are not usually observable. thus studying infectious disease transmission using an agent-based model can be a useful tool. in the remainder of the paper we model the spread of an infectious disease across the simulated hospital contact networks described above. once created, we use the contact network in a model of the spread of an infectious disease in the hospital as follows: agents can be in one of three states, susceptible to being infected (s), infected and able to infect others (i), or recovered (and therefore immune) (r). we assume that each infected agent recovers after 5 periods which is in-line with the infectious period for influenza. once recovered the agent enters state r and is therefore immune to further transitions to the infected state. initially, all agents in the model are in state s. agents may be vaccinated against infection. vaccinations occur only in the initial period of the model. once vaccinated, an agent moves immediately from state s to state r and is thus immune for the remainder of the model. in the initial period of the model, each agent in state s (all agents that have not been vaccinated) is subject to infection with probability α 0 = 0.005. these are the agents of our model that seed the potential epidemic. once these initial infections occur we assume that each contact in our network occurs once in each subsequent period of the model. if a contact occurs between a state i and a state s agent, the state s agent transitions to state i with probability α, which we vary across experiments. we continue the model until no agents remain in state i. in each period of the model we calculate the fraction of agents in each worker category in each state (s, i, or r). for each of the results reported below we run 100 replications of each parameter set or computational experiment reported. the results reported are averages over these replications. in addition to the results reported below, we have studied a wide range of parameters for our model and find the results reported below to be robust to changes in all of the parameters within reasonable bounds. 6 the purpose of the model is to estimate m(γ j ) and the externality generated for the network of contacts in the uihc shadow data and, in turn, to identify the classes of workers most important to vaccinate. this is a two step process. first we perform a series of base-line models as described above with none of the healthcare workers and patients vaccinated. from this baseline, we observe the rate of infection for each class of agents in the hospital population (15 worker groups and patients for a total of 16 groups). we denote the infection rate of group k in the base model as a function of the transmission rate α as π k b (α) and the overall infection rate in the entire population as a function of α as π 0 b (α). second, we want to calculate the average and marginal infections generated by each group. to do so, we change the vaccination rate for each group, one group at a time, and re-run the model. as an example, we run the model with all floor nurses vaccinated and no other vaccinations and observe π 0 1 (α). then we run the model with all housekeepers vaccinated and no one else and observe π 0 2 (α) and so on for each group. we then compare the change in the average number of infections between the models, δ(b, k) = (π 0 b (α)−π 0 k (α))n , which is the difference between the overall infection rate in the base model with no vaccinations and the overall infection rate in the model with all of group k vaccinated, multiplied by the total population size n . now, using the notation described above, the change in the number of infections δ(b, k) is equal to the change in number of people vaccinated, n k , multiplied by the probability that each of these agents becomes infected if not vaccinated, multiplied by the number of additional infections each infected agent would generate. simplifying, if we assume each agent infects the same number of individuals, we can write the average number of secondary infections generated per infected person in group k, as a k and write: one can then find an estimate of the average secondary infections per infected person as:â effectively, this process removes each group from the hospital, one at a time. we then can observe the effect of each individual worker group on the size of the modeled epidemic. instead of vaccinating all agents of a group at once, we can vaccinate a fraction of a group. as we change this fraction at intervals n k we can view the effect of increases in vaccination rates for each group (one at a time). we then have an estimate of the marginal infections prevented per vaccination as: we now proceed in two steps. first we investigate the effect of each group in total on the epidemic process by vaccinating an entire group one at a time and calculatinĝ a k for each group. we then choose a sample of interesting groups and study the details of the epidemic process as we vary the number of vaccinations performed in each of these groups, at specified intervals between 0 and 100 %. interestingly, as we vary the percent of each group vaccinated, we will see that there are different outcomes across these different groups in terms of marginal infections generated, probability of infection and the overall effect of a vaccination (in terms of reducing the number of infections). we begin by varying the transmission rate, α, over the range [0.0004, 0.006] and observing the base infection rate π 0 b (α). the results are displayed in fig. 1 . as one can see in the figure, a sufficiently large transmission rate is needed to generate an epidemic of reasonable size. further, as expected, the number of infections generated monotonically increases as a function of the transmission rate, α. our primary interest is in intermediate ranges of epidemic outbreaks. if the transmission rate is too high then almost everyone in a population needs to be vaccinated in order to reach herd immunity. and, if the transmission rate is too low, then there is not a large need to worry about vaccine priority. thus, we concentrate on two intermediate levels of the transmission rate α = 0.0030 and α = 0.0035. with no vaccinations, these levels yield an epidemic where between one-third and one-half of the population is infected over the course of the epidemic. we now find the average effect of vaccinations across the hospital worker groups using the procedure described above for α = 0.0030 and α = 0.0035. we present results for the average "secondary infections" generated per infected person,â k , and the product of average infections and probability of infection, which yields the "decrease in infections per vaccination," δ(b,k) n k in tables 3 and 4 . from the decrease in infections per vaccination we have an indication of how much the vaccination of an individual group member is contributing to preventing the spread of an epidemic. the results of this experiment suggest where efforts should be directed in the event of an influenza vaccine shortage or in the event of the development of new disease for which a vaccine may be developed (e.g., avian flu, swine flu, etc.) but is initially in short supply until mass quantities may be made available. note that some of the groups have vaccinations that prevent less than one infection per vaccination. this occurs because these groups have a low probability of infection and sufficiently low number of average infections that each member generates if infected. groups with large decreases in infections per vaccination are the ones to prioritize in times of a vaccine shortage, assuming equal costs of infection. in these experiments we see three clear groups that stand above the others in terms of the effect of vaccinations. for the parameters of the experiments, each vaccination of a unit clerk, social worker, and phlebotomist, results in a decrease of 3.1 infections or more on average for α = 0.0030 and of 2.2 or more for α = 0.0035. 7 in addition vaccinating unit clerks is extremely effective; each unit clerk vaccination results in a decrease of over 7 infections for α = 0.0030 and over 6 infections for α = 0.0035. somewhat surprisingly, some of the groups that are seen as central to the functioning of a hospital play a very small or moderate role in spreading an infectious disease. vaccinating staff physicians results in a lower than average decrease in infections. we revisit this result in our discussion of transmission rates later in the paper. also of note, as one would expect, as the transmission rate increases, the probability of infection increases. but, this has the effect of making individual vaccinations less beneficial on average. note thatâ k is smaller for each group for a higher transmission rate. this has the effect of lowering the variance of average infections. for the α = 0.0030 case above the standard deviation is 1.79, and for the α = 0.0035 case, the standard deviation is 1.41. as the transmission rate increases, a larger fraction of individuals are infected throughout the population. thus there are more opportunities for each individual to be infected if she has not already been infected. vaccinating a given person in the population will only prevent one of these multiple channels for infection. so, as the infection rate increases, the effectiveness of a vaccination becomes more uniform across the groups. this has direct policy applications. an infectious disease that is highly contagious could best be met with a uniform vaccination strategy since each individual in the population will create a similar level of infections on average. but an infectious disease with a low level of contagiousness could most effectively be met with a targeted vaccination campaign (bansal and pourbohloul 2007) . we next look at the marginal effect of a vaccination as the number of vaccinations increase. we present results in figs. 2, 3, 4 , 5, 6 and 7 for five interesting worker group categories for the same two transmission rates discussed above. unit clerks, social workers and phlebotomists are chosen because of their large number of secondary infections generated. we also choose floor nurses and staff physicians because of interest in the effect of worker groups with primary patient care responsibilities. in fig. 2 we plot the marginal infections prevented per vaccination as a function of the number of vaccinations performed for the five groups. recall that the number of marginal infections is the additional number of infections that an agent generates if the agent becomes infected. in fig. 3 we plot the probability of infection for the five groups. and, in fig. 4 we plot the product of marginal infections and probability of infection which yields the total number of infections prevented per vaccination. these figures all consider a transmission rate of 0.0030. figures 5, 6 , and 7 plot the same relationships for a transmission rate of 0.0035. we begin with marginal infections. recall that marginal infections may be increasing or decreasing in the number of vaccinations performed for small numbers of vaccinations (boulier et al. 2007 ). (here the number of vaccinations performed is small relative to the entire population as we are only vaccinating some members of one of the 15 groups. so, even if we vaccinate an entire group, this is a small number relative to the entire population.) here, we see two interesting outcomes. first, we see that marginal infections for both unit clerks and floor nurses increase as more vaccinations are performed. for these two groups, each additional vaccination prevents a larger and larger number of infections. this is particularly extreme for a transmission rate of 0.0030 and the case of unit clerks where a small number of vaccinations results the product of the two previously discussed statistics yields the number of infections prevented per vaccination. again, unit clerks display a unique relationship in that the number of infections prevented per vaccination dramatically increases in the number of vaccinations performed. the other four groups result in much flatter plots. thus again there is little difference between the marginal and the average for these groups. there are two interesting points to be made from these results. the first is that the effect of vaccinating unit clerks in our data is most important both from a marginal and an average perspective regardless of how many vaccinations have been performed. particularly, the marginal effect of vaccinating a unit clerk increases at a greater rate than the probability of infection for a non-vaccinated unit clerk falls. thus, it is always more beneficial to vaccinate one more unit clerk as opposed to a worker from another group (assuming transmission rates are equal). the second is that there is little difference between the average and the marginal effect of a vaccination for the other groups considered here. this second point can be interpreted as good news from a policy making perspective in the sense that the optimal allocation of vaccinations does not switch as more of a group is vaccinated. in other words it is not the case that group a is the optimal group to target up to some vaccination percentage, after which group b should be targeted. switching such as that would indicate a much more complicated solution to the optimal vaccine allocation problem. here, because there is little difference between the marginal and the average effect of a vaccination for most groups, one can pragmatically target the groups with the largest average effect of a vaccination. we now move to discuss the important features of the contact network that creates the externality. as we will see below, it is not just the number of contacts that an agent has but also which specific agents and groups the agent contacts, as well as who the agent's contacts connect to in turn. we begin by looking at some basic statistics of the contacts in our data in table 5 . for each group, the table displays the total number of contacts, the percentage of total contacts that are with members of an agent's own group, and the number of patient contacts. total contacts and contacts with patients could be directly correlated with the likelihood of being infected and with passing on infections. the percentage of contacts within one's own group can indicate how varied one's network is and how widespread one's connections are. for instance having few contacts within one's own group provides the possibility of introducing an infection to other groups within the hospital. in addition the table also contains a common network characteristic measure, betweenness centrality, which we discuss below. on a network, the geodesic distance, g a,b , between nodes a and b is the length of the shortest path between the two nodes that traverses connections on the network measured as the number of connections traversed (or "hops" required) to reach node b from node a. as one measure of the centrality of a node on a graph one can calculate the average distance to all other nodes on the graph. thus if a graph g is composed of n nodes, average distance for node a is calculated as: a short average distance indicates that a node is close to other nodes on average an thus may be likely to be infected and to pass on infections. thus it is sometimes considered a measure of the centrality or importance of a node in a network. betweenness is another measure of the centrality of a node in a network. let c i jk be the proportion of all geodesics linking node j and node k which pass through node i. let c i be the sum of all c i jk for i = j = k. letc i be the maximum possible value for c i . (normalized) betweenness for node i, b i is then: betweenness for node i is therefore a measure of the proportion of shortest paths between nodes that go through node i. 8 in the the table below we report group level values for betweenness centrality. the values are created in the following manner. first we create a network using the same methods as described above. second, we then calculate the betweenness centrality measurement for each agent in the simulated network. third, we calculate the average value for each hospital worker group and report the result in the table below. this network variable is likely to be an indicator of importance for disease transmission in the network. if a hospital worker group has a high average betweenness value, then the nodes in this group are potentially important in passing infections on to other nodes as it plays a crucial role in location along many of the shortest paths between nodes in a network. as such, this measure should be closely related to the marginal infections that a group generates. what is most interesting in table 5 is the lack of a clear relationship between any of the variables in the first three columns and our previously-listed most important groups (highlighted in bold). the only measure that consistently aligns well is the betweenness measure. for a moment concentrate on the values in the first three columns. each of these three plausibly important characteristics fail to display a meaningful relationship with the average or marginal infections generated. if we concentrate on the top three most important groups, some have relatively large numbers of contacts (unit clerks), although not the largest, while others have contacts significantly below the average (phlebotomists). some have large numbers of patient contacts (phlebotomists) while others have some of the smallest number of patient contacts (unit clerks and social workers). one interesting thing that appears in the table is that phlebotomists have almost all of their contacts with patients and other phlebotomists. thus the network position of phlebotomists, is in some sense, very similar to that of patients. overall, what these observations imply is that there is not likely to be a simple relationship (or a small set of simple relationships) indicating which individuals are most important to vaccinate by looking at easily observed worker interaction patterns. instead the relationship depends on the intricate and complex web of relationships that make up the entire contact network of the hospital. this is what is captured in the betweenness centrality measure. betweenness measures the percentage of shortest paths in the entire network on which an agent is located. if we remove agents with high betweenness measures (by vaccinations) from the disease propagation network, we disrupt the flow of an epidemic. for concreteness, the correlation between the betweeness measure and the average infections generated is about 0.84 for both α = 0.0030 and α = 0.0035. as mentioned above, the primary focus of this paper concerns the effect of network position on the marginal infections generated within a hospital. here, we consider two robustness checks to the results presented above. first, we discuss a comparison of the magnitude of the above described network effects and the effect of transmission rates on marginal infections for an interesting example group, staff physicians. second, we do an additional set of experiments using an additional data set that includes observed repeated contacts. we perform these robustness checks for two main reasons: first, because of the job responsibilities of different worker groups, the different worker groups may have different transmission rates, durations of contacts, or frequency of contacts creating another source of heterogeneity in infections. as a few examples, a staff physician may be more likely to transmit an infection over the course of a patient exam that includes a series of physical contacts, compared to a nurse who has a brief arm's-length conversation with a hospital transporter. a floor nurse may have multiple interactions with the same patient during the course of the day. 9 we begin this analysis by varying the transmission rate of staff physicians. (recall that staff physicians created a lower than average number of infections in our earlier model.) again, this is primarily to account for the fact that physicians may have longer duration contacts with patients and the contacts may more frequently involve physical touch. we now assign a special transmission rate to staff physicians that we denote as α p . this will be the transmission rate for any contact between a physician and another agent where one of the agents is infected. we use α = 0.0030 for all other contacts in the population and vary α p from α p = 0.0030 and α p = 0.0200. as we do this we again measure average infections as described earlier and show the resulting average infections for unit clerks and staff physicians in fig. 8 . before we present the results we note that there are alternative ways to model this scenario. for instance, we could re-weight our contact matrix. if we knew, for instance, that physician to patient contacts lasted twice as long as other pairs of contacts in the hospital worker population, we could re-weight each physician to patient contact by a factor of two. but note that, on average, this is equivalent to increasing the transmission probability by a factor of two because the expected number of new infections is the number of susceptible to infected contacts in a period multiplied by the transmission rate. a doubling of the number of contacts is equivalent to a doubling of the transmission rate. 10 recall from our results above that when the transmission rates are equal the average secondary infections created by unit clerks are slightly greater than 8 and slightly greater than 1.5 for staff physicians. as we increase the transmission rate for staff physicians we see several things: first, as you would expect, the average infections for staff physicians increases. but the change is not large. for α p = 0.0200 the average secondary infections created by staff physicians is 2.86, slightly less than a two-fold increase and still well below the level of average infections noted for unit clerks when the transmission rates are equal. for unit clerks, as α p increases, the average infections of unit clerks drops rapidly. this occurs for at least two reasons: one is that unit clerks become less important relative to staff physicians as α p increases. another is that, as α p increases, the overall infection rates increase, and as we reported earlier, this causes average infections to become more uniform across groups. overall, the level of average infections between these two groups does not become similar until the α p increases to about 0.0150, a five-fold increase in the staff physician transmission rate. and, the average infections of staff physicians does not become greater than that of unit clerks until α p = 0.0200. most interesting about these results is the magnitude of the network effects relative to the magnitude of the transmission rates. in the case of staff physicians and unit clerks it takes a 5-6 times increase in the transmission rate of staff physicians to "make up" the difference in network position. this suggests that the network effect differences in average infections are very important in understanding overall transmission patterns. as a second robustness check we use an additional cut of our data that includes observed repeated contacts. in some instances members of the hospital population were observed to have come in contact more than once during the observation period. we use this data as one way to include the weighting of contacts mentioned above into the analysis. table 6 displays the observed contacts that were observed and had occurred previously in our data. as you see in the table many of these contacts involved repeated interactions with patients (often by members of the nursing staff). we re-perform the analysis above with these additional contacts added into the data. the only additional change is that we modify the transmission rate to α = 0.0020 so that the total number of infections in the population remains nearly constant compared to the non-repeat contact data. (recall from earlier in the paper that a larger epidemic smooths out differences in the population groups and makes the average and marginal values more similar across groups. thus we control for epidemic size by varying the transmission rate.) we present the results in table 7 . you will notice in the table that most of the groups that previously had the largest effect still do. unit clerks are still the most important group to vaccinate, but the difference in magnitude between unit clerks and other important groups is less than in the the previous model. further, as one would expect, groups that have more repeated contacts, such as all types of nurses, become more important. the important point though, is that the relative ranking of most of the groups changes very little. of course this is partially due to there being relatively few repeated contacts in the data set. taking these two robustness checks in combination, they demonstrate two important things. first, network structure is at least as important as transmission rate in determining the course of an epidemic in our data set. second, while the data we have collected is not perfect in terms of comprehensiveness, the relative ranking of group importance appears relatively robust to alternative measures of network structure. a a minimum, taken in combination, these results suggest a need for a greater emphasis on network based data collection in order to better understand both micro level and macro level epidemiology. we now make a small shift in focus. generally, a hospital's primary goal is to restore or improve patient health. thus prioritizing healthcare worker vaccinations so as to best protect patients may be a legitimate goal of a hospital. in other words hospital administrators may care about protecting patients from infection as much, or more, than they do about protecting the entire hospital population from infection. of course these may be closely related goals. in addition, in a large scale epidemic, it may be of great importance to have a healthy staff of physicians to treat patients. thus protecting physicians may be another important goal in vaccine priority within a hospital. in tables 8 and 9 we display the same relationships as displayed in our initial results section above but this time only with regard to patient infections generated and staff physician infections generated, not infections in the entire hospital. in this analysis we see very similar results to the overall population results. beginning with patients, the four groups (unit clerk, social worker, physical and occupational therapist, and phlebotomist) that played the most important role in transmitting to the hospital population as a whole also play an important role in their effect on patients specifically. however, we see some difference in the two groups of models. first, groups that have more direct patient contacts increase in importance. for instance, phlebotomists replace unit clerks as the most important group. second new groups emerge as important for transmitting to patients. for instance, hospital transporters are among the top four groups in transmitting to patients but are significantly below the average in terms of transmissions to the general population. with this in mind it seems that giving vaccination priority to health care workers with direct patient contacts is more important for protecting patients than it is for protecting the general population, as one would expect. but still, some of the groups with the largest impact on infecting patients have few direct patient contacts (unit clerks and social workers, for example). for staff physicians we see similar results. again, the four most important groups remain important, but other groups such as floor nurses increase in importance when considering staff physicians specifically. to summarize the results of this section, the same groups that create infections in the general population also create infections in the patient and staff physician population. but, groups that have direct patient or staff physician contacts have increased importance. still, one should not ignore other groups central to the network of the hospital that have only few direct patient or staff physician contacts (e.g., social workers and unit clerks). we utilize a newly collected data set on contacts of health care workers at a large university hospital to estimate network effects for infectious disease transmission. interestingly the most important groups to vaccinate tend to have heterogeneous contacts throughout the hospital. groups such as social workers and unit clerks are very important to vaccinate even though they have been given low priority in past vaccine campaigns because of their relatively limited number of patient contacts. for instance, the cdc recommended in their interim influenza vaccination recommendations in 2004-2005 that influenza vaccine priority be given to "health-care workers involved in direct patient care" and further stated, "persons who are not included in one of the priority groups described above should be informed about the urgent vaccine supply situation and asked to forego or defer vaccination." this mismatch of scientific results and past policy decisions suggests that future research in this area is warranted especially when one considers the public health dangers associated with the emergence of avian flu, a more lethal version of swine flu, or recent dangers such as sars. with that stated, we want to be careful to recognize that one reason to vaccinate primary care providers is to assure individuals are available to care for the sick. this important incentive is outside the scope of our model. the results of this paper lead to important public policy considerations. specifically, hospital workers with a low probability of infection may be likely to ignore recommendations for vaccination even if they are central to the spread of an infectious disease. one way to increase the overall vaccination level is with a subsidy program. but, as the results in this paper show, not all hospital workers are equal in terms of the positive externality generated by a vaccination. because of the heterogeneous contacts throughout the hospital, some workers are more important to the spread of an infectious disease than others. thus if hospitals and other public health organizations want to efficiently distribute vaccines they need to target specific worker groups, perhaps by allocating subsidies, on the basis of discrepancies in probability of infection and marginal infections generated. this paper is the first to use specific micro-level contact data within a hospital to guide policy makers and public health officials in this endeavor. to be clear, these results are not meant to be specifically calibrated to measure the exact effect of vaccinations in these groups. instead our hope is that the orderings of the hospital worker groups (which are robust across the parameters that we have explored) indicate where public health officials can effectively intervene in order to prevent widespread epidemics within hospitals. and these experiments reveal interesting and surprising groupings. prior to this study, as quoted above, it had been argued that groups like unit clerks be excluded from influenza vaccine campaigns, in times of vaccine shortages, because of their minimal patient contacts. the results of this study suggest that decisions such as these need to be more fully explored. a comparative analysis of influenza vaccination programs article 23 cdc (2002) epidemiology and prevention of vaccine-preventable diseases, 7th edn. public health foundation prevention and control of influenza: recommendations of the advisory committee on immunization practice (acip) model parameters and outbreak control for sars the spatial dynamics of host parasitoid systems incidence and recall of influenza in a cohort of glasgow healthcare workers during the 1993-94 epidemic: results of serum testing and questionnaire optimal tax/subsidy combinations for flu season disease eradication: private versus public vaccination meta)population dynamics of infectious diseases selected aspects of the socioeconomic impact of nosocomial infections: morbidity mortality, cost, and prevention a contibution to the mathematical theory of epidemics epidemiology economic, diseases infectious prospects of the control of disease prevention and control of influenza: recommendations of the advisory committee on immunization practices, morbidity and mortality weekly report 55 mortality associated with influenza and respiratory syncytial virus in the united states controlling nosocomial infection based on the structure of hospital social networks perspective: human contact patter ns and the spread of airborne infectious diseases barriers to influenza vaccination acceptance. a survey of physicians and nurses key: cord-355024-v5lahyw4 authors: van seventer, jean maguire; hochberg, natasha s. title: principles of infectious diseases: transmission, diagnosis, prevention, and control date: 2016-10-24 journal: international encyclopedia of public health doi: 10.1016/b978-0-12-803678-5.00516-6 sha: doc_id: 355024 cord_uid: v5lahyw4 infectious disease control and prevention relies on a thorough understanding of the factors determining transmission. this article summarizes the fundamental principles of infectious disease transmission while highlighting many of the agent, host, and environmental determinants of these diseases that are of particular import to public health professionals. basic principles of infectious disease diagnosis, control, and prevention are also reviewed. an infectious disease can be defined as an illness due to a pathogen or its toxic product, which arises through transmission from an infected person, an infected animal, or a contaminated inanimate object to a susceptible host. infectious diseases are responsible for an immense global burden of disease that impacts public health systems and economies worldwide, disproportionately affecting vulnerable populations. in 2013, infectious diseases resulted in over 45 million years lost due to disability and over 9 million deaths (naghavi et al., 2015) . lower respiratory tract infections, diarrheal diseases, hiv/ aids, malaria, and tuberculosis (tb) are among the top causes of overall global mortality (vos et al., 2015) . infectious diseases also include emerging infectious diseases; diseases that have newly appeared (e.g., middle east respiratory syndrome) or have existed but are rapidly increasing in incidence or geographic range (e.g., extensively drug-resistant tuberculosis (xdr tb) and zika virus (morse, 1995) . infectious disease control and prevention relies on a thorough understanding of the factors determining transmission. this article summarizes some of the fundamental principles of infectious disease transmission while highlighting many of the agent, host, and environmental determinants of these diseases that are of particular import to public health professionals. a classic model of infectious disease causation, the epidemiological triad (snieszko, 1974) , envisions that an infectious disease results from a combination of agent (pathogen), host, and environmental factors ( figure 1 ). infectious agents may be living parasites (helminths or protozoa), fungi, or bacteria, or nonliving viruses or prions. environmental factors determine if a host will become exposed to one of these agents, and subsequent interactions between the agent and host will determine the exposure outcome. agent and host interactions occur in a cascade of stages that include infection, disease, and recovery or death (figure 2(a) ). following exposure, the first step is often colonization, the adherence and initial multiplication of a disease agent at a portal of entry such as the skin or the mucous membranes of the respiratory, digestive, or urogenital tract. colonization, for example, with methicillin-resistant staphylococcus aureus in the nasal mucosa, does not cause disease in itself. for disease to occur, a pathogen must infect (invade and establish within) host tissues. infection will always cause some disruption within a host, but it does not always result in disease. disease indicates a level of disruption and damage to a host that results in subjective symptoms and objective signs of illness. for example, latent tb infection is only infectionevidenced by a positive tuberculin skin test or interferon gamma release assaybut with a lack of symptoms (e.g., cough or night sweats) or signs (e.g., rales on auscultation of the chest) of disease. this is in contrast to active pulmonary tb (disease), which is accompanied by disease symptoms and signs. recovery from infection can be either complete (elimination of the agent) or incomplete. incomplete recovery can result in both chronic infections and latent infections. chronic infections are characterized by the continued detectable presence of an infectious agent. in contrast, latent infections are distinguished by an agent which can remain quiescent in host cells and can later undergo reactivation. for example, varicella zoster virus, the agent causing chicken pox, may reactivate many years after a primary infection to cause shingles. from a public health standpoint, latent infections are significant in that they represent silent reservoirs of infectious agent for future transmission. when a potential host is exposed to an infectious agent, the outcome of that exposure is dependent upon the dynamic relationship between agent determinants of infectivity, pathogenicity, and virulence, and intrinsic host determinants of susceptibility to infection and to disease (figure 2(b) ). environmental factors, both physical and social behavioral, are extrinsic determinants of host vulnerability to exposure. environment disease figure 1 the epidemiological triad model of infectious disease causation. the triad consists of an agent (pathogen), a susceptible host, and an environment (physical, social, behavioral, cultural, political, and economic factors) that brings the agent and host together, causing infection and disease to occur in the host. infectivity is the likelihood that an agent will infect a host, given that the host is exposed to the agent. pathogenicity refers to the ability of an agent to cause disease, given infection, and virulence is the likelihood of causing severe disease among those with disease. virulence reflects structural and/or biochemical properties of an infectious agent. notably, the virulence of some infectious agents is due to the production of toxins (endotoxins and/or exotoxins) such as the cholera toxin that induces a profuse watery diarrhea. some exotoxins cause disease independent of infection, as for example, the staphylococcal enterotoxins that can cause foodborne diseases. agent characteristics can be measured in various ways. infectivity is often quantified in terms of the infectious dose 50 (id 50 ), the amount of agent required to infect 50% of a specified host population. id 50 varies widely, from 10 organisms for shigella dysenteriae to 10 6 -10 11 for vibrio cholerae (gama et al., 2012; fda, 2012) . infectivity and pathogenicity can be measured by the attack rate, the number of exposed individuals who develop disease (as it may be difficult to determine if someone has been infected if they do not have outward manifestations of disease). virulence is often measured by the case fatality rate or proportion of diseased individuals who die from the disease. the outcome of exposure to an infectious agent depends, in part, upon multiple host factors that determine individual susceptibility to infection and disease. susceptibility refers to the ability of an exposed individual (or group of individuals) to resist infection or limit disease as a result of their biological makeup. factors influencing susceptibility include both innate, genetic factors and acquired factors such as the specific immunity that develops following exposure or vaccination. the malaria resistance afforded carriers of the sickle cell trait exemplifies how genetics can influence susceptibility to infectious disease (aidoo et al., 2002) . susceptibility is also affected by extremes of age, stress, pregnancy, nutritional status, and underlying diseases. these latter factors can impact immunity to infection, as illustrated by immunologically naïve infant populations, aging populations experiencing immune senescence, and immunocompromised hiv/aids patients. mechanical and chemical surface barriers such as the skin, the flushing action of tears, and the trapping action of mucus are the first host obstacles to infection. for example, wound infection and secondary sepsis are serious complications of severe burns which remove the skin barrier to microbial entry. lysozyme, secreted in saliva, tears, milk, sweat, and mucus, and gastric acid have bactericidal properties, and vaginal acid is microbicidal for many agents of sexually transmitted infections (stis). microbiome-resident bacteria (a.k.a. commensal bacteria, normal flora) can also confer host protection by using available nutrients and space to prevent pathogenic bacteria from taking up residence. the innate and adaptive immune responses are critical components of the host response to infectious agents ( table 1) . each of these responses is carried out by cells of a distinct hematopoietic stem cell lineage: the myeloid lineage gives rise to innate immune cells (e.g., neutrophils, macrophages, dendritic cells) and the lymphoid lineage gives rise to adaptive immune cells (e.g., t cells, b cells). the innate immune response is an immediate, nonspecific response to broad groups of pathogens. by contrast, the adaptive immune response is initially generated over a period of 3-4 days, it recognizes specific pathogens, and it consists of two main branches: (1) t cell-mediated immunity (a.k.a. cell-mediated immunity) and (2) b cellmediated immunity (a.k.a. humoral or antibody-mediated immunity). the innate and adaptive responses also differ in table 1 comparison of innate and adaptive immunity innate immune response adaptive immune response immediate response; initiated within seconds gradual response; initially generated over 3-4 days (primary response) targets groups of pathogens targets-specific pathogens no memory memory progression from one stage to the next is dependent upon both agent properties of infectivity, pathogenicity, and virulence, and host susceptibility to infection and disease, which is in large part due to both protective and adverse effects of the host immune response. credit: modification of original by barbara mahon, md, mph. that the latter has memory, whereas the former does not. as a consequence of adaptive immune memory, if an infectious agent makes a second attempt to infect a host, pathogenspecific memory t cells, memory b cells, and antibodies will mount a secondary immune response that is much more rapid and intense than the initial, primary response and, thus, better able to inhibit infection and disease. immune memory is the basis for the use of vaccines that are given in an attempt to stimulate an individual's adaptive immune system to generate pathogen-specific immune memory. of note, in some cases the response of the immune system to an infectious agent can contribute to disease progress. for example, immunopathology is thought to be responsible for the severe acute disease that can occur following infection with a dengue virus that is serotypically distinct from that causing initial dengue infection (screaton et al., 2015) . an immune host is someone protected against a specific pathogen (because of previous infection or vaccination) such that subsequent infection will not take place or, if infection does occur, the severity of disease is diminished. the duration and efficacy of immunity following immunization by natural infection or vaccination varies depending upon the infecting agent, quality of the vaccine, type of vaccine (i.e., live or inactivated virus, subunit, etc.), and ability of the host to generate an immune response. for example, a single yellow fever vaccination appears to confer lifelong immunity, whereas immune protection against tetanus requires repeat vaccination every 10 years (staples et al., 2015; broder et al., 2006) . in malariaendemic areas, natural immunity to malaria usually develops by 5 years of age and, while protective from severe disease and death, it is incomplete and short-lived (langhorne et al., 2008) . functionally, there are two basic types of immunization, active and passive. active immunization refers to the generation of immune protection by a host's own immune response. in contrast, passive immunization is conferred by transfer of immune effectors, most commonly antibody (a.k.a. immunoglobulin, antisera), from a donor animal or human. for example, after exposure to a dog bite, an individual who seeks medical care will receive both active and passive postexposure immune prophylaxis consisting of rabies vaccine (to induce the host immune response) and rabies immune globulin (to provide immediate passive protection against rabies). an example of natural passive immunization is the transfer of immunity from mother to infant during breastfeeding. vaccination does not always result in active immunization; failure of vaccination can be due to either host or vaccine issues. individuals who are immunosuppressed as, for example, a result of hiv infection, malnutrition, immunosuppressive therapy, or immune senescence might not mount a sufficient response after vaccination so as to be adequately immunized (protected). similarly, vaccination with an inadequate amount of vaccine or a vaccine of poor quality (e.g., due to break in cold chain delivery) might prevent even a healthy individual from becoming immunized. environmental determinants of vulnerability to infectious diseases include physical, social, behavioral, cultural, political, and economic factors. in some cases, environmental influences increase risk of exposure to an infectious agent. for example, following an earthquake, environmental disruption can increase the risk of exposure to clostridium tetani and result in host traumatic injuries that provide portals of entry for the bacterium. environmental factors promoting vulnerability can also lead to an increase in susceptibility to infection by inducing physiological changes in an individual. for example, a child living in a resource-poor setting and vulnerable to malnutrition may be at increased risk of infection due to malnutritioninduced immunosuppression. table 2 provides examples of some of the many environmental factors that can facilitate the emergence and/or spread of specific infectious diseases. a unique characteristic of many infectious diseases is that exposure to certain infectious agents can have consequences for other individuals, because an infected person can act as a source of exposure. some pathogens (e.g., sti agents) are directly transmitted to other people, while others (e.g., vectorborne disease (vbd) agents) are transmitted indirectly. from a public health standpoint, it is useful to define stages of an infectious disease with respect to both clinical disease and potential for transmission ( figure 3 ). with respect to disease, the incubation period is defined as the time from exposure to an infectious agent until the time of first signs or symptoms of disease. the incubation period is followed by the period of clinical illness which is the duration between first and last disease signs or symptoms. with respect to transmission of an infectious agent, the latent (preinfectious) period is the duration of time between exposure to an agent and the onset of infectiousness. it is followed by the infectious period (a.k.a. period of communicability) which is the time period when an infected person can transmit an infectious agent to other individuals. in parasitic infections, the latent and infectious periods are commonly referred to as the prepatent period and patent period, respectively. the duration of disease stages is unique for each type of infection and it can vary widely for a given type of infection depending upon agent, host, and environmental factors that affect, for example, dose of the inoculated agent, route of exposure, host susceptibility, and agent infectivity and virulence. knowledge of the timing of disease stages is of key importance in the design of appropriate control and prevention strategies to prevent the spread of an infectious disease. for example, efforts to control the recent ebola west africa outbreak through contact tracing and quarantine were based on knowledge that the infectious period for ebola does not begin until the start of the period of clinical illness, which occurs up to 21 days following exposure (figure 3(a) ; pandey et al., 2014) . a carrier is, by definition, an infectious individual who is not showing clinical evidence of disease and, thus, might unknowingly facilitate the spread of an infectious agent through a population. incubatory carriers exist when the incubation period overlaps with the infectious period, as can occur in some cases of chicken pox (figure 3(b) ). convalescent carriers occur when the period of infectiousness extends beyond the period of clinical illness (figure 3(c) ). carriers of this type can be a significant issue in promoting the spread of certain enteric infections, such as those caused by the bacterium, v. cholerae. healthy carriers, infected individuals that remain asymptomatic but are capable of transmitting an infectious agent, occur commonly with many infectious diseases (e.g., meningococcal meningitis and typhoid fever) and are also significant challenges to disease control ( figure 3 (d)). a variety of terms are used to describe the occurrence of an infectious disease within a specific geographic area or population. sporadic diseases occur occasionally and unpredictably, while endemic diseases occur with predictable regularity. levels of endemicity can be classified as holoendemic, hyperendemic, mesoendemic, or hypoendemic depending upon whether a disease occurs with, respectively, extreme, high, moderate, or low frequency. for some infectious diseases, such as malaria, levels of endemicity are well defined and used as parameters for identifying disease risk and implementing control activities. malaria endemicity is quantified based upon rates of palpable enlarged spleens in a defined (usually pediatric) age group: holoendemic >75%, hyperendemic 51-75%, mesoendemic 11-50%, and hypoendemic <10% (hay et al., 2008 ). an epidemic refers to an, often acute, increase in disease cases above the baseline level. an epidemic may reflect an escalation in the occurrence of an endemic disease or the appearance of a disease that did not previously exist in a population. the term outbreak is often used synonymously with epidemic but can occasionally refer to an epidemic occurring in a more limited geographical area; for example, a foodborne illness associated with a group gathering. by contrast, a pandemic is an epidemic that has spread over a large geographic region, encompassing multiple countries or continents, or extending worldwide. influenza commonly occurs as a seasonal epidemic, but periodically it gives rise to a global pandemic, as was the case with 2009 h1n1 influenza. two fundamental measures of disease frequency are prevalence and incidence. prevalence is an indicator of the number of existing cases in a population as it describes the proportion of individuals who have a particular disease, measured either at a given point in time (point prevalence) or during a specified time period (period prevalence). in contrast, incidence (a.k.a. incidence rate) is a measurement of the rate at which new cases of a disease occur (or are detected) in a population over a given time period. usually measured as a proportion (number infected/number exposed), attack rates are often calculated during an outbreak. in some circumstances, a secondary attack rate is calculated to quantify the spread of disease to susceptible exposed persons from an index case (the case first introducing an agent into a setting) in a circumscribed population, such as in a household or hospital. during the 2003 sars epidemic, secondary attack rates in toronto hospitals were high but varied from 25% to 40% depending upon the hospital ward (cdc, 2003b) . the basic reproductive number (basic reproductive ratio; r 0 ) is a measure of the potential for an infectious disease to spread through an immunologically naïve population. it is defined as the average number of secondary cases generated by a single, infectious case in a completely susceptible population. in reality, for most infectious diseases entering into a community, some proportion of the population is usually immune (and nonsusceptible) due to previous infection and/or immunization. thus, a more accurate reflection of the potential for community disease spread is the effective reproductive number (r) which measures the average number of new infections due to a single infection. in general, for an epidemic to occur in a population, r must be >1 so that the number of cases continues to increase. herd immunity (a.k.a. community immunity) refers to population-level resistance to an infectious disease that occurs when there are enough immune individuals to block the chain of infection/transmission. as a result of herd immunity, susceptible individuals who are not immune themselves are indirectly protected from infection ( figure 4 ). vaccine hesitancy, the choice of individuals or their caregivers to delay or decline vaccination, can lead to overall lower levels of herd immunity. outbreaks of measles in the united states, including a large 2014 measles outbreak at an amusement park in california, highlight the phenomena of vaccine refusal and associated increased risk for vaccine-preventable diseases among both nonvaccinated and fully vaccinated (but not fully protected) individuals (phadke et al., 2016) . an important public health consequence of herd immunity is that immunization coverage does not need to be 100% for immunization programs to be successful. the equation r ¼ r 0 (1 à x) (where x equals the immune portion of the population) indicates the level of immunization required to prevent the spread of an infectious disease through a population. the proportion that needs to be immunized depends on the pathogen (table 3) . when the proportion immunized (x) reaches a level such that r < 1, a chain of infection cannot be sustained. thus, ro and r can be used to calculate the target immunization coverage needed for the success of vaccination programs. proper diagnosis of infectious illnesses is essential for both appropriate treatment of patients and carrying out prevention and control surveillance activities. two important properties that should be considered for any diagnostic test utilized are sensitivity and specificity. sensitivity refers to the ability of the test to correctly identify individuals infected with an agent ('positive in disease'). a test that is very sensitive is more likely to pick up individuals with the disease (and possibly some without the disease); a very sensitive test will have few false negatives. specificity is the ability of the test to correctly identify individuals not infected by a particular agent ('negative in health'); high specificity implies few false positives. often, screening tests are highly sensitive (to capture any possible cases), and confirmatory tests are more specific (to rule out false-positive screening tests). broadly, laboratory diagnosis of infectious diseases is based on tests that either directly identify an infectious agent or provide evidence that infection has occurred by documenting agent-specific immunity in the host ( figure 5 ). identification of an infecting agent involves either direct examination of host specimens (e.g., blood, tissue, urine) or environmental specimens, or examination following agent culture and isolation from such specimens. the main categories of analyses used in pathogen identification can be classified as phenotypic, revealing properties of the intact agent, nucleic acid-based, determining agent nucleic acid (dna or rna) characteristics and composition, and immunologic, detecting microbial antigen or evidence of immune response to an agent ( figure 5 ). direct phenotypic analyses include both macroscopic and/or microscopic examination of specimens to determine agent morphology and staining properties. cultured material containing large quantities of agent can undergo analyses to determine characteristics, such as biochemical enzymatic activity (enzymatic profile) and antimicrobial sensitivity, and to perform phage typing, a technique which differentiates bacterial strains according to the infectivity of strain-specific bacterial viruses (a.k.a. bacteriophages). nucleic acid-based tests often make use of the polymerase chain reaction (pcr) to amplify agent dna or complementary dna (cdna) synthesized from messenger rna (mrna). the ability of pathogen-specific pcr primers to generate an amplification product can confirm or rule out involvement of a specific pathogen. sequencing of amplified dna fragments can also assist with pathogen identification. restriction fragment analysis, as by pulse-field gel electrophoresis of restriction enzyme-digested genomic dna isolated from cultured material, can yield distinct 'dna fingerprints' that can be used for comparing the identities of bacteria. the cdc pulsenet surveillance program uses dna fingerprinting as the basis for detecting and defining foodborne disease outbreaks that can sometimes be quite widely dispersed (cdc, 2013) . most recently, next-generation sequencing technologies have made whole-genome sequencing a realistic subtyping method for use in foodborne outbreak investigation and surveillance (deng et al., 2016) . the objective of immunologic analysis of specimens is to reveal evidence of an agent through detection of its antigenic components with agent-specific antibodies. serotyping refers to the grouping of variants of species of bacteria or viruses based on shared surface antigens that are figure 4 herd immunity occurs when one is protected from infection by immunization occurring in the community. using influenza as an example, the top box shows a population with a few infected individuals (shown in red) and the rest healthy but unimmunized (shown in blue); influenza spreads easily through the population. the middle box depicts the same population but with a small number who are immunized (shown in yellow); those who are immunized do not become infected, but a large proportion of the population becomes infected. in the bottom box, a large proportion of the population is immunized; this prevents significant transmission, including to those who are unimmunized. source: national institute of allergy and infectious diseases (niaid). identified using immunologic methodologies such as enzymelinked immunosorbent assay (elisa) and western blotting. immunologic assays are also used to look for evidence that an agent-specific immune response has occurred in an exposed or potentially exposed individual. serologic tests detect pathogen-specific b cell-secreted antibodies in serum or other body fluids. some serologic assays simply detect the ability of host antibodies to bind to killed pathogen or components of pathogen (e.g., elisa). others rely on the ability of antibodies to actually neutralize the activity of live microbes; as, for example, the plaque reduction neutralization test which determines the ability of serum antibodies to neutralize virus. antibody titer measures the amount of a specific antibody present in serum or other fluid, expressed as the greatest dilution of serum that still gives a positive test in whatever assay is being employed. intradermal tests for identification of t cell-mediated immediate type (type i) hypersensitivity or delayed type (type iv) hypersensitivity responses to microbial antigen can be used to diagnose or support the diagnosis of some bacterial, fungal, and parasitic infections, such as, the mantoux (tuberculin) test for tb. based on the classic model of leavell and clark (1965) , infectious disease prevention activities can be categorized as primary, secondary, or tertiary. primary prevention occurs at the predisease phase and aims to protect populations, so that infection and disease never occur. for example, measles immunization campaigns aim to decrease susceptibility following exposure. the goal of secondary prevention is to halt the progress of an infection during its early, often asymptomatic stages so as to prevent disease development or limit its severity; steps important for not only improving the prognosis of individual cases but also preventing infectious agent transmission. for example, interventions for secondary prevention of hepatitis c in injection drug user populations include early diagnosis and treatment by active surveillance and screening (miller and dillon, 2015) . tertiary prevention focuses on diseased individuals with the objective of limiting impact through, for example, interventions that decrease disease progression, increase functionality, and maximize quality of life. broadly, public health efforts to control infectious diseases focus on primary and secondary prevention activities that reduce the potential for exposure to an infectious agent and increase host resistance to infection. the objective of these activities can extend beyond disease control, as defined by the 1997 dahlem workshop on the eradication of infectious diseases, to reach objectives of elimination and eradication (dowdle, 1998 ; box 1). as noted earlier, the causation and spread of an infectious disease is determined by the interplay between agent, host, and environmental factors. for any infectious disease, this interplay requires a specific linked sequence of events termed the chain of infection or chain of transmission ( figure 6 ). the chain starts with the infectious agent residing and multiplying in some natural reservoir; a human, animal, or part of the environment such as soil or water that supports the existence of the infectious agent in nature. the infectious agent leaves the reservoir via a portal of exit and, using some mode of transmission, moves to reach a portal of entry into a susceptible host. a thorough understanding of the chain of infection is crucial for the prevention and control of any infectious disease, as breaking a link anywhere along the chain will stop transmission of the infectious agent. often more than one intervention can be effective in controlling a disease, and the approach selected will depend on multiple factors such as economics and ease with which an intervention can be executed in a given setting. it is important to realize that the potential for rapid and far-reaching movement of infectious agents that has accompanied globalization means that coordination of intervention activities within and between nations is required for optimal prevention and control of certain diseases. the cause of any infectious disease is the infectious agent. as discussed earlier, many types of agents exist, and each can be characterized by its traits of infectivity, pathogenicity, and virulence. a reservoir is often, but not always, the source from which the agent is transferred to a susceptible host. for example, bats are both the reservoir for marburg virus and a source of infection for humans and bush animals including african gorillas. however, because morbidity and mortality due to marburg infection is significant among these bush animals, they cannot act as a reservoir to sustain the virus in nature (they die too quickly), although they can act as a source to transmit marburg to humans. infectious agents can exist in more than one type of reservoir. the number and types of reservoirs are important determinants of how easily an infectious disease can be prevented, controlled, and, in some cases, eliminated or eradicated. animal, particularly wild animal, reservoirs, and environmental reservoirs in nature can be difficult to manage and, thus, can pose significant challenges to public health control efforts. in contrast, infectious agents that only occur in human reservoirs are among the 1997 dahlem workshop on the eradication of infectious diseases defined a continuum of outcomes due to public health interventions targeting infectious diseases: "1) control, the reduction of disease incidence, prevalence, morbidity or mortality to a locally acceptable level as a result of deliberate efforts; continued intervention measures are required to maintain the reduction (e.g. diarrheal diseases), 2) elimination of disease, reduction to zero of the incidence of a specified disease in a defined geographical area as a result of deliberate efforts; continued intervention measures are required (e.g. neonatal tetanus), 3) elimination of infections, reduction to zero of the incidence of infection caused by a specific agent in a defined geographical area as a result of deliberate efforts; continued measures to prevent re-establishment of transmission are required (e.g. measles, poliomyelitis),4) eradication, permanent reduction to zero of the worldwide incidence of infection caused by a specific agent as a result of deliberate efforts; intervention measures are no longer needed (e.g. smallpox), and 5) extinction, the specific infectious agent no longer exists in nature or in the laboratory (e.g. none)" (dowdle, 1998) . the chain of infection (a.k.a. chain of transmission). one way to visualize the transmission of an infectious agent though a population is through the interconnectedness of six elements linked in a chain. public health control and prevention efforts focus on breaking one or more links of the chain in order to stop disease spread. those most easily targeted, as illustrated by the success of smallpox eradication. humans are the reservoir for many common infectious diseases including stis (e.g., hiv, syphilis) and respiratory diseases (e.g., influenza). humans also serve as a reservoir, although not always a primary reservoir, for many neglected tropical diseases (ntds) as, for example, dracunculiasis (a.k.a. guinea worm). from a public health standpoint, an important feature of human reservoirs is that they might not show signs of illness and, thus, can potentially act as unrecognized carriers of disease within communities. the classic example of a human reservoir is the cook mary mallon (typhoid mary); an asymptomatic chronic carrier of salmonella enterica serovar typhi who was linked to at least 53 cases of typhoid fever (soper, 1939) . animals are a reservoir for many human infectious diseases. zoonosis is the term used to describe any infectious disease that is naturally transmissible from animals to humans. these diseases make up approximately 60% of all infectious diseases, and an estimated 75% of recently emerging infectious diseases (burke et al., 2012) . zoonotic reservoirs and sources of human disease agents include both domestic (companion and production) animals (e.g., dogs and cows) and wildlife. control and prevention of zoonotic diseases requires the concerted efforts of professionals of multiple disciplines and is the basis for what has become known as the one health approach (gibbs, 2014) . this approach emphasizes the interconnectedness of human health, animal health, and the environment and recognizes the necessity of multidisciplinary collaboration in order to prevent and respond to public health threats. inanimate matter in the environment, such as soil and water, can also act as a reservoir of human infectious disease agents. the causative agents of tetanus and botulism (clostridium tetani and c. botulinum) are examples of environmental pathogens that can survive for years within soil and still remain infectious to humans. legionella pneumophila, the etiologic agent of legionnaires' disease, is part of the natural flora of freshwater rivers, streams, and other bodies. however, the pathogen particularly thrives in engineered aquatic reservoirs such as cooling towers, fountains, and central air conditioning systems, which provide conditions that promote bacterial multiplication and are frequently linked to outbreaks. soil and water are also sources of infection for several protozoa and helminth species which, when excreted by a human reservoir host, can often survive for weeks to months. outbreaks of both cryptosporidiosis and giardiasis commonly occur during summer months as a result of contact with contaminated recreational water. soil containing roundworm (ascaris lumbricoides) eggs is an important source of soil-transmitted helminth infections in children. early steps in preventing exposure to an infectious agent include interventions to control or eliminate the agent within its reservoir, to neutralize or destroy the reservoir, and/or to stop the agent from exiting its reservoir. central to these interventions are surveillance activities that routinely identify disease agents within reservoirs. when humans are the reservoir, or source, of an infectious agent, early and rapid diagnosis and treatment are key to decreasing the duration of infection and risk of transmission. both active surveillance and passive surveillance are used to detect infected cases and carriers. some readily communicable diseases, such as ebola, can require isolation of infected individuals to minimize the risk of transmission. as part of the global effort to eradicate dracunculiasis, several endemic countries have established case containment centers to provide treatment and support to patients with emerging guinea worms to keep them from contaminating water sources and, thereby, exposing others (hochberg et al., 2008) . contact tracing and quarantine are other activities employed in the control of infections originating from a human reservoir or source. during the west africa ebola outbreak, key control efforts focused on the tracing and daily follow-up of healthy individuals who had come in contact with ebola patients and were potentially infected with the virus (pandey et al., 2014) . one health emphasizes the importance of surveillance and monitoring for zoonotic pathogens in animal populations. for some diseases (e.g., rift valley fever) epizootics (analogous to epidemics, but in animal populations) can actually serve as sentinel events for forecasting impending human epidemics. once animal reservoirs (and sources) of infection are identified, approaches to prevention and control include reservoir elimination and prevention of reservoir infection. zoonotic diseases exist in nature in predictably regular, enzootic cycles and/or epizootic cycles and are transmitted to humans via distinct pathways. the focus of prevention and control activities for these diseases reflects the extent to which a zoonotic pathogen has evolved to become established in human populations (wolfe et al., 2007) . for some zoonotic diseases (e.g. anthrax, nipah, rabies), primary transmission always occurs from animals, with humans acting as incidental (dead end) hosts; control of these diseases thus concentrates on preventing animal-to-animal and, ultimately, animal-to-human transmission. currently, most human cases of avian influenza are the result of human infection from birds; human-to-human transmission is extremely rare. thus, reservoir elimination by culling infected poultry flocks is a recommended measure for controlling avian influenza in birds and preventing sporadic infection of humans (cdc, 2015) . other zoonotic diseases demonstrate varying degrees of secondary human-to-human transmission following primary transmission (a.k.a. spillover) from animals. both rates of spillover and the ability to sustain human-tohuman transmission can vary widely between zoonoses and, in consequence, control strategies can also be quite different. for example, outbreaks of ebola arise following an initial bush animal-to-human transmission event, and subsequent human-to-human transmission is often limited (feldmann and geisbert, 2011) . in contrast, the four dengue viruses originally emerged from a sylvatic cycle between non-human primates and mosquitoes, and are now sustained by a continuous human-mosquito-human cycle of transmission with outbreaks occurring as a result of infected individuals entering into naïve populations (vasilakis et al., 2011) . thus, while ebola outbreak prevention efforts would include limiting contact with bush animals, such efforts would not be useful for prevention of dengue outbreaks. hiv is an example of a virus that emerged from an ancestral animal virus, simian immunodeficiency virus, but has evolved so that it is now hiv is an example of a virus that emerged from an ancestral animal virus, simian immunodeficiency virus, but has evolved so that it is now only transmitted human to human (faria et al., 2014) . infectious agents exit human and animal reservoirs and sources via one of several routes which often reflect the primary location of disease; respiratory disease agents (e.g., influenza virus) usually exit within respiratory secretions, whereas gastrointestinal disease agents (e.g., rotavirus, cryptosporidium spp.) commonly exit in feces. other portals of exits include sites from which urine, blood, breast milk, and semen leave the host. for some infectious diseases, infection can naturally occur as a result of contact with more than one type of bodily fluid, each of which uses a different portal of exit. while infection with the sars virus most frequently occurred via contact with respiratory secretions, a large community outbreak was caused by the spread of virus in a plume of diarrhea (yu et al., 2004) . control interventions targeting portals of exit and entry are discussed below. there are a variety of ways in which infectious agents move from a natural reservoir to a susceptible host, and several different classification schemes are used. the scheme below categorizes transmission as direct transmission, if the infective form of the agent is transferred directly from a reservoir to an infected host, and indirect transmission, if transfer takes place via a live or inanimate intermediary (box 2). direct physical contact between the skin or mucosa of an infected person and that of a susceptible individual allows direct transfer of infectious agents. this is a mode of transmission for most stis and many other infectious agents, such as bacterial and viral conjunctivitis (a.k.a. pink eye) and ebola virus disease. direct droplet transmission occurs after sneezing, coughing, or talking projects a spray of agent-containing droplets that are too large to remain airborne over large distances or for prolonged periods of time. the infectious droplets traverse a space of generally less than 1 m to come in contact with the skin or mucosa of a susceptible host. many febrile childhood diseases, including the common cold, are transferred this way. diseases spread by direct contact and droplet transmission require close proximity of infected and susceptible individuals and, thus, commonly occur in settings such as households, schools, institutions of incarceration, and refugee/displaced person camps. infectious agents spread exclusively in this manner are often unable to survive for long periods outside of a host; direct transmission helps to ensure transfer of a large infective dose. direct contact to an agent in the environment is a means of exposure to infectious agents maintained in environmental reservoirs. diseases commonly transmitted in this manner include those in which the infectious agent enters a susceptible host via inhalation (e.g., histoplasmosis) or through breaks in the skin following a traumatic event (e.g., tetanus). animal bites are another way in which some infectious agents are directly transferred, through broken skin. this is the most common means of infection with rabies virus. transplacental (a.k.a. congenital, vertical) and perinatal transmissions occur during pregnancy and delivery or breastfeeding, respectively. classic examples include mother-to-child transmission of the protozoa toxoplasma gondii during pregnancy, hiv during pregnancy, delivery, or breastfeeding, and zika virus during pregnancy (rasmussen et al., 2016) . case finding and contact tracing are public health prevention and control activities aimed at stopping the spread of infectious diseases transmitted by either direct contact or direct spread of droplets. once identified, further activities to limit transmission to susceptible individuals can involve definitive diagnosis, treatment, and, possibly, isolation of active cases and carriers, and observation, possible quarantine, or prophylactic vaccination or treatment of contacts. patient education is an important feature of any communicable infectious disease control effort. environmental changes, such as decreasing overcrowded areas and increasing ventilation, can also contribute to limiting the spread of some infectious diseases, particularly respiratory diseases. central to prevention of transplacental and perinatal infectious disease transmission is avoidance of maternal infection and provision of early diagnosis and treatment of infected women prior to or during pregnancy. for example, public health efforts targeting congenital toxoplasmosis focus on preventing pregnant women from consuming undercooked meat or contacting cat feces that may be contaminated. current who guidelines for prevention of mother-to-child hiv transmission recommend that hiv-infected pregnant and breastfeeding women should be maintained on antiretrovirals (who, 2013) . there are three main categories of indirect transmission: biological, mechanical, and airborne. box 3 provides definitions of the different types of hosts, vectors, and vehicles involved in the life cycle of agents that are transmitted indirectly. biological transmission occurs when multiplication and/or development of a pathogenic agent within a vector (e.g., biological vector or intermediate host) is required for the agent to become infectious to humans. the time that is necessary for these events to occur is known as the extrinsic incubation period; in contrast to the intrinsic incubation period which is the time required for an exposed human host to become infectious. indirect transmission by mosquito vectors is the primary mode of transmission of a large number of viruses (arthropod-borne viruses or arboviruses) of public health concern (e.g., west nile, zika). a number of ntds are also transmitted by biological vectors including lymphatic filariasis (a.k.a. elephantiasis) by mosquitoes. ticks are biological vectors for many bacterial etiological agents (e.g., lyme disease and ehrlichiosis), and the parasitic agent causing babesiosis. the infectious agent of the helminthic ntds, schistosomiasis, and dracunculiasis are transmitted indirectly via intermediate freshwater snail and copepod hosts, respectively. mechanical transmission does not require pathogen multiplication or development within a living organism. it occurs when an infectious agent is physically transferred by a live entity (mechanical vector) or inanimate object (vehicle) to a susceptible host. classic examples of diseases spread by mechanical vector transmission are shigellosis (transmission of shigella spp. on the appendages of flies) and plague (transmission of yersinia pestis by fleas). many diarrheal diseases are transmitted by the fecal-oral route with food and water often acting as vehicles (figure 7) . other types of vehicles for infectious disease agents are biologic products (e.g., blood, organs for transplant) and fomites (inanimate objects such as needles, surgical instruments, door handles, and bedding). transfusion-related protozoal infection resulting in chagas disease has been of increasing concern to the us blood banks that have instituted screening measures (cdc, 2007) . airborne transmission involves aerosolized suspensions of residue (less than five microns in size, from evaporated aerosol droplets) or particles containing agents that can be transported over time and long distance and still remain infective. tb is a classic example of an infectious disease often spread by airborne transmission. vbds comprise approximately 17% of the global burden of infectious diseases (townson et al., 2005) . for some diseases (e.g., dengue, zika, chagas), chemoprophylaxis and immunoprophylaxis are not prevention and control options, leaving vector control as the primary means of preventing disease transmission. integrated vector management is defined by the who as, "a rational decision-making process to optimize the use of resources for vector control" (who, 2012) . there are four major categories of ivm vector control strategies: biological, chemical, environmental, and mechanical. ivm interventions are chosen from these categories based upon available resources, local patterns of disease transmission, and ecology of local disease vectors. two key elements of ivm are collaboration within the health sector and with other sectors (e.g., agriculture, engineering) to plan and carry out vector control activities, and community engagement to promote program sustainability. another core element is the integrated approach which often permits concurrent targeting of multiple vbds, as some diseases are transmitted by a single, common vector, and some vector control interventions can target several different vectors. in addition, combining interventions serves not only to reduce reliance on any single intervention, but also to reduce the selection pressure for insecticide and drug resistance. table 4 , adapted from the 2012 who handbook for ivm, illustrates some of the many types of ivm activities and provides examples of vbds that might be controlled by such interventions (who, 2012) . diarrheal diseases primarily result from oral contact with water, food, or other vehicles contaminated with pathogenic agents originating from human or animal feces. most ($88%) of diarrhea-associated deaths are attributable to unsafe drinking water, inadequate sanitation, and insufficient hygiene (black et al., 2003; prüss-üstün et al., 2008) . interruption of fecaloral transmission through provision of safe water and adequate sanitation, and promotion of personal and domestic hygiene are fundamental to diarrhea prevention and control. fecaloral transmission of a diarrheal agent can occur via one of several routes. in 1958, wagner and lanoix developed a model of major transmission depicted in what has become known as the 'f-diagram,' based on steps within the fecal-oral flow of transmission starting with the letter 'f': fluids (drinking water), definitive host: a host in which a parasite reproduces sexually. humans are definitive hosts for roundworms. by strict definition, mosquitoes are the definitive host of malaria as they are the organism in which sexual reproduction of the agent protozoa, plasmodium spp., occurs. reservoir host: a host that serves to sustain an infectious pathogen as a potential source of infection for transmission to humans. note that a reservoir host will not succumb to infection. lowland gorillas and chimpanzees can be infected by ebola virus, but they are not a reservoir host as they suffer devastating losses from infection. bats are a suspected reservoir for ebola virus. intermediate dead-end host: a host from which infectious agents cannot be transmitted to other susceptible hosts. humans are a dead-end host for west nile virus which normally circulates between mosquitoes and certain avian species. vector: a generic term for a living organism (e.g., biological vector or intermediate host) involved in the indirect transmission of an infectious agent from a reservoir or infected host to a susceptible host. biological vector: a vector (often arthropod) in which an infectious organism must develop or multiply before the vector can transmit the organism to a susceptible host. aedes spp. mosquitoes are a biological vector for dengue, chikungunya, and zika. mechanical vector: a vector (often arthropod) that transmits an infectious organism from one host to another but is not essential to the life cycle of the organism. the house fly is a mechanical vector in the diarrheal disease shigellosis as it carries feces contaminated with the shigella spp. bacterium to a susceptible person. vehicles: inanimate objects that serve as an intermediate in the indirect transmission of a pathogen from a reservoir or infected host to a susceptible host. these include food, water, and fomites such as doorknobs, surgical instruments, and used needles. fingers, flies, fields (crops and soil), floods (representative of surface water in general), and food (wagner and lanoix, 1958 ; figure 7 ). other f's that can be considered include facilities (e.g., settings where transmission is likely to occur such as daycare centers) and fornication. the f-diagram is useful for depicting where water, sanitation, and hygiene (wash) interventions act as barriers in the fecal-oral flow of diarrheal pathogens. safe excreta disposal and handling act as primary barriers to transmission by preventing fecal pathogens from entering the environment. once the environment has become contaminated with pathogen-containing feces, secondary and tertiary barriers to transmission include water treatment, safe transport and storage of water, provision of sewage systems to control flooding, fly control, and good figure 7 the 'f-diagram' illustrates major direct and indirect pathways of fecal-oral pathogen transmission and depicts the roles of water, sanitation, and hygiene interventions in providing barriers to transmission. primary barriers prevent contact with feces, and secondary barriers prevent ingestion of feces. source: water, engineering and development centre (wedc), loughborough university. personal and domestic hygiene (e.g., food hygiene) practices (requiring adequate water quantity) ( figure 7) . as with ivm, the control of diarrheal diseases increases with integration of control measures to achieve multiple barriers to fecal-oral transmission. the basic approach to preventing transmission of an infectious agent from a contaminated vehicle is to prevent contamination of, decontaminate, or eliminate the vehicle. food is a common vehicle for infectious agents, and it can potentially become contaminated at any step along the food production chain of production, processing, distribution, and preparation. production refers to the growing of plants for harvest and raising animals for food. an example of contamination at this step includes the use of fecally contaminated water for crop irrigation. processing refers to steps such as the chopping, grinding, or pasteurizing of food to convert it into a consumer product; if the external surface of a melon is contaminated, chopping it into pieces for sale can result in contamination of the fruit. distribution, in which food is transferred from the place where it was produced and/or processed to the consumer, can result in contamination if, for example, the transportation vehicle is not clean. finally, preparation is the step in which food is made ready to eat; not cleaning a cutting board after cutting raw chicken can result in microbial pathogen crosscontamination of other food items. food hygiene is the term used to describe the conditions and activities employed to prevent or limit microbial contamination of food in order to ensure food safety. decontamination includes sterilization, the destruction of all microbial agents, and disinfection, the destruction of specific agents. control of airborne diseases focuses on regulating environmental airflow and air quality to minimize contact with infectious droplet nuclei. in health-care settings, negative pressure isolation rooms and exhaust vents can be used to manipulate airflow. recirculating, potentially infectious air can undergo high-efficiency particulate air (hepa) filtration and/or be mixed with 'clean' (noncontaminated) air to remove or dilute the concentration of infectious particle to below the infectious dose. health-care workers should use n95 masks. on commercial aircraft, airborne pathogen transmission is minimized by methods including regulating airflow to prevent widespread dispersal of airborne microbes throughout the cabin, hepa filtering recirculating air, and mixing recirculating air with fresh air (considered sterile) (dowdall et al., 2010) . the portal of entry refers to the site at which the infectious agent enters a susceptible host and gains access to host tissues. many portals of entry are the same as portals of exit and include the gastrointestinal, genitourinary, and respiratory tracts, as well as compromised skin and mucous membrane surfaces. some infectious agents can naturally enter a susceptible host by more than one portal. for example, the three forms of human anthrax can be distinguished according to the route of agent entry: cutaneous anthrax due to entry through the skin, gastrointestinal anthrax resulting from ingestion of spores, and pulmonary anthrax following inhalation of spores. standard infection control practices target portals of exit (and entry) of infectious agents from human reservoirs and sources. cdc guidelines suggest two levels of precautions to stop transmission of infectious agents: standard precautions and transmission-based precautions (siegel et al., 2007) . standard precautions prevent transmission of infectious agents that can be acquired by contact with blood, body fluids, nonintact skin, and mucous membranes. they can be used to prevent transmission in both health-care and non-health-care settings, regardless of whether infection is suspected or confirmed. hand hygiene is a major component of these precautions, along with use of personal protective equipment (ppe). common ppe include gloves, gowns, face protection (e.g., eye-protecting face shields), and respiratory protection using n95 masks to prevent inhalation of airborne infectious particles (e.g., from mycobacterium tuberculosis). of note, depending on the circumstance, ppe can be used to prevent dispersal of infectious agents from their source by providing a barrier to the portal of exit, or to protect a susceptible individual by placing a barrier to a portal of entry. respiratory hygiene/cough etiquette is used to prevent spread of infection by respiratory droplets. main elements of respiratory hygiene/cough etiquette include covering the nose and mouth area with one's elbow during coughing or sneezing or using a surgical mask to limit dissemination of infectious respiratory secretions, and hand hygiene after contact with respiratory secretions. other components of standard precautions include needle stick and sharp injury prevention, safe injection practices, cleaning and disinfection of potentially contaminated equipment and other objects, and safe waste management. a susceptible host is an individual who is at risk of infection and disease following exposure to an infectious agent. as discussed previously, there are many determinants of host susceptibility, including both innate factors determined by the genetic makeup of the host and, acquired factors such as agent-specific immunity and malnutrition. important prevention and control interventions that target the susceptible host include both those that address determinants of susceptibility in the host (e.g., immunoprophylaxis, provision of adequate nutrition, treatment of underlying diseases) and those that target an infecting agent (e.g., chemoprophylaxis). immunoprophylaxis encompasses both active immunization by vaccination and passive immunization through provision of pathogen-specific immunoglobulin. malnutrition is a strong risk factor for morbidity and mortality due to diarrheal disease, and a vicious cycle exists between infectious diarrheal disease leading to malnutrition and impaired immune function which, in turn, promotes increased susceptibility to infection (keusch et al., 2006) . consequently, breastfeeding and safe complementary feeding play crucial roles in protecting infants and young children from infectious diseases, particularly in resource-poor settings. micronutrients are required for normal immune function, and vitamin a and zinc supplementations have been shown to decrease some types of infections in children deficient in these micronutrients (mayo-wilson et al., 2014; imdad et al., 2010) . in certain circumstances, chemoprophylaxis is employed to protect a susceptible host in anticipation of, or following exposure to an infectious agent. antimalarial drugs are routinely used in combination with personal protective measures to prevent malaria in travelers and established guidelines exist for antibiotic prophylaxis prior to surgery. another important element in the prevention and control of infections is the recognition and management of patients with underlying diseases and conditions that can weaken host barriers to infection. for example, tb is the leading opportunistic infection in hivinfected individuals, and antiretroviral therapy reduces risk of developing tb and mortality due to tb disease. infectious complications are a major cause of morbidity and mortality in cancer and transplant patients, often resulting from immunosuppression that can be primary or related to drug and/or radiation therapy. infectious disease control is also critical in individuals with compromised physical barriers to microbes as, for example, burn patients and patients with cystic fibrosis. dr william h stewart, the one-time surgeon general of the united states, has been quoted (perhaps mistakenly) as saying in the 1960s "it is time to close the book on infectious diseases, and declare the war against pestilence won (spellberg, 2008) ." these words clearly do not hold true today, and public health practitioners wage an ever-growing fight against emerging pathogens, drug-resistant organisms, and vaccine-preventable diseases. in this light, it is all the more important that we have the tools needed to understand transmission dynamics and implement effective prevention and control programs. clear definitions of terminology and elucidation of fundamental principles lay the foundation for effective public health interventions. hopefully, this article helps strengthen the armamentarium of the public health practitioner. narrative review: tetanus-a health threat after natural disasters in developing countries protective effects of the sickle cell gene against malaria morbidity and mortality revisiting leishmaniasis in the time of war: the syrian conflict and the lebanese outbreak estimating the reproduction number of ebola virus (ebov) during the 2014 outbreak in west africa chagas disease and transfusion medicine: a perspective from non-endemic countries prediction, assessment of the rift valley fever activity in east and southern africa 2006-2008 and possible vector control strategies environmental (saprozoic) pathogens of engineered water systems: understanding their ecology for risk assessment and management case-control study of an acute aflatoxicosis outbreak a multistate outbreak of escherichia coli o157:h7-associated bloody diarrhea and hemolytic uremic syndrome from hamburgers. the washington experience where and why are 10 million children dying every year? preventing tetanus, diphtheria, and pertussis among adolescents: use of tetanus toxoid, reduced diphtheria toxoid and acellular pertussis vaccines recommendations of the advisory committee on immunization practices (acip) dengue fever seroprevalence and risk factors a review of zoonotic disease surveillance supported by the armed forces health surveillance center extreme water-related weather events and waterborne disease morbidity and mortality surveillance in rwandan refugees-burundi and zaire multistate outbreak of monkeypox-illinois, indiana, and wisconsin update: severe acute respiratory syndrome-toronto, canada blood donor screening for chagas disease -united states avian influenza in birds smallpox: disease, prevention, and intervention genomic epidemiology: wholegenome-sequencing-powered surveillance and outbreak investigation of foodborne bacterial pathogens air travel and tb: an airline perspective the principles of disease elimination and eradication tracing seeds, in particular fenugreek (trigonella foenum-graecum) seeds, in relation to the on request of european food safety authority, question no efsa-q-2011-00817, issued on 05 climatic and environmental patterns associated with hantavirus pulmonary syndrome, four corners region, united states past, present, and future of japanese encephalitis hiv epidemiology. the early spread and epidemic ignition of hiv-1 in human populations bad bug book, foodborne pathogenic microorganisms and natural toxins ebola haemorrhagic fever immune subversion and quorum-sensing shape the variation in infectious dose among bacterial pathogens the evolution of one health: a decade of progress and challenges for the future outpatient antibiotic use in europe and association with resistance: a cross-national database study of mice and men: lyme disease and biodiversity avian flu from an occupational health perspective measuring malaria endemicity from intense to interrupted transmission a national outbreak of salmonella enteritidis infections from ice cream. the investigation team cultural contexts of ebola in northern uganda the role of case containment centers in the eradication of dracunculiasis in togo and ghana anisakidosis: perils of the deep vitamin a supplementation for preventing morbidity and mortality in children from 6 months to 5 years of age new evidence of the effects of agro-ecologic change on malaria transmission effect of irrigation and large dams on the burden of malaria on a global and regional scale diarrheal diseases immunity to malaria: more questions than answers preventive medicine for the doctor in his community; an epidemiologic approach zinc supplementation for preventing mortality, morbidity, and growth failure in children aged 6 months to 12 years of age early diagnosis improves outcomes in hepatitis c schistosomiasis among recreational users of factors in the emergence of infectious diseases global, regional, and national age-sex specific all-cause and cause-specific mortality for 240 causes of death, 1990-2013: a systematic analysis for the global burden of disease study global epidemiology of hepatitis b and hepatitis c in people who inject drugs: results of systematic reviews strategies for containing ebola in west africa severe acute respiratory syndrome association between vaccine refusal and vaccine-preventable diseases in the united states: a review of measles and pertussis urban habituation, ecological connectivity and epidemic dampening: the emergence of hendra virus from flying foxes (pteropus spp.) the natural history of ebola virus in safer water, better health: costs, benefits and sustainability of interventions to protect and promote health cofactors for the acquisition of hiv-1 among heterosexual men: prospective cohort study of trucking company workers in kenya zika virus and birth defects -reviewing the evidence for causality infection with chikungunya virus in italy: an outbreak in a temperate region new insights into the immunopathology and control of dengue virus infection guideline for isolation precautions: preventing transmission of infectious agents in health care settings a political and social history of hiv in south africa urban lymphatic filariasis the effects of environmental stress on outbreaks of infectious diseases of fishes* the curious career of typhoid mary dr. william h. stewart: mistaken or maligned? yellow fever vaccine booster doses: recommendations of the advisory committee on immunization practices schistosomiasis and water resources development: systematic review, meta-analysis, and estimates of people at risk exploiting the potential of vector control for disease prevention fever from the forest: prospects for the continued emergence of sylvatic dengue virus and its impact on public health global, regional, and national incidence, prevalence, and years lived with disability for 301 acute and chronic diseases and injuries in 188 countries, 1990-2013: a systematic analysis for the global burden of disease study excreta disposal for rural areas and small communities. who consolidated guidelines on the use of antiretroviral drugs for treating and preventing hiv infection large outbreak of cryptosporidium hominis infection transmitted through the public water supply origins of major human infectious diseases impact of deforestation and agricultural development on anopheline ecology and malaria epidemiology evidence of airborne transmission of the severe acute respiratory syndrome virus key: cord-328181-b2o05j3j authors: nunez-corrales, s.; jakobsson, e. title: the epidemiology workbench: a tool for communities to strategize in response to covid-19 and other infectious diseases date: 2020-07-25 journal: nan doi: 10.1101/2020.07.22.20159798 sha: doc_id: 328181 cord_uid: b2o05j3j covid-19 poses a dramatic challenge to health, community life, and the economy of communities across the world. while the properties of the virus are similar from place to place, the impact has been dramatically different from place to place, due to such factors as population density, mobility, age distribution, etc. thus, optimum testing and social distancing strategies may also be different from place to place. the epidemiology workbench provides access to an agent-based model in which demographic, geographic, and public health information a community together with a social distancing and testing strategy may be input, and a range of possible outcomes computed, to inform local authorities on coping strategies. the model is adaptable to other infectious diseases, and to other strains of coronavirus. the tool is illustrated by scenarios for the cities of urbana and champaign, illinois, the home of the university of illinois at urbana-champaign. our calculations suggest that massive testing is the most effective strategy to combat the likely increase in local cases due to mass ingress of a student population carrying a higher viral load than that currently present in the community. mathematical models of infectious disease epidemiological dynamics can be provide valuable assistance to public health officials and health care providers in assessing the likely seriousness of an epidemic or its potential to grow into a pandemic, and later in allocating resources to counter the spread of the disease [105] . simulations that trace either prior or projected time courses make use of various mathematical and computational techniques, including (non-exhaustively) differential equation models, statistical regression and curve fitting, network propagation dynamics, and direct representation of human actors and their actions by means of agent-based models. the sir model in particular along with its various adaptations [100] has remained successful, at least in part, due to its universality as 1 evidenced by its empirical adequacy across multiple epidemics, and by its formal robustness when connecting microscale host-pathogen related events and macroscale disease observables [28, 10] . stochastic versions of the sir model show that adding noise to the system changes the predicted onset of an epidemic [113] , the stability of its endemic equilibrium [123] , the value of its effective reproduction number [57] or its duration [72] when contrasted against the deterministic one. this is significant not only at the theoretical level when studying the stability and asymptotic representativeness of deterministic vs stochastic sir models under various noise regimes, but at the policy making level where computational epidemiology may form the basis of informed decisions under policy, budgetary and other types of constraints [46] . moreover, the sir has been extended spatially to account for the diffusion-like properties associated with geographic patterns observed during epidemics. while the qualitative (and some quantitative) properties of the traditional and the spatial sir models remain largely shared, spatial versions appear to be numerically susceptible to how these capture spatial interactions [101] . as with any other diffusion process in some space, we are usually interested on the ability of a disease to cover larger shares of the population as time marches on. detailed analysis of deterministic and stochastic sir models with spatial components [16] indicates the existence of solutions corresponding to traveling waves that propagate the disease among point-like processes. it has also been shown that spatial sir models can account for the effects of long-distance travel by replacing diffusion operators containing local processes with appropriate integro-differential ones that capture non-local dispersal processes [117] . agent-based models (abms) constitute a family of models where sets of active entities (i.e. agents) interact collectively by following prescribed individual rules intended to portray the emergent dynamics of a real social system [20] . in computational epidemiology, abms have been used and comparatively evaluated against sir models of various kinds. analysis of the behavior of both classes of models suggests that for many purposes the two classes will give qualitatively the same result, but that agent-based models have an advantage in ease of accounting for heterogeneity in subpopulations where that is significant [5, 96] . furthermore, the sir differential equations model is derivable from asymptotic limit of an sir abm model through diffusion approximation [17] . a notable coordinated effort to develop agent-based models of a flu pandemic was the models of infectious disease agents study funded by the national institutes of health [51] . more recently the attention of the world was dramatically drawn to the need for public health interventions in the case of covid-19 by a simulation model projecting 2.2 million deaths from covid-19 in the united states, and 510,000 in the u.k., in the absence of such interventions [41] . since then, models continue to be refined as more data are analyzed [3] . it is important to note that there are enormous local geographic variations in the incidence of, and deaths from, covid-19 [23] . these local variations imply a need for local models, to enable local authorities to construct appropriate strategies of social distancing and testing for mitigation of the effects of covid-19. the work described in this paper is designed to meet this need. wicked policy problems are characterized by 1) complexity of elements to be accounted for and their relationship to each other, 2) uncertainty in relationship to the description of the problem and the consequences of actions, and 3) divergence within the affected community of values and interests [55] . by all three measures of wickedness -complexity, uncertainty, and divergence-covid-19 is a highly wicked problem and will continue to be at least until there is an effective and universally available vaccine. dimensions of complexity in covid-19 emerge from all of the multiple ways in which people interact with each other in such a way as to breathe the same air, and from the consequent trade-offs. these trade-offs involve public health, economics, every aspect of community life, and levels of emotional stress in individuals-a multi-level hierarchy of perspectives involving psychology, sociology, economics, health care, and politics. correspondingly, we expect covid-19 modeling efforts to include a growing number of these concerns while remaining actionable and scientifically useful. we expect the complexity of such models to grow, but to do so in a manner that remains intellectually transparent [38] about what is stated in them. to this extent, models become critical components within the top-level decision support system necessary to regain situational control during the current pandemic. dimensions of uncertainty abound. as noted above, there is enormous geographic variation in the documented impact of the disease, and variation even in the apparent fundamental parameters of the virus -transmissibility, latency, and virulence-for reasons that are not yet understood. contributors to the uncertainty may be genetic variation in human populations [26] , genetic variation in the virus as it continues to evolve [107] , variation in childhood vaccine regimens from one nation to another [82] , variations in weather patterns [61] , and variations in testing rates and disease reporting accuracy and criteria [60] . also, there is an element of pure chance -whether or not a particular community was "seeded" with infectious individuals, and how many. dimensions of divergence are in some ways clear, and in some ways complicated. the clearest divergence is between the imperative to save lives by social distancing and the costs of social distancing to the community-both economic costs [15, 110] and also costs that are less tangible due to how wealth moves across the global economy [79] . early on in this crisis most of the world appears to have made the choice of that we would throw our economies into depression [12] and restrict many community activities we value [40, 97] in order to save the lives of the probably less than 1% of the population who would die from infection should no social distancing constraints be imposed. at the time of writing, this choice is constantly being reconsidered, or at least recalibrated [50] . more, and more open discussions are being held on increasing economic activity even at the acknowledged cost of more infections and even deaths. one is reminded of a famous comedy routine when comedian jack benny, whose comic persona was as a notorious cheapskate, is held at gunpoint by a robber who demands "your money or your life!" this is followed by a long silence, a repeated demand, and a response by benny, "i'm thinking!". it seems that covid-19 has the whole world thinking about the trade-offs between economics -and other aspects of community life-and lives. with respect to divergence, covid-19 seems as wicked as possible. the economic dimensions of community life can be measured in dollars. the many other dimensions have no common units of measurement, so their relative value is literally incalculable. and yet we are forced to decide about what to value. another way to look at a wicked policy problem is a one where the space of potential solution alternatives contains far more social dilemmas than solutions. we may thus define a social dilemma is a situation where multiple agents have (explicit or implicit) stakes in the resolution of a problem, stakes are tied to multiple value systems (and not just shared, "objective" technical considerations for example), and a proposed solution contains value contradictions that get translated to unacceptable potential losses were that solution to materialize; at the same time, the social dilemma also provides consequences if a solution fails to appear .conversely, a perfect solution to a wicked problem is a point of total satisfaction of constraints at all levels of representation of the problem. this is, of course, an idealization; in practice some constraints must be eliminated, relaxed or ignored to find a collective solution. in summary, a social problem is wicked if the density of true solutions is low in the space of all solution alternatives and the search for them can be described as unstructured or even counter-incentivized at best. thus, the final element of covid-19 as a truly wicked problem is that, although it is insoluble, we must make our best effort to solve it. the consequences of not trying to solve this intractable problem, of simply guessing at answers guided only by intuition, are far worse than the consequences of being guided by imperfect models. 3 building a multi-objective model for covid-19: the agent-based route based on the discussion above, our current research efforts have focused on the development of an integrated simulation model capable of a) accurately reflecting known dynamics of the current pandemic and the qualitative results of other models, b) simulating data-driven stochastic heterogeneity across agent populations to more realistically reflect the variability of underlying human populations when the model is applied, c) integrating economic considerations in association with observable features of the pandemic, d) allowing detailed simulation of known public policy measures at different times, intensities and dates, and e) providing a simple interface for non-expert users to configure and interpret. in relation to the latter point (e), we envision assisting the decision-making process in two steps: first by providing some metaphor or visual proxy for users to construct intuitions by running specific scenarios, and then by translating some of these intuitions into fullyfledged computing and analysis tasks. succinctly, we wish to facilitate decision making processes that are both robust and adaptive [66] while helping decision makers to avoid falling into the "illusion of control", or the false belief in the causal relation between computing consequences with a model and immediately improving their decision-making abilities [63] . at the same time, we remain painfully aware of the intrinsic difficulties posed by imperfect data, imperfect implementation of public policy measures, and uncertain timelines for when and for how long to apply measures under unknown timelines for availability of vaccines [104] . ee expect our model to be beneficial when a) decision makers are fully aware of the underlying simplifications we have made, b) model outcomes are contrasted and adjusted with incoming data during an unfolding situation, c) experts assisting decision makers carefully determine and document how data produced by these simulations is analyzed and translated into tentative recommendations [70] . to this extent, we have focused our efforts on providing modeling tools for population centers with 100,000 inhabitants or less. our choice is motivated by the geographic distribution of cities and towns across the us [45] and by the apparent inverse correlation between population size and rurality. this is significant since the push for urbanization seems to have driven rural cities and towns to more precarious health systems than their urban counterparts [95] : one can expect covid-19 propagation to be slower due to lower population densities, but the impact to be at least similar or stronger due to age distribution and availability of health care facilities [8] , with a special emphasis on availability of icus [58] . in our model, agents interact and traverse a discrete 2d torus composed of connected lattice points that represent geographic locations. agent actions and decisions are governed by random variables with suitable distributions. a single execution (i.e. a scenario run) of a parametrization of the model corresponds to a possible world, while a simulation (i.e. a scenario) comprises an ensemble of multiple executions with the same parametrizations where outcomes correspond to distributions of agent states and observable quantities must be computed through averages. the choice of geometry presupposes that agents move across a common landscape at all times, and no agents enter or leave it. this simplification of the geographical landscape, while in general unrealistic, is not uncommon [93, 9] and provides two advantages: a) it naturally matches intuitions behind interactive particle systems driven by mass action principles [75] such as in the sir model and b) when translated into code, no boundary checks need to be performed by the agents. lattice sites are connected, from the perspective of an agent, by a moore neighborhood in an effort to reduce the effect of discretization artifacts [64] . we note here that our model presupposes a homogeneous population density as a means of ensure representativeness of processes within the geographical domain. although accounting for variable population density areas in the same scenario is possible, our approach simplifies implementation aspects and prevents artifacts for model outcomes that may be strongly density dependent [121] in both epidemiological and economic aspects. our model does not explicitly contain a rich representation of locations where agents are drawn to and act as temporal sinks. instead, we chose to model agent tropism through randomized agent dwell times. dwell time (τ dw ) refers here to the minimum amount of time an agent susceptible to covid-19 contagion needs to spend in a given location to acquire the 5 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. (which was not certified by peer review) virus. based on existing estimates [36] , we have chosen τ dw = 15 minutes, which corresponds to one time step t s , s ≥ 0. average total dwell time t dw = k dw · τ dw for an agent corresponds to the (integer) average number of steps an agent will dwell on a single location. since our model assumes a day as a usual reporting unit in decision-making activities, all parameters stated in days are internally rescaled to τ dw units (i.e. 1 day = 96 simulation steps). agent dwell times are set at creation time using a random deviate from poisson(k; λ = k dw ). to configure of a scenario, a collection of demographic and disease parameters must be specified. age structure appears to be strongly associated with differences in covid-19 fatality rates [19, 32, 37, 98] . the model requires estimates both of the distribution and observed fatality proportions p sex f and p age f per sex (i.e. male and female) and age (i.e. every ten-year intervals) respectively. co-morbidities are introduced in a similar manner by means of the age and sex structure of the population as a collection of positive multipliers k sex f , k age f per relevant condition, and aggregate clinical data about their prevalence per age and sex. the total number of agents n at the onset of the simulation remains constant at all times, except when an influx of new agents is simulated. to do so, the number of new agents entering the population correspond to a proportion p new of the existing ones. in addition, one must specify when the agents will be introduced t s = t new and how long it takes them to enter the space τ new . in this manner our model can account for seasonal population increases driven by, for instance, agricultural production or the start of semesters in university towns. after time t s = t new + τ new , the simulation will contain approximately n = n + p new · n agents. to account for population density ρ pop , the model specifies the width w and height h of the lattice that will contain the agents. we presuppose that agents move across the lattice one jump at a time if their dwell time is exhausted. the size of the lattice should be also adjusted based on the mobility and transportation patterns of individuals within the enclosed region of interest -that is, excluding realistic density fluctuations due to commuters that spend most of their time outside of the simulated region. the effect of such adjustment is equal to re-scaling the population density by a mobility pre-factor k mobl . after these calculations have been performed, the model should ensure for t new = ∞ that intuitively, the effect of greater mobility is equivalent to increased population density, or correspondingly, to traversing a smaller space. while our model does not include the effect of road networks or vehicle use, a carefully constructed average for k mobl can provide a sufficiently adequate approximation. disease-wise, the model comprises six critical parameters. first, the initial proportion of agents p iexp that are exposed to the disease. the model assumes that their introduction occurs at the onset of the disease incubation period. then, the incubation period τ incb and the recovery time τ recv are inputs correspond to average observed or estimated values in clinical 6 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . patients [69] . in the simulation, each agents is initialized with individual incubation and recovery times drawn from poisson(k; λ = τ incb ) and poisson(k; λ = τ recv ) respectively. our reasoning behind this choice rests on the fact that a) the time at which symptoms manifest across patients depends on a common organismal response to the pathogen dependent on the activation of known (and yet unknown) molecular pathways [2, 103, 111] , and at the same time on the intra-population variations that are found across individuals due to their specific genetic make-up and context. thus, we treat symptoms onset as a homogeneous poisson process for simplicity even when the described coupling exists. an improvement to our current model would entail, if the reasoning above holds, computing observables using a general poisson point process by assuming that the radon-nikodym density exists [29] . fourth, the proportion of individuals who remain asymptomatic p asym is accounted for and utilized when stochastically deciding the fate of exposed agents. the role of asymptomatic patients remains heavily investigated [11, 85] and appears to play a crucial role in disease mitigation for covid-19 [18, 33, 43, 122] . from literature data, the proportion of asymptomatic patients appears to vary greatly across countries and demographics (e.g. [4, 6] ), although the matter is far from settled. in this sense, our model is intended to be applied by using data starting at the most local level if possible, and only moving to larger geographical instances when data cannot be obtained by means of statistically robust antibody testing. fifth, the proportion of severe cases p sev is significant for the model due to its relation to hospital capacity. the definition of severity used here is that reported in [118] ; we suggest similar guidelines on estimating the proportion severe cases should be followed. another proxy for severity may comprise the number of non-icu and icu admissions, and their ratio [86] ; in general, the need of hospitalization implies that clinical evaluation of a patient raises enough concerns as to consider the possibility of transitioning from non-icu stage to the icu stage of care [94] . to account for saturation of health care services, the model males use of the proportion of beds proportional to population density p beds , and we assume that only severe cases are hospitalized. if a severe patient cannot be hospitalized due to saturation, then its probability of fatality rises by a factor to be determined empirically; for reference, our model presupposes a four-fold increase. at present, our model does not provide estimates of icu occupancy. finally, the model utilizes the probability of contagion p cont per agent per interaction. this quantity can be obtained from field data or other (more coarsed-grained) epidemiological models at the onset from estimates of the basic reproduction number r 0 by observing that where n s,i t is the average number of contacts between susceptible (s) and infectious (i) agents, and d is the duration of infectiousness of the disease. recalling the sir model 7 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . we observe that γ = d −1 and our view of r 0 is that of a preliminary estimate for initial calibration at the onset of the period of interest. for reporting purposes, we favor the effective reproductive number r(t), and consequently provide information about the observable for n s,i (t) such that since one agent represents multiple individuals in the region of interest, it becomes necessary to compensate for this renormalization process. for such purpose, we provide a scaling factor associated to the representativeness of the model k r given with a scale 1:r by we found empirically γ ≈ 1.58489, such that rescaling the probability of contagion to obtain p cont = p cont /k r leads to in addition, r(t) ∝ ρ pop (eq. 1), which implies that k mob must also modulate r(t). the final expression becomes at the model level, observables are interrogated across the agent population, agent step actions scheduled and the step number updated. at each step, agents move one space across the 2d torus after exhausting their dwell time per location. all agents posses an internal state σ that stores information relevant to the disease, its economic aspects and various control structures. their motion is driven by a random walk within their moore neighborhood, unless their state has been set to isolation. isolation means not moving across the space regardless of dwell times. more than one agent can inhabit one lattice site, which forms the basis for determining when a susceptible agent becomes exposed and the infectious cycle 8 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . starts. prior to performing stage-dependent computations from the disease perspective, agents compute the consequences of policy measures and adjust various elements of their internal states relevant to epidemiological and economic actions to follow. our model departs from the usual compartments of the sir and extends it in order to account for a more fine grained variety of significant infection stages. agent disease states are as follows: susceptible all agents (except those marked as initially exposed) start as the susceptible population. when susceptible agents share the same lattice site, these may come in contact with other exposed, symptomatic (i.e. due to voluntary or involuntary breaking of quarantine with probability 1 − p isoeff ) or undetected asymptomatic agents. if at least one agent is infectious, the agent changes its state σ to exposed as dictated by bernoulli(σ, p cont ). unless quarantined due to a policy measure, susceptible agents move freely across the lattice. exposed in the absence of any policy measures impacting exposed agents, these continue to explore lattice sites until their incubation period given by poisson(x, λ = τ incb ) is exhausted. at that time, agents become asymptomatic as dictated by bernoulli(σ, p asym ) or symptomatic detected, otherwise. asymptomatic asymptomatic agents continue moving through the space and remain infectious until the recovery period given by poisson(x, λ = τ recv ) is exhausted. at that point, the agent enters the population of those recovered. symptomatic when an agent becomes symptomatic, it is immediately marked as detected and quarantined. regardless of stringency of testing policies, the definition of confirmed case depends at the minimum on being both symptomatic and a positive identification via some form of testing (i.e. rt-pcr qualitative or quantitative, serological [91] ). symptomatic agents follow two possible trajectories. in the first one, agents convalesce without becoming severe until their recovery time is exhausted an recuperate. in the second one, agents become severe as dictated by bernoulli(σ, p sev ). in terms of the impact of saturation of health care services, research is needed to determine lethality of patients outside hospitals and other facilities. however, using existing ethical guidelines that provide heuristics of fair resource allocation for beds and ventilator equipment [39] as a proxy, we estimate that lethality increases (conservatively) by at least a factor of 5. asymptomatic detected asymptomatic agents detected by some widespread systematic testing strategy are immediately quarantined in place, and wait until their recovery time is exhausted. at that point, they join the population of recovered agents. . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . severe agents that enter the severe stage represent patients that require some form of hospitalization, and for some of them, use of mechanical ventilators; these remain under perfect quarantine. lethality, computed per age and sex population structures as p f = p age f · p sex f is used to determine whether a severe agent becomes a fatality as given by bernoulli(σ, p f ). agents, on the other hand, may survive until recovery. recovered agents that have recovered leave quarantine and move again freely across the lattice. our model does not consider the probability of re-infection, but this may need to be included in the future [92] . at the moment of writing, this aspect of covid-19 remains speculative and uncertain for human patients [34, 90, 115] despite encouraging evidence obtained from experiments using rhesus macaques [14] . deceased agent that count as fatalities do not interact with other agents or undergo any further epidemiological significant events until the end of the simulation. in order to increase model realism, we consider the effect of inbound infectious cases of two sorts: people that live within the community but have to travel and work outside of it in a steady stream daily that maintains the overall population density stable, and people that move seasonally within the community, potentially bringing in more cases that have a different viral load. this last case describes, for example, the opening of university campuses for instruction where several people relocate to adjacent towns. for the first situation, the model includes the probability of susceptible people becoming infected with a daily rate that determines the infectious stage depending on bernoulli(x, r ibnd ). in the second type of infectious cases, at time t mass a proportion p mass of the current population will enter the simulation space during a time period τ mass with a probability of being in the exposed state of p mass . covid-19 has forced a frantic search for public policy measure combinations capable of containing viral transmission, and ideally quelling its progression altogether [13, 42, 47, 56, 67, 71, 89, 99, 106, 115, 116, 120] . all measures reviewed and emerging across literature roughly belong to four main classes of measures: 1) those that aim to reduce at any instant the density of individuals at locations with potentially high concentration of people by means of imposed self-isolation of non-essential workers and cancellation of activities involving massive amounts of people, 2) those that reduce the likelihood of viral exposure for those qualified as essential workers for whom close social interactions are inescapable, 3) those intended to detect and isolate positive virus carriers through application of molecular or serological testing and 4) those that seek to reconstruct interaction histories in which a positive infectious patient may have had an active role in unknowingly spreading the disease. it has become increasingly clear that these measures are essential yet hard to sustain for long periods of time. on the one hand, various degrees of negative psychological impact have 10 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . been recently reported [31] , in particular impacts that decrease adherence to public policy measures [59] which are expected to naturally arise after periods of prolonged confinement under a collective crisis; world war ii critics of air raid shelter policies constitute a significant precedent [80] . on the hand, mounting concerns on lasting economic impacts [76, 83] materialize the wickedness of the covid-19 pandemic and the cost of measures to address it [22] , concerns that which include social protection of workers [44] , the labor market [30] , patterns of work [65] , gender equality [7] , nation-state economics [74] , and monetary policy [21] to name a few. motivated by the latter, our work attempts to model the individual variability expected when these types of measures are implemented, their various impacts in terms of disease and economy collective observables, and the potential outcomes of combining them in various manners. we note that societal and economic impacts of covid-19 differ from those in other pandemics due to the tight coupling of global events and the effect of near-instantaneous digital communication. we are changing the pandemic while living it. while our model does not provide mechanisms to state the associated control problem in cybernetics terms, emerging literature (e.g. [35, 48, 87, 119] ) suggests that such approach may be possible, and even essential to provide solutions that account for the complexities involved in politically and socially driven environments. self-isolation in our model is captured by establishing a period in which a proportion p isol of agents in mobile states (i.e. susceptible, exposed, asymptomatic) remains at a fixed location for a well-defined period of time. self-isolation starts at time t isol and extends for a period τ isol , after which motion across space is restored. agents isolate with effectiveness p isoeff , representing the probability that when in contact with another infective agent the final probability of contagion becomes p cont = p isoeff · p cont . social distancing is modeled as a distance-dependent adjustment constant δ( ) that adjusts the probability of contagion p cont depending on linear distance assumed between agents within the same cell such that p cont = δ( ) · p cont . based on recent experiments on the effect of air turbulence on droplet dispersion [24] , we assume a decreasing sigmoid profile after 1.5 meters. hence, 11 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . where k is a constant that adjust the decrease rate of the sigmoid function. for the purposes of the covid-19 model, k = 10.0. the value of can be adjusted also to account for the effect of other interventions that decrease contagion probability per contact by decreasing the effective viral load, such as the use of various types of face masks [62] . similar to self-isolation, testing in the simulation operates by distributing a target percentage of the population into a given period. susceptible and asymptomatic agents can be tested; in our simulation, we do not re-test those who recover. a symptomatic case is treated immediately as tested, representing the case where a patient reaches a health provider and a test is applied to determine the correspondence between symptoms and the disease. testing proceeds in the following manner. once time t test is reached, symptomatic and asymptomatic agents are selected with a probability p test proportional to the period τ test . this testing process simulates massive testing policies without any statistical design or underlying population structure. we simulate forms of automated contact tracing by means of a set of known prior contacts. we assume a delay of two days for contact follow-up once an infected patient has been discovered. once an agent is marked as positive, all of its contacts are evaluated and classified either as susceptible (negative), symptomatic or asymptomatic detected. contact tracing utilizes the same start time and period as testing. along with an epidemiological model, we have included economic factors tied to disease stages. after some investigation around statistical theories in economics [27] , we decided to implement a simple model where value creation in terms of exchanges between money and products or services [84] are computed in connection with the progression of the disease. our model is an oversimplification of economic systems, and as such, its goal is to provide a numerical intuition about the immediate effects on the accumulation of capital by individuals and the public sector during the period of interest. thus, the economic model makes no assumptions about wealth distribution, wealth inequality or other societal factors, and as such only aims at portraying the impact of an epidemic on transactional capital gains or losses at the private and public levels. regarding public value [81] , we observe that the complex web of actions across individuals and institutions make construction of a detailed model expensive in connection with infectious disease dynamics. growing literature on the subject is indicative of the latter [25, 53, 78, 112] . to the best of our current experience, proper treatment of the current situation would require a different class of model based on fractional operators coupling 12 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . https://doi.org/10.1101/2020.07.22.20159798 doi: medrxiv preprint epidemiological [1] and econometric aspects [109, 73] capable of accounting for short-and long-term memory macroeconomic effects [108] . our take on the matter can be stated through the following somewhat intuitive principles: 1. disease stages that allow interactions should cause non-linear positive externalities in terms of collectively amplified public value. the more frequent interaction exchanges, the higher the materialization of public value as a function of non-linear effects of reaching throughput efficiencies that both maximize economies of scale and translate into deep capital accumulation and redistribution across the public sector. 2. disease stages that forbid interactions but do not put individual lives at risk should cause linear negative externalities associated to both increased unemployment ripple effects and inability to reach throughput thresholds capable of amplifying value creation. 3. disease stages that both forbid interactions and put individual lives at risk should cause non-linear negative externalities due to increased unemployment ripple effects, inability to reach throughput thresholds capable of amplifying value creation and the saturation of alternatives under an increasingly severe public health crisis. 4. in addition, we estimate the cost of performing one test as part of the public cost. the purpose of this observable is to provide an account of testing as a public measure versus other actions for which their cost is harder to account for. in order to materialize the above principles, our approach is inspired by that of inputoutput matrices utilized to account for combined economy-ecosystem calculations [52] . the equivalent of the matrix m in our model expresses economic outputs per disease stages. viewed as a matrix computation, the components of the input vector u correspond to proportions of the agent population per infectious stage, while the output vector v is of the form v = (v priv , v pub ). depending on the disease stage, vector components are computed aggregating the value of interactions or individually. by a suitable homotopic transformation t [88], we approximate non-linear effects of market dynamics, thus the final value of v becomes t [v] = (v α priv priv , v α pub pub ). the matrix components m ij and both α priv and α pub are model inputs. our model captures during its execution various observables every time step. all values are aggregates and no particular agent information is stored; in the future, this can be of value if the model is extended with network information. after a simulation has completed, the following classes of observables are recorded in a csv format: . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. step number, population size disease (fraction) susceptible, exposed, asymptomatic, symptomatic quarantined, asymptomatic quarantined, severe, recovered, diseased measures self-isolated, tested, traced epidemiology effective reproductive number economics cumulative private value, cumulative public value the epidemiology workbench is implemented using python (v 3.6) using the mesa agentbased simulation library [77] . in batch processing mode, our model receives a json file with all the parameters described above and a number indicating the number of cores to use during the simulation. once a sanity check is performed, the parameters are used in conjunction with the multiprocessing batch running features provided by mesa. we also provide a parameterizable web-based dashboard to explore individual runs. the objective behind this corresponds to enabling decision-making users to progressively gain intuitions behind each parameter, and not to provide operational information. our code is openly accessible through github 1 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . cases can provide information leading to estimates, or values provided by trusted sources at the most local level possible can be used. incubation time is likely to be blurred by multiple confusion variables captured in our model as a poisson distribution. despite its stochasticity, the model supposes that large population sizes group tightly towards a mean value. this assumption may need to be revised retrospectively later. for the proportion of asymptomatic patients, contact tracing and structured antibody testing can provide specific local information. as a general calibration protocol, we suggest the following steps: 1. adjust grid size based on fixed population density until r 0 matches the best known value for the area in the first days of the model (steps = 96) and > 30 runs. population density in the model loosely includes average mobility patterns, and cell sizes reflect the distance traversed every 15 minutes. also -but not recommended-the probability of contagion may be used to calibrate. this may imply unknown population conditions and should be used only to test hypotheses about individual variations that manifest in the ability of covid-19 to spread. 2. execute the model to the point where the number of symptomatic agents corresponds to one representative agent. for example, with an abm of 1000 agents and a population of 100k individuals, the critical infected agent-to-population ratio is 1:100. use the point in time rounded to the nearest integer day as point of departure for policy measures. in practice, this implies executing the model in excess of as many days as the longest known incubation period. 3. if policy measures have been introduced, use date above as the reference point for their introduction. to develop scenarios, we strongly recommend starting from the most recently calibrated model that includes policies as well as using a model without measures as a basis for counterfactual arguments. the cities of urbana (pop. 42,214) 2 and champaign (pop. 88,909) 3 comprise a population of approximately 132,000 inhabitants. figure 1 describes the current percent distribution per age group. in addition, distribution per sex is 50.5% males and 49.5% females. however, the population across both cities undergoes a seasonal decrease on mid may due to the end of the spring semester and an increase in early august with the start of the fall semester in the university of illinois at urbana-champaign. from the more than 50,000 students enrolled on campus on may 4 , we estimate that around 30,000 leave during summer to return for the fall semester. effectively, the summer population amounts to around 100,000 inhabitants. we used covid-19 age-and sex-dependent mortality values as reported by cdc until june 10. covid-19 data was obtained from the champaign-urbana public health district (cu-phd) 5 and cdc for mortality data 6 . sex-dependent mortality was established at 61.8% for males and 38.2% for females. the first local case in the community was reported on march 8, and state-wide shelter-in-place measures were applied on march 21 7 . later on, mandatory mask usage was established on may 1 8 . by april 21, cumulative cases had reached 0.1% of the population, and by july 8 it had increased to 1%. the local r(t) value at the peak cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . period between april 21 and may 17 reached an estimated value of 1.2; after these measures, it remained around 0.91. we used a probability of contagion per interaction of 0.004 every 15 minutes if there is at least one person infected at the same location as the susceptible one. this value, although computed from data, appears to reflect compliance with various sanitation practices among the population. based on the fluctuation in local data, we have estimated an inbound probability of 0.0002 new cases due to members of the community becoming exposed elsewhere. cu-phd performs strict contact tracing across all cases, hence for all simulations we assume contact tracing remains active across all simulations. google mobility data were used to estimate the average effective shelter-in-place value to a 45% of the population with 0.8% efficacy of self-isolation. the severity was similarly estimated at 5% based on local case information. similarly, a starting value of r(t) = 1.2 corresponded to a grid of 190 × 225 cells and 1000 agents and k mob ≈ 0.4781; this last value appears to be consistent with the regular use of public transport in the area and local observed shelter-at-home patterns. at the start of the simulation, the agent-to-inhabitant ratio is approximately 1:100. hence, the calibration starts with one exposed agent. differences between r(t) in april and july correspond to = 1.89. intuitively, this implies that the effect of wearing a mask may roughly translate into increasing the distance 0.39 meters beyond what who recommends for proper social distancing. however, this cannot and should not be interpreted as to relax mask usage in any manner. in regard to asymptomatic patients, we have established a conservative value of 35% based on prior studies [49] . the following sequence of events was assumed toward the start of classes this fall: our goal for simulating the impact of mass ingress was to determine, based on various measures, the viability of preserving public health under variations of measures a week after a significant portion of the population has been tested. to determine the latter, we obtained data for r(t), symptomatic, asymptomatic, severe and deceased fractions of the population. in addition, we collected information about economic impact using our input-output matrix model. a total of 6 simulations explore the following parameter settings: . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . 1. shelter-in-place continued/removed on day 117, 2. massive testing performed to 25%, 50% and 75% of the population in the enlarged community. we also computed a counterfactual case corresponding to no massive testing and lifting of shelter-at-home orders to compare against a backdrop without measures. model calibration results (figure 2 ) correspond to the parametrization publicly available at the github project repository 9 . we computed a total of 7 scenarios (shelter-at-home times testing levels plus counterfactual), each one with an ensemble of 30 independent runs. execution of these scenarios was performed on amazon ec2 infrastructure using a c5a.8xlarge non-dedicated instance with 36 processors and 64 gb ram. we used an ubnutu 18.04 86 image as the choice of operating system. calibration cpu time with an ensemble of similar size for the first 40 days is, on average, 16.3±0.4 minutes, and the average execution time of a complete scenario is 65.2 ± 1.3 minutes. preliminary profiling indicates that random number generation using the scipy library [114] explains most of the execution time. no attempts were made to further speed up our code by compiling it using cython [102] or numba [68] . our goal in these simulations was not to reproduce exactly the case curves observed in the community, but to obtain a picture that remains quantitatively and qualitatively rigorous. confidence intervals are calculated at 95% when present. we provide scripts to automate the setup 9 see: https://bit.ly/2bhcuv3 18 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . of the amazon ec2 instance with model installation 10 , the execution of all scenarios 11 and their visualization 12 for reproducibility purposes. the following convention is applied to all figures: dark red corresponds to 25% testing, teal to 50% and dark blue to 75%; the counterfactual case is colored in purple. a dotted line corresponds to the start of mass ingress, a dot-dash line to the end of mass ingress and the start of massive testing, and a dashed line to the end of massive testing. all figures related to disease stages start at day 80. model results indicate that outbound exposed individuals coupled with local fluctuations appear to drive the behavior of this pandemic for the cities of urbana and champaign. the combination of contact tracing and public health management by cu-phd, compliance with health and sanitary measures and rapid implementation of shelter-in-place measures have prevented the pandemic from escalating in the region. considering both cities as a closed system, adequate health management appears to be ultimately responsible for the small number of severe cases and hospitalizations in the region. the effect of masks, based on the information obtained from our simulation, appears to have a significant effect on the value r(t) according to fig. 2 . our simulation of the reopening of the university of illinois local campus with a higher viral load suggests that an increase in cases should be observable independent of the testing regime or relaxation measures. the future impact of this increase, however, is not. figure 3 indicates that lifting all shelter-at-home restrictions and performing testing has a significant impact regardless of the testing level, while preserving shelter-at-home measures along with any testing level can drastically sustain r(t) slightly below pre-mask order values. we note that r(t) decreases in all cases, which can be explained by contact tracing-based testing. this suggests that even hybrid education modes (in presence + online) constitute significantly better alternatives than full campus reopening. testing intensity matters. in terms of the outcome after day 137, testing intensity determines the last value of active cases during two weeks after testing. note that even when the testing instant itself has passed, capturing a larger and larger number of positive cases (particularly asymptomatic ones) drastically reduces the infectious population (figure ??) . even when shelter-at-home measures have been lifted, testing reduces the fraction of the population classified as active cases at least higher but close to its value prior to mass ingress (25%), roughly equal its prior value (50%) or lower than the prior value (75%). lifting shelter-at-home measures has a significant impact on the magnitude of both the peak of active cases and the value after two weeks have passed since testing occurred. plans to test the student population once per week at scale, although not simulated here, appear to be a most effective solution to further tame the curve. . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . the main mechanism massive testing addresses in general, according to our simulations, is the removal of infectious individuals from the population, in particular those who are asymptomatic. in general, the estimated proportion of asymptomatic patients is a significant driver of the contagion in our model. when the population increases, many more individual contacts are possible within the same geographical area, and the lag induced by the incubation time translates into observing the impact of testing at least a week later. figure 5 compares the asymptomatic fraction of the population across a baseline simulation without massive testing or some degree of sheltering. as in the previous case, shelter-inplace measures have an effect on the growth rate of the asymptomatic population, but even a testing intensity of 25% appears to lower it significantly compared to the counterfactual case. at case severity of 5%, more hospitalizations may be expected six days after day 130 (mass ingress), particularly if shelter-at-home orders have been lifted ( figure 6 ). the impact of testing, while it cannot be fully distinguished due to the overlap of confidence intervals in our simulations our analysis of economic impact focuses on two per-capita averages: cumulative private value (without any egress) and cumulative public value. while this part of our proposed model is experimental and requires further analysis, we proceed to state the current results. first, we studied the effect of the pandemic only on individual accumulation of private value (figure 7) . mass ingress appears to renormalize temporarily the distribution and removing shelter-at-home measures, predictably, increases the final value at day 153, but only by approximately five units. all testing levels appear to comprise a relatively tight bundle, which can be interpreted either as an artifact due to the model being simplistic, or as the fact that the impact of testing levels on cumulative private value in the context of a 21 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. pandemic under control (as in urbana and champaign cities) is limited. even when these differences are small, they do exist. testing at low levels (i.e. 25% and 50%) reduces the number of isolated people compared to testing at a broader scale (i.e. 70%). however, when an epidemic process remains under control, public health benefits largely appears overshadow individual losses, contingent on the validity of this approach. in the case of per-capita cumulative public value, the model predicts negative outcomes even for an epidemic under control (figure ) . this appears to align with public expenditure needed to mitigate any economic and societal lockdown necessary to stop the spread of the disease, as well as the negative externalities leading to less public transactions taking place on 22 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. the copyright holder for this preprint this version posted july 25, 2020. . https://doi.org/10.1101/2020.07.22.20159798 doi: medrxiv preprint systems that were designed to support a certain minimum load to remain profitable: public finances tend to be, in general, inelastic for that reason. when shelter-at-home measures are lifted, a natural order of solutions arises from worst-case (our counterfactual, purple) to best-case (75% testing, blue): strong testing reduces the long-term impact of active, asymptomatic and severe cases. in this situation, however, the effect of mass ingress appears to be to re-bundle the behavior as r(t) increases again. if shelter-at-home is preserved, an interesting situation arises: doing nothing appears to be a good economic solution. we believe the main reason behind this result to be that, for the case modeled here, the social and economic cost of the lockdown is higher due to the situation being under control; however, the material gains computed by the model between are rather small, and preserving public health over economics is a better-long term strategy. despite this, strong massive testing still provides a next best solution from the point of view of economics, and the best strategy from the public health perspective; this is evidenced by the diverging curves in fig. 8(b) . we speculate, based on our current simulation outcomes, that the ordering in the economic cost profile of a pandemic during its exponential phase should be similar to that of fig. 8(a) , but with the divergence observed in (b). we reported here the construction of an agent-based workbench using the mesa modeling framework capable of capturing epidemic processes alongside public policy measures. the model is fully stochastic, entailing the computation of observables of different kinds. while computationally expensive, its formulation allows to easily obtain quantities that appear to be useful in the process of combating an epidemic. we applied our workbench to understanding the possible epidemiological profile of two cities, urbana and champaign, in the context 23 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. (which was not certified by peer review) the copyright holder for this preprint this version posted july 25, 2020. . of the reopening of the university of illinois local campus next fall. our simulations indicate that at least 50% testing of the local population is needed to sustain the pressure of mass ingress of individuals with a higher viral load compared to the local one. more generally, contemporary management of an epidemic demands changing the mode of interaction across as much as the population as possible from those requiring physical proximity to those that do not. although digital technologies provide mechanisms to preserve safe spatial distancing, temporal distancing can also be intelligently used to reduce the probability of contagion. in terms of economics, public health measures must be privileged over financial concerns, since the panorama appears similarly bleak during the early phases of an epidemic, and strict measures possibly provide the best solution during the exponential phase. our model contains the following key limitations. first, only one measure per type can be specified at the moment, instead of a sequence of dates paired with values corresponding to the measured (or expected) effect of measures of the same type. in the example above, we used an approximation of shelter-in-place for the entire simulation period (april 21-september 12), even when the state of illinois ordered phase 4 re-openings on june 26. another critical element missing in our model corresponds to preferential shelter-in-place per age group. even when mortality appears to more strongly impact the elderly, local mortality is low. this may be due to higher compliance of that population with shelter-athome and other sanitary measures, including wearing masks in public. variable viral loads per disease stage [54] are missing in our model, but these are harder to calibrate due to the biosafety and time elements involved in quantitative pcr-rt. nevertheless, we do foresee situations where this may be possible and pertinent. finally, our model assumes agents have an effectively infinite memory to remember whom they have had contact with. contact tracing has a stringent limit when performed manually, which can be expanded greatly by means of various information technologies. hence, our model is not realistic in this sense, since it does not distinguish between these two cases: when an epidemic has reached certain critical mass, active cases will be underestimated. computationally, the epidemiology workbench is limited by the lack of true concurrency in mesa, thereby impacting scaling properties of our simulation. distributing the agents across multiple compute nodes in python requires architectural changes in mesa beyond the scope of our work. at present, efforts are under way to develop an elixir-based abm platform capable of addressing this limitation as part of the spec collaboration in the computational social science community. another critical bottleneck is random number generation, for which various strategies may be applied, including the use of (approximately) irrational numbers as coefficients of fourier series. a final aspect underscored by our research, in particular in the context of covid-19, is the need for anticipatory mechanisms driving public policy measures. in this sense, simulation methods such as the one presented here or others lack convenience when introducing external events and the execution cost of recomputing complex models, and appear to make sense only at early stages of a epidemic process: once the disease spread has reached its exponential phase, the need moves from prediction to probabilistically qualified estimations of short term measure effectiveness. this suggests a different class of stochastic methods that 24 . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. (which was not certified by peer review) the copyright holder for this preprint this version posted july 25, 2020. . not only predict expected trends but also recommend measures based on their effectiveness, similar to those used in high-speed trading of financial derivatives and futures. the complexity of the socio-technical character of the global economy and society demands these more powerful methods to successfully address the wicked character of a pandemic, in particular this one. as of now, our efforts concentrate on seeking viable ways to package and deploy the epidemiology workbench across various cyberinfrastructure resources in order to make it available to other small cities (including training resources) and, particularly, to communities with strong presence of underrepresented minorities whose public health planning resources are heavily constrained. . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. (which was not certified by peer review) the copyright holder for this preprint this version posted july 25, 2020. . https://doi.org/10.1101/2020.07.22.20159798 doi: medrxiv preprint . cc-by-nd 4.0 international license it is made available under a is the author/funder, who has granted medrxiv a license to display the preprint in perpetuity. (which was not certified by peer review) the copyright holder for this preprint this version posted july 25, 2020. . https://doi.org/10.1101/2020.07.22.20159798 doi: medrxiv preprint on a comprehensive model of the novel coronavirus (covid-19) under mittag-leffler derivative the possible immunological pathways for the variable immunopathogenesis of covid-19 infections among healthy adults, elderly and children special report: the simulations driving the world's response to covid-19 investigating the impact of asymptomatic carriers on covid-19 transmission. medrxiv comparing large-scale computational approaches to epidemic modeling: agent-based versus structured metapopulation models asymptomatic coronavirus infection: mers-cov and sars-cov-2 (covid-19) the impact of covid-19 on gender equality rural america and coronavirus epidemic: challenges and solutions landscape epidemiology modeling using an agent-based model and a geographic information system elementary proof of convergence to the mean-field model for the sir process presumed asymptomatic carrier transmission of covid-19 covid-induced economic uncertainty covid-19: emerging protective measures lack of reinfection in rhesus macaques infected with sars-cov-2. biorxiv sectoral effects of social distancing. available at ssrn deterministic and stochastic models for recurrent epidemics agent-based derivation of the sir-differential equations covid-19: the case for health-care worker screening to prevent hospital transmission. the lancet effects of latency and age structure on the dynamics and containment of covid-19. medrxiv agent-based modeling: methods and techniques for simulating human systems. proceedings of the national academy of sciences fighting the covid-19 emergency and relaunching the european economy: debt monetization and recovery bonds the economy on ice: meeting the economic challenges during and after the covid-19 crisis geographical tracking and mapping of coronavirus disease covid-19/severe acute respiratory syndrome coronavirus 2 (sars-cov-2) epidemic and associated events around the world: how 21st century gis technologies are supporting the global fight against outbreaks and epidemics turbulent gas clouds and respiratory pathogen emissions: potential implications for reducing transmission of covid-19 the economic ripple effects of covid-19. unpublished manuscript. available at the world bank development policy and covid-19-eseminar series comparative genetic analysis of the novel coronavirus (2019-ncov/sars-cov-2) receptor ace2 in different populations statistical theories of income and wealth distribution. economics: the open-access the sir epidemic model from a pde point of view stochastic geometry and its applications labor markets during the covid-19 crisis: a preliminary view social psychological measurements of covid-19: coronavirus perceived threat, government response, impacts, and experiences questionnaires cmmid covid-19 working group, et al. age-dependent effects in the transmission and control of covid-19 epidemics. medrxiv covid-19: identifying and isolating asymptomatic people helped eliminate virus in italian village covid-19-new insights on a rapidly changing epidemic optimal covid-19 epidemic control until vaccine deployment. medrxiv how to restart? an agent-based simulation model towards the definition of strategies for covid-19" second phase demographic science aids in understanding the spread and fatality rates of covid-19 model transparency and validation: a report of the ispor-smdm modeling good research practices task force-7 fair allocation of scarce medical resources in the time of covid-19 the effectiveness of moral messages on public health behavioral intentions during the covid-19 pandemic impact of non-pharmaceutical interventions (npis) to reduce covid19 mortality and healthcare demand the global community needs to swiftly ramp up the response to contain covid-19 asymptomatic transmission, the achilles' heel of current strategies to control covid-19 social protection and jobs responses to covid-19: a real-time review of country measures us city-size distribution and space ethics, information technology, and public health: duties and challenges in computational epidemiology face masks for the public during the covid-19 crisis optimal quarantine strategies for covid-19 control models spread of sars-cov-2 in the icelandic population variation in government responses to covid-19 modeling targeted layered containment of an influenza pandemic in the united states ecological pricing and economic efficiency combining behavioral economics and infectious disease epidemiology to mitigate the covid-19 outbreak temporal dynamics in viral shedding and transmissibility of covid-19 wicked problems in public policy quantifying the impact of physical distance measures on the transmission of covid-19 in the uk threshold behaviour of a stochastic sir model an older population increases estimated covid-19 death rates in rural america measuring two distinct psychological threats of covid-19 and their unique impacts on wellbeing and adherence to public health behaviors health security capacities in the context of covid-19 outbreak: an analysis of international health regulations annual report data from 182 countries. the lancet projecting the transmission dynamics of sars-cov-2 through the postpandemic period aerosol filtration efficiency of common fabrics used in respiratory cloth masks computer-assisted decision making: performance, beliefs, and the illusion of control moore and more and symmetry the impact of covid-19 on us employment and hours: real-time estimates with homebase data coping with the wickedness of public policy problems: approaches for decision making under deep uncertainty effect of non-pharmaceutical interventions to contain covid-19 in china numba: a llvm-based python jit compiler the incubation period of coronavirus disease 2019 (covid-19) from publicly reported confirmed cases: estimation and application the complexities of agent-based modeling output analysis. the journal of artificial societies and social simulation scientific and ethical basis for social-distancing interventions against covid-19. the lancet. infectious diseases long-time behavior of a stochastic sir model fractional state space analysis of economic systems the cost of covid-19: a rough estimate of the 2020 us gdp impact. special edition policy brief exploring synthetic mass action models covid-19, risk, fear, and fall-out mesa: an agent-based modeling framework 3 the economic impact of covid-19 the global macroeconomic impacts of covid-19: seven scenarios air raid shelter policy and its critics in britain before the second world war public value inside: what is public value creation? correlation between universal bcg vaccination policy and reduced morbidity and mortality for covid-19: an epidemiological study the world after covid-19 and its impact on global economy the quantity theory of the value of money estimating the asymptomatic proportion of coronavirus disease 2019 (covid-19) cases on board the diamond princess cruise ship, yokohama, japan projecting hospital utilization during the covid-19 outbreaks in the united states controlling the transmission dynamics of covid-19 generalized degree and nonlinear problems quarantine alone or in combination with other public health measures to control covid-19: a rapid review the covid-19 pandemic in the us: a clinical update world health organization et al. laboratory testing for coronavirus disease (covid-19) in suspected human cases: interim guidance will we see protection or reinfection in covid-19? abloom: location behaviour, spatial patterns, and agent-based modelling intensive care management of coronavirus disease 2019 (covid-19): challenges and recommendations. the lancet respiratory medicine structural urbanism contributes to poorer health outcomes for rural america heterogeneity and network structure in the dynamics of diffusion: comparing agent-based and differential equation models moral decision-making and mental health during the covid-19 pandemic estimating the infection and case fatality ratio for coronavirus disease (covid-19) using age-adjusted data from the outbreak on the diamond princess cruise ship covid-19 epidemic in switzerland: on the importance of testing, contact tracing and isolation extending the sir epidemic model modelling sir-type epidemics by odes, pdes, difference equations and cellular automata-a comparative study fast numerical computations with cython covid-19 infection: the perspectives on immune responses forecasting models for coronavirus disease (covid-19): a survey of the stateof-the-art mathematical modeling of infectious disease dynamics only strict quarantine measures can curb the coronavirus disease (covid-19) outbreak in italy covid-19: epidemiology, evolution, and cross-disciplinary perspectives long and short memory in economics: fractional-order difference and differentiation fractional calculus in economic growth modeling. the portuguese case the benefits and costs of using social distancing to flatten the curve for covid-19 molecular basis of covid-19 relationships in different species: a one health perspective. microbes and infection susceptible-infected-recovered (sir) dynamics of covid-19 and economic impact stability of a stochastic sir system scipy 1.0: fundamental algorithms for scientific computing in python positive rt-pcr test results in discharged covid-19 patients: reinfection or residual? can we contain the covid-19 outbreak with the same measures as for sars? epidemic waves of a spatial sir model in combination with random dispersal and non-local dispersal estimating clinical severity of covid-19 from the transmission dynamics in wuhan, china a cybernetics-based dynamic infection model for analyzing sars-cov-2 infection stability and predicting uncontrollable risks. medrxiv taking the right measures to control covid-19. the lancet infectious diseases role of density and field in spatial economics. contemporary issues in urban and regional economics covid-19 transmission through asymptomatic carriers is a challenge to containment. influenza and other respiratory viruses stochastic sir model with jumps the authors wish to thank the following individuals for their contributions and support: key: cord-340827-vx37vlkf authors: jackson, matthew o.; yariv, leeat title: chapter 14 diffusion, strategic interaction, and social structure date: 2011-12-31 journal: handbook of social economics doi: 10.1016/b978-0-444-53187-2.00014-0 sha: doc_id: 340827 cord_uid: vx37vlkf abstract we provide an overview and synthesis of the literature on how social networks influence behaviors, with a focus on diffusion. we discuss some highlights from the empirical literature on the impact of networks on behaviors and diffusion. we also discuss some of the more prominent models of network interactions, including recent advances regarding interdependent behaviors, modeled via games on networks. jel classification codes: d85, c72, l14, z13 how we act, as well as how we are acted upon, are to a large extent influenced by our relatives, friends, and acquaintances. this is true of which profession we decide to pursue, whether or not we adopt a new technology, as well as whether or not we catch the flu. in this chapter we provide an overview of research that examines how social structure impacts economic decision making and the diffusion of innovations, behaviors, and information. we begin with a brief overview of some of the stylized facts on the role of social structure on diffusion in different realms. this is a rich area of study that includes a vast set of case studies suggesting some important regularities. with that empirical perspective, we then discuss insights from the epidemiology and random graph literatures that help shed light on the spread of infections throughout a society. contagion of this form can be thought of as a basic, but important, form of social interaction, where the social structure largely determines patterns of diffusion. this literature presents a rich understanding of questions such as: "how densely connected does a society have to be in order to have an infection reach a nontrivial fraction of its members?," "how does this depend on the infectiousness of the disease?," "how does it depend on the particulars of the social network in place?," "who is most likely to become infected?," and "how widespread is an infection likely to be?," among others. the results on this apply beyond infectious diseases, and touch upon issues ranging from the spread of information to the proliferation of ideas. while such epidemiological models provide a useful look at some types of diffusion, there are many economically relevant applications in which a different modeling approach is needed, and, in particular, where the interaction between individuals requires a game theoretic analysis. in fact, though disease and the transmission of certain ideas and bits of information can be modeled through mechanical or purely probabilistic sorts of diffusion processes, there are other important situations where individuals take decisions and care about how their social neighbors or peers behave. this applies to decisions of which products to buy, which technology to adopt, whether or not to become educated, whether to learn a language, how to vote, and so forth. such interactions involve equilibrium considerations and often have multiple potential outcomes. for example, an agent might care about the proportion of neighbors adopting a given action, or might require some threshold of stimulus before becoming convinced to take an action, or might want to take an action that is different from that of his or her neighbors (e.g., free-riding on their information gathering if they do gather information, but gathering information him or herself if neighbors do not). here we provide an overview of how the recent literature has modeled such interactions, and how it has been able to meld social structure with predictions of behavior. there is a large body of work that identifies the effects of social interactions on a wide range of applications spanning fields: epidemiology, marketing, labor markets, political science, and agriculture are only a few. while some of the empirical tools for the analysis of social interaction effects have been described in block, blume, durlauf, and ioannides (chapter 18, this volume) , and many of their implementations for research on housing decisions, labor markets, addictions, and more, have been discussed in ioannides (chapter 25, this volume), epple and romano (chapter 20, this volume), topa (chapter 22, this volume) , fafchamps (chapter 24, this volume), jackson (chapter 12, this volume), and munshi (chapter 23, this volume), we now describe empirical work that ties directly to the models that are discussed in the current chapter. in particular, we discuss several examples of studies that illustrate how social structure impacts outcomes and behaviors. the relevant studies are broadly divided into two classes. first, there are crosssectional studies that concentrate on a snapshot of time and look for correlations between social interaction patterns and observable behaviors. this class relates to the analysis below of strategic games played by a network of agents. while it can be very useful in identifying correlations, it is important to keep in mind that identifying causation is complicated without the fortuitous exogenous variation or structural underpinnings. second, there are longitudinal studies that take advantage of the inherent dynamics of diffusion. such studies have generated a number of interesting observations and are more suggestive of some of the insights the theoretical literature on diffusion has generated. nonetheless, these sorts of studies also face challenges in identifying causation because of potential unobserved factors that may contemporaneously influence linked individuals. the empirical work on these topics is immense and we provide here only a narrow look of the work that is representative of the type of studies that have been pursued and relate to the focus of this chapter. studies that are based on observations at one point of time most often compare the frequency of a certain behavior or outcome across individuals who are connected as opposed to ones that are not. for example, glaeser, sacerdote, and scheinkman (1996) showed that the structure of social interactions can help explain the cross-city variance in crime rates in the u.s.; bearman, moody, and stovel (2004) examined the network of romantic connections in high-school, and its link to phenomena such as the spread of sexually transmitted diseases (see the next subsection for a discussion of the spread of epidemics). such studies provide important evidence for the correlation of behaviors with characteristics of individuals' connections. in the case of diseases, they provide some direct evidence for diffusion patterns. with regards to labor markets, there is a rich set of studies showing the importance of social connections for diffusing information about job openings, dating back to rees (1966) and rees and schultz (1970) . influential studies by granovetter (1973 granovetter ( , 1985 granovetter ( , 1995 show that even casual or infrequent acquaintances (weak ties) can play a role in diffusing information. those studies were based on interviews that directly ask subjects how they obtained information about their current jobs. other studies, based on outcomes, such as topa (2001), conley and topa (2002) , and bayer, ross, and topa (2008) , identify local correlations in employment status within neighborhoods in chicago, and consider neighborhoods that go beyond the geographic but also include proximity in other socioeconomic dimensions, examining the extent to which local interactions are important for employment outcomes. bandiera, barankay, and rasul (2008) create a bridge between network formation (namely, the creation of friendships amongst fruit pickers) and the effectiveness of different labor contracts. the extensive literature on networks in labor markets 1 documents the important role of social connections in transmitting information about jobs, and also differentiates between different types of social contacts and shows that even weak ties can be important in relaying information. there is further (and earlier) research that examines the different roles of individuals in diffusion. important work by katz and lazarsfeld (1955) (building on earlier studies of lazarsfeld, berelson, and gaudet (1944) , merton (1948) , and others), identifies the role of "opinion leaders" in the formation of various beliefs and opinions. individuals are heterogeneous (at least in behaviors), and some specialize in becoming well informed on certain subjects, and then information and opinions diffuse to other less informed individuals via conversations with these opinion leaders. lazarsfeld, berelson, and gaudet (1944) study voting decisions in an ohio town during the 1940 u.s. presidential campaign, and document the presence and significance of such opinion leaders. katz and lazarsfeld (1955) interviewed women in decatur, illinois, and asked about a number of things such as their views on household goods, fashion, movies, and local public affairs. when women showed a change in opinion in follow-up interviews, katz and lazarsfeld traced influences that led to the change in opinion, again finding evidence for the presence of opinion leaders. diffusion of new products is understandably a topic of much research. rogers (1995) discusses numerous studies illustrating the impacts of social interactions on the diffusion of new products, and suggests various factors that impact which products succeed and which products fail. for example, related to the idea of opinion leaders, feick and price (1987) surveyed 1531 households and provided evidence that consumers recognize and make use of particular individuals in their social network termed "market mavens," those who have a high propensity to provide marketplace and shopping information. whether or not products reach such mavens can influence the success of a product, independently of the product's quality. tucker (2008) uses micro-data on the adoption and use of a new video-messaging technology in an investment bank consisting of 2118 employees. tucker notes the effects of the underlying network in that employees follow the actions of those who either have formal power, or informal influence (which is, to some extent, endogenous to a social network). in the political context, there are several studies focusing on the social sources of information electors choose, as well as on the selective mis-perception of social information they are exposed to. a prime example of such a collection of studies is huckfeldt and sprague (1995), who concentrated on the social structure in south bend, indiana, during the 1984 elections. they illustrated the likelihood of socially connected individuals to hold similar political affiliations. in fact, the phenomenon of individuals connecting to individuals who are similar to them is observed across a wide array of attributes and is termed by sociologists homophily (for overviews see mcpherson, smith-lovin, and cook, 2001, jackson, 2007 , as well as the discussion of homophily in jackson, chapter 12 in this volume). while cross-sectional studies are tremendously interesting in that they suggest dimensions on which social interactions may have an impact, they face many empirical challenges. most notably, correlations between behaviors and outcomes of individuals and their peers may be driven by common unobservables and therefore be spurious. given the strong homophily patterns in many social interactions, individuals who associate with each other often have common unobserved traits, which could lead them to similar behaviors. this makes it difficult to draw (causal) conclusions from empirical analysis of the social impact on diffusion of behaviors based on cross-sectional data. 2 given some of the challenges with causal inference based on pure observation, laboratory experiments and field experiments are quite useful in eliciting the effects of real-world networks on fully controlled strategic interactions, and are being increasingly utilized. as an example, leider, mobius, rosenblat, and do (2009) elicited the friendship network among undergraduates at a u.s. college and illustrated how altruism varies as a function of social proximity. in a similar setup, goeree, mcconnell, mitchell, tromp, and yariv (2010) elicited the friendship network in an all-girls school in pasadena, ca, together with girls' characteristics and later ran dictator games with recipients who varied in social distance. they identified a "1/d law of giving," in that the percentage given to a friend was inversely related to her social distance in the network. 3 various field experiments, such as those by duflo and saez (2003) , karlan, mobius, rosenblat, and szeidl (2009), dupas (2010) , beaman and magruder (2010) , and feigenberg, field, and pande (2010) , also provide some control over the process, while working with real-world network structures to examine network influences on behavior. 4 another approach that can be taken to infer causal relationships is via structural modeling. as an example, one can examine the implications of a particular diffusion model for the patterns of adoption that should be observed. one can then infer characteristics of the process by fitting the process parameters to best match the observed outcomes in terms of behavior. for instance, banerjee, chandrasekhar, duflo, and jackson (2010) use such an approach in a study of the diffusion of microfinance participation in rural indian villages. using a model of diffusion that incorporates both information and peer effects, they then fit the model to infer the relative importance of information diffusion versus peer influences in accounting for differences in microfinance participation rates across villages. of course, in such an approach one is only as confident in the causal inference as one is confident that the model is capturing the essential underpinnings of the diffusion process. the types of conclusions that have been reached from these cross sectional studies can be roughly summarized as follows. first, in a wide variety of settings, associated individuals tend to have correlated actions and opinions. this does not necessarily embody diffusion or causation, but as discussed in the longitudinal section below, there is significant evidence of social influence in diffusion patterns as well. second, individuals tend to associate with others who are similar to themselves, in terms of beliefs and opinions. this has an impact on the structure of social interactions, and can affect diffusion. it also represents an empirical quandary of the extent to which social structure influences opinions and behavior as opposed to the reverse (that can partly be sorted out with careful analysis of longitudinal data). third, individuals fill different roles in a society, with some acting as "opinion leaders," and being key conduits of information and potential catalysts for diffusion. longitudinal data can be especially important in diffusion studies, as they provide information on how opinions and behaviors move through a society over time. they also help sort out issues of causation as well as supply-specific information about the extent to which behaviors and opinions are adopted dynamically, and by whom. such data can be especially important in going beyond the documentation of correlation between social connections and behaviors, and illustrating that social links are truly the conduits for information and diffusion if one is careful to track what is observed by whom at what point in time, and can measure the resulting changes in behavior. for example, conley and udry (2008) show that pineapple growers in ghana tend to follow those farmers who succeed in changing their levels of use of fertilizers. through careful examination of local ties, and the timing of different actions, they trace the influence of the outcome of one farmer's crop on subsequent behavior of other farmers. more generally, diffusion of new technologies is extremely important when looking at transitions in agriculture. seminal studies by ryan and gross (1943) and griliches (1957) examined the effects of social connections on the adoption of a new behavior, specifically the adoption of hybrid corn in the u.s. looking at aggregate adoption rates in different states, these authors illustrated that the diffusion of hybrid corn followed an s-shape curve over time: starting out slowly, accelerating, and then ultimately decelerating. 5 foster and rosenzweig (1995) collected household-level panel data from a representative sample of rural indian households having to do with the adoption and profitability of high-yielding seed varieties (associated with the green revolution). they identified significant learning-by-doing, where some of the learning was through neighbors' experience. in fact, the observation that adoption rates of new technologies, products, or behaviors exhibit s-shaped curves can be traced to very early studies, such as tarde (1903) , who discussed the importance of imitation in adoption. such patterns are found across many applications (see mahajan and peterson (1985) and rogers (1995) ). understanding diffusion is particularly important for epidemiology and medicine for several reasons. for one, it is important to understand how different types of diseases spread in a population. in addition, it is crucial to examine how new treatments get adopted. colizza, barrat, barthelemy, and vespignani (2006, 2007) tracked the spread of severe acute respiratory syndrome (sars) across the world combining census data with data on almost all air transit during the years 2002-2003. they illustrated the importance of structures of long-range transit networks for the spread of an epidemic. coleman, katz, and menzel (1966) is one of the first studies to document the role of social networks in diffusion processes. the study looked at the adoption of a new drug (tetracycline) by doctors and highlighted two observations. first, as with hybrid corn, adoption rates followed an s-shape curve over time. second, adoption rates depended on the density of social interactions. doctors with more contacts (measured according to the trust placed in them by other doctors) adopted at higher rates and earlier in time. 6 diffusion can occur in many different arenas of human behavior. for example christakis and fowler (2007) document influences of social contacts on obesity levels. they studied the social network of 12,067 individuals in the u.s. assessed repeatedly from 1971 to 2003 as part of the framingham heart study. concentrating on bodymass index, christakis and fowler found that a person's chances of becoming obese increased by 57% if he or she had a friend who became obese, by 40% if he or she had a sibling who became obese, and by 37% if they had a spouse who became obese in a previous period. the study controls for various selection effects, and takes advantage of the direction of friendship nominations to help sort out causation. for example, christakis and fowler find a significantly higher increase of an individual's body mass index in reaction to the obesity of someone that the individual named as a friend compared to someone who had named the individual as a friend. this is one method of sorting out causation, since if unobserved influences that were common to the agents were at work, then the direction of who mentioned the other as a friend would not matter, whereas direction would matter if it indicated which individuals react to which others. based on this analysis, christakis and fowler conclude that obesity spreads very much like an epidemic with the underlying social structure appearing to play an important role. it is worth emphasizing that even with longitudinal studies, one still has to be cautious in drawing causal inferences. the problem of homophily still looms, as linked individuals tend to have common characteristics and so may be influenced by common unobserved factors, for example, both being exposed to some external stimulus (such as advertising) at the same time. this then makes it appear as if one agent's behavior closely followed another's, even when it may simply be due to both having experienced a common external event that prompted their behaviors. aral, muchnik, and sundararajan (2009) provide an idea of how large this effect can be, by carefully tracking individual characteristics and then using propensity scores (likelihoods of having neighbors with certain behaviors) to illustrate the extent to which one can over-estimate diffusion effects by not accounting for common backgrounds of connected individuals. homophily not only suggests that linked individuals might be exposed to common influences, it also makes it hard to disentangle which of the following two processes is at the root of observed similarities in behavior between connected agents. it could be that similar behavior in fact comes from a process of selection (assortative pairing), in which similarity precedes association. alternatively, it could be a consequence of a process of socialization, in which association leads to similarity. in that respect, tracking connections and behaviors over time is particularly useful. kandel (1978) concentrated on adolescent friendship pairs and examined the levels of homophily on four attributes (frequency of current marijuana use, level of educational aspirations, political orientation, and participation in minor delinquency) at various stages of friendship formation and dissolution. she noted that observed homophily in friendship dyads resulted from a significant combination of both types of processes, so that individuals emulated their friends, but also tended to drop friendships with those more different from themselves and add new friendships to those more similar to themselves. 7 in summary, let us mention a few of the important conclusions obtained from studies of diffusion. first, not only are behaviors across socially connected individuals correlated, but individuals do influence each other. while this may sound straightforward, it takes careful control to ensure that it is not unobserved correlated traits or influences that lead to similar actions by connected individuals, as well as an analysis of similarities between friends that can lead to correlations in their preferences and the things that influence them. second, in various settings, more socially connected individuals adopt new behaviors and products earlier and at higher rates. third, diffusion exhibits specific patterns over time, and specifically there are many settings where an "s"-shaped pattern emerges, with adoption starting slowly, then accelerating, and eventually asymptoting. fourth, many diffusion processes are affected by the specifics of the patterns of interaction. we now turn to discussing various models of diffusion. as should be clear from our description of the empirical work on diffusion and behavior, models can help greatly in clarifying the tensions at play. given the issues associated with the endogeneity of social relationships, and the substantial homophily that may lead to correlated behaviors among social neighbors, it is critical to have models that help predict how behavior should evolve and how it interacts with the social structure in place. we start with some of the early models that do not account for the underlying network architecture per-se. these models incorporate the empirical observations regarding social influence through the particular dynamics assumed, or preferences posited, and generate predictions matching the aggregate empirical observations regarding diffusion over time of products, diseases, or behavior. for example, the so-called s-shaped adoption curves. after describing these models, we return to explicitly capturing the role of social networks. one of the earliest and still widely used models of diffusion is the bass (1969) model. this is a parsimonious model, which can be thought of as a "macro" model: it makes predictions about aggregate behavior in terms of the percentage of potential adopters of a product or behavior who will have adopted by a given time. the current rate of change of adoption depends on the current level and two critical parameters. these two parameters are linked to the rate at which people innovate or adopt on their own, and the rate at which they imitate or adopt because others have, thereby putting into (theoretical) force the empirical observation regarding peers' influence. if we let g(t) be the percentage of agents who have adopted by time t, and m be the fraction of agents in the population who are potential adopters, a discrete time version of the bass model is characterized by the difference equation where p is a rate of innovation and q is a rate of imitation. to glean some intuition, note that the expression p (m ã� g(t ã� 1)) represents the fraction of people who have not yet adopted and might potentially do so times the rate of spontaneous adoption. in the expression qã°m ã� gã°t ã� 1ã�ã� gã°tã�1ã� m , the rate of imitation is multiplied by two factors. the first factor, (m ã� g(t ã� 1)), is the fraction of people who have not yet adopted and may still do so. the second expression, g tã�1 ã° ã� m , is the relative fraction of potential adopters who are around to imitate. if we set m equal to 1, and look at a continuous time version of the above difference equation, we get where g is the rate of diffusion (times the rate of change of g). solving this when p > 0 and setting the initial set of adopters at 0, g(0) â¼ 0, leads to the following expression: this is a fairly flexible formula that works well at fitting time series data of innovations. by estimating p and q from existing data, one can also make forecasts of future diffusion. it has been used extensively in marketing and for the general analysis of diffusion (e.g., rogers (1995)), and has spawned many extensions and variations. 8 if q is large enough, 9 then there is a sufficient imitation/social effect, which means that the rate of adoption accelerates after it begins, and so g(t) is s-shaped (see figure 1 ), matching one of the main insights of the longitudinal empirical studies on diffusion discussed above. the bass model provides a clear intuition for why adoption curves would be s-shaped. indeed, when the adoption process begins, imitation plays a minor role (relative to innovation) since not many agents have adopted yet and so the volume of adopters grows slowly. as the number of adopters increases, the process starts to accelerate as now innovators are joined by imitators. the process eventually starts to slow down, in part simply because there are fewer agents left to adopt (the term 1ã�g(t) in (1) eventually becomes small). thus, we see a process that starts out slowly, then accelerates, and then eventually slows and asymptotes. the bass model described above is mechanical in that adopters and imitators are randomly determined; they do not choose actions strategically. the empirical observation that individuals influence each other through social contact can be derived through agents' preferences, rather than through some exogenously specified dynamics. diffusion in a strategic context was first studied without a specific structure for interactions. broadly speaking, there were two approaches taken in this early literature. in the first, all agents are connected to one another (that is, they form a complete network). effectively, this corresponds to a standard multi-agent game in which payoffs to each player depend on the entire profile of actions played in the population. the second approach has been to look at interactions in which agents are matched to partners in a random fashion. diffusion on complete networks. granovetter (1978) considered a model in which n agents are all connected to one another and each agent chooses one of two actions: 0 or 1. associated with each agent i is a number n i . this is a threshold such that if at least n i other agents take action 1 then i prefers action 1 to action 0, and if fewer than n i other agents take action 1 then agent i prefers to take action 0. the game exhibits what are known as strategic complementarities. for instance, suppose that the utility of agent i faced with a profile of actions (x 1 , . . ., x n ) 2 {0, 1} n is described by: where c i is randomly drawn from a distribution f over [0,1]. c i can be thought of as a cost that agent i experiences upon choosing action 1 (e.g., a one-time switching cost from one technology to the other, or potential time costs of joining a political revolt, etc.). the utility of agent i is normalized to 0 when choosing the action 0. when choosing the action 1, agent i experiences a benefit proportional to the fraction of other agents choosing the action 1 and a cost of c i . granovetter considered a dynamic model in which at each stage agents best respond to the previous period's distribution of actions. if in period t there was a fraction x t of agents choosing the action 1, then in period t ã¾ 1 an agent i chooses action 1 if and only if his or her cost is lower than nx t ã�x t i n ã�1 , the fraction of other agents taking action 1 in the last period. for a large population, then corresponds to an (approximate) equilibrium of a large population. the shape of the distribution f determines which equilibria are tipping points: equilibria such that only a slight addition to the fraction of agents choosing the action 1 shifts the population, under the best response dynamics, to the next higher equilibrium level of adoption (we return to a discussion of tipping and stable points when we consider a more general model of strategic interactions on networks below). note that while in the bass model the diffusion path was determined by g(t), the fraction of adopters as a function of time, here it is easier to work with f(x), corresponding to the fraction of adopters as a function of the previous period's fraction x. although granovetter (1978) does not examine conditions under which the time series will exhibit attributes like the s-shape that we discussed above, by using techniques from jackson and yariv (2007) we can derive such results, as we now discuss. keeping track of time in discrete periods (a continuous time analog is straightforward), the level of change of adoption in the society is given by thus, to derive an s-shape, we need this quantity to initially be increasing, and then eventually to decrease. assuming differentiability of f, this corresponds to the derivative of d(x t ) being positive up to some x and then negative. the derivative of f(x) ã� x is f 0 (x) ã� 1 and having an s-shape corresponds to f 0 being greater than 1 up to some point and then less than 1 beyond that point. for instance, if f is concave with an initial slope greater than 1 and an eventual slope less than 1, this is satisfied. note that the s-shape of adoption over time does not translate into an s-shape of f -but rather a sort of concavity. 10 the idea is that we initially need a rapid level of change, which corresponds to an initially high slope of f, and then a slowing down, which corresponds to a lower slope of f. fashions and random matching. a different approach than that of the bass model is taken by pesendorfer (1995) , who considers a model in which individuals are randomly matched and new fashions serve as signaling instruments for the creation of matches. he identifies particular matching technologies that generate fashion cycles. pesendorfer describes the spread of a new fashion as well as its decay over time. in pesendorfer's model, the price of the design falls as it spreads across the population. once sufficiently many consumers own the design, it is profitable to create a new design and thereby render the old design obsolete. in particular, demand for any new innovation eventually levels off as in the above two models. information cascades and learning. another influence on collective behavior derives from social learning. this can happen without any direct complementarities in actions, but due to information flow about the potential payoffs from different behaviors. if people discuss which products are worth buying, or which technologies are worth adopting, books worth reading, and so forth, even without any complementarities in behaviors, one can end up with cascades in behavior, as people infer information from others' behaviors and can (rationally) imitate them. as effects along these lines are discussed at some length in jackson (chapter 12, this volume) and goyal (chapter 15, this volume), we will not detail them here. we only stress that pure information transmission can lead to diffusion of behaviors. we now turn to models that explicitly incorporate social structure in examining diffusion patterns. we start with models that stem mostly from the epidemiology literature and account for the underlying social network, but are mechanical in terms of the way that disease spreads from one individual to another (much like the bass model described above). we then proceed to models in which players make choices that depend on their neighbors' actions as embedded in a social network; for instance, only adopting an action if a certain proportion of neighbors adopt as well (as in granovetter's setup), or possibly not adopting an action if enough neighbors do so. many models of diffusion and strategic interaction on networks have the following common elements. there is a finite set of agents n â¼ {1, . . ., n}. agents are connected by a (possibly directed) network g 2 {0, 1} nã�n . we let n i (g) {j : g ij â¼ 1} be the neighbors of i. the degree of a node i is the number of her neighbors, d i jn i (g)j. when links are determined through some random process, it is often useful to summarize the process by the resulting distribution of degrees p, where p(d) denotes the probability a random individual has a degree of d. 11, 12 each agent i 2 n takes an action x i . in order to unify and simplify the description of various models, we focus on binary actions, so that x i 2 {0, 1}. actions can be metaphors for becoming "infected" or not, buying a new product or not, choosing one of two activities, and so forth. some basic insights about the extent to which behavior or an infection can spread in a society can be derived from random graph theory. random graph theory provides a tractable base for understanding characteristics important for diffusion, such as the structure and size of the components of a network, maximally connected subnetworks. 13 before presenting some results, let us talk through some of the ideas in the context of what is known as the reed-frost model. 14 consider, for example, the spread of a disease. initially, some individuals in the society are infected through mutations of a germ or other exogenous sources. consequently, some of these individuals' neighbors are infected through contact, while others are not. this depends on how virulent the disease is, among other things. in this application, it makes sense (at least as a starting point) to assume that becoming infected or avoiding infection is not a choice; 11 such a description is not complete, in that it does not specify the potential correlations between degrees of different individuals on the network. see galeotti, goyal, jackson, vega-redondo, and yariv (2010) for more details. 12 in principle, one would want to calibrate degree distributions with actual data. the literature on network formation, see bloch and dutta (chapter 16, this volume) and jackson (chapter 12, this volume), suggests some insights on plausible degree distributions p(d). 13 formally, these are the subnetworks projected induced by maximal sets c n of nodes such any two distinct nodes in c are path connected within c. that is, for any i,j 2 c, there exist i 1 , . . ., i k 2 c such that g ii1 â¼ g i1i2 â¼ . . . â¼ g ikã�1ik â¼ g ikj â¼ 1. 14 see jackson (2008) for a more detailed discussion of this and related models. i.e., contagion here is nonstrategic. in the simplest model, there is a probability p ! 0 that a given individual is immune (e.g., through vaccination or natural defenses). if an individual is not immune, it is assumed that he or she is sure to catch the disease if one of his or her neighbors ends up with the disease. in this case, in order to estimate the volume of those ultimately infected, we proceed in two steps, depicted in figure 2 . first, we delete a fraction p of the nodes that will never be infected (these correspond to the dotted nodes in the figure) . then, we note that the components of the remaining network that contain the originally infected individuals comprise the full extent of the infection. in particular, if we can characterize what the components of the network look like after removing some portion of the nodes, we have an idea of the extent of the infection. in figure 2 , we start with one large connected component (circumvented by a dotted line) and two small-connected components. after removing the immune agents, there is still a large connected component (though smaller than before), and four small components. thus, the estimation of the extent of infection of the society is reduced to the estimation of the component structure of the network. a starting point for the formal analysis of this sort of model uses the canonical random network model, where links are formed independently, each with an identical probability p > 0 of being present. this is sometimes referred to as a "poisson random network" as its degree distribution is approximated by a poisson distribution if p is not excessively large; and has various other aliases such as an "erdã¶s-renyi random graph," a "bernoulli random graph," or a "g(n,p)" random graph (see jackson, chapter 12 in this volume, for more removing immune agents background). ultimately, the analysis boils down to considering a network on (1ã�p)n nodes with an independent link probability of p, and then measuring the size of the component containing a randomly chosen initially infected node. clearly, with a fixed set of nodes, and a positive probability p that lies strictly between 0 and 1, every conceivable network on the given set of nodes could arise. thus, in order to say something specific about the properties of the networks that are "most likely" to arise, one generally works with large n where reasoning based on laws of large numbers can be employed. for example, if we think of letting n grow, we can ask for which p's (that are now dependent on n) a nonvanishing fraction of nodes will become infected with a probability bounded away from 0. so, let us consider a sequence of societies indexed by n and corresponding probabilities of links p(n). erdã¶s and renyi (1959 renyi ( , 1960 proved a series of results that characterize some basic properties of such random graphs. in particular, 15 â�¢ the threshold for the existence of a "giant component," a component that contains a nontrivial fraction of the population, is 1/n, corresponding to an average degree of 1. that is, if p(n) over 1/n tends to infinity, then the probability of having a giant component tends to 1, while if p(n) over 1/n tends to 0, then the probability of having a giant component tends to 0. â�¢ the threshold for the network to be connected (so that every two nodes have a path between them) is log(n)/n, corresponding to an average degree that is proportional to log(n). the logic for the first threshold is easy to explain, though the proof is rather involved. to heuristically derive the threshold for the emergence of a giant component, consider following a link out of a given node. we ask whether or not one would expect to be able to find a link to another node from that one. if the expected degree is much smaller than 1, then following the few (if any) links from any given node is likely to lead to dead-ends. in contrast, when the expected degree is much higher than 1, then from any given node, one expects to be able to reach more nodes, and then even more nodes, and so forth, and so the component should expand outward. note that adjusting for the factor p of the number of immune nodes does not affect the above thresholds as they apply as limiting results, although the factor will be important for any fixed n. between these two thresholds, there is only one giant component, so that the next largest component is of a size that is a vanishing fraction of the giant component. this is intuitively clear, as to have two large components requires many links within each component but no links between the two components, which is an unlikely event. in that sense, the image that emerges from figure 2 of one large connected component is reasonably typical for a range of parameter values. these results then tell us that in a random network, if average degree is quite low (smaller than 1), then any initial infection is likely to die out. in contrast, if average degree is quite high (larger than log(n)), then any initial infection is likely to spread to all of the susceptible individuals, i.e., a fraction of 1 ã� p of the population. in the intermediate range, there is a probability that the infection will die out and also a probability that it will infect a nontrivial, but limited, portion of the susceptible population. there, it can be shown that for such random networks and large n, the fraction of nodes in the giant component of susceptible nodes is roughly approximated by the nonzero q that solves here, q is an approximation of the probability of the infection spreading to a nontrivial fraction of nodes, and also of the percentage of susceptible nodes that would be infected. 16 this provides a rough idea of the type of results that can be derived from random graph theory. there is much more that is known, as one can work with other models of random graphs (other than ones where each link has an identical probability), richer models of probabilistic infection between nodes, as well as derive more information about the potential distribution of infected individuals. it should also be emphasized that while the discussion here is in terms of "infection," the applications clearly extend to many of the other contexts we have been mentioning, such as the transmission of ideas and information. fuller treatment of behaviors, where individual decisions depend in more complicated ways on neighbors' decisions, are treated in section 4.3. the above analysis of diffusion presumes that once infected, a node eventually infects all of its susceptible neighbors. this misses important aspects of many applications. in terms of diseases, infected nodes can either recover and stop transmitting a disease, or die and completely disappear from the network. transmission will also generally be probabilistic, depending on the type of interaction and its extent. 17 similarly, if we think of behaviors, it might be that the likelihood that a node is still actively transmitting a bit of information to its neighbors decreases over time. ultimately, we will discuss models that allow for rather general strategic impact of peer behavior (a generalization of the approach taken by granovetter). but first we discuss some aspects of the epidemiology literature that takes steps forward in that direction by considering two alternative models that keep track of the state of nodes and are more explicitly dynamic. the common terminology for the possible states that 16 again, see chapter 4 in jackson (2008) for more details. 17 probabistic transmission is easily handled in the above model by simply adjusting the link probability to reflect the fact that some links might not transmit the disease. a node can be in are: susceptible, where a node is not currently infected or transmitting a disease but can catch it; infected, where a node has a disease and can transmit it to its neighbors; and removed (or recovered), where a node has been infected but is no longer able to transmit the disease and cannot be re-infected. the first of the leading models is the "sir" model (dating to kermack and mckendrick, 1927) , where nodes are initially susceptible but can catch the disease from infected neighbors. once infected, a node continues to infect neighbors until it is randomly removed from the system. this fits well the biology of some childhood diseases, such as the chicken pox, where one can only be infected once. the other model is the "sis" model (see bailey, 1975) , where once infected, nodes can randomly recover, but then they are susceptible again. this corresponds well with an assortment of bacterial infections, viruses, and flus, where one transitions back and forth between health and illness. the analysis of the sir model is a variant of the component-size analysis discussed above. the idea is that there is a random chance that an "infected" node infects a given "susceptible" neighbor before becoming "removed." roughly, one examines component structures in which instead of removing nodes randomly, one removes links randomly from the network. this results in variations on the above sorts of calculations, where there are adjusted thresholds for infection depending on the relative rates of how quickly infected nodes can infect their neighbors compared to how quickly they are removed. in contrast, the sis model involves a different sort of analysis. the canonical version of that model is best viewed as one with a random matching process rather than a social network. in particular, suppose that a node i in each period will have interactions with d i other individuals from the population. recall our notation of p(d) describing the proportion of the population that has degree d (so d interactions per period). the matches are determined randomly, in such a way that if i is matched with j, then the probability that j has degree d > 0 is given bỹ where hã�i represents the expectation with respect to p. 18 this reflects the fact that an agent is proportionally more likely to be matched with other individuals who have lots of connections. to justify this formally, one needs an infinite population. indeed, with any finite population of agents with heterogeneous degrees, the emergent networks will generally exhibit some correlation between neighbors' degrees. 19 individuals who have high degrees will have more interactions per period and will generally be more likely to be infected at any given time. an important calculation 18 we consider only individuals who have degree d > 0, as others do not participate in the society. 19 see the appendix of currarini, jackson, and pin (2009) for some details along this line. then pertains to the chance that a given meeting will be with an infected individual. if the infection rate of degree d individuals is r(d ), the probability that any given meeting will be with an infected individual is y, where the chance of meeting an infected individual in a given encounter then differs from the average infection rate in the population, which is just r â¼ p p d ã° ã�r d ã° ã�, because y is weighted by the rate at which individuals meet each other. a standard version of contagion that is commonly analyzed is one in which the probability of an agent of degree d becoming infected is where n 2 (0, 1) is a rate of transmission of infection in a given period, and is small enough so that this probability is less than one. if n is very small, this is an approximation of getting infected under d interactions with each having an (independent) probability y of being infected and then conditionally (and independently) having a probability n of getting infected through contact with a given infected individual. the last part of the model is that in any given period, an infected individual recovers and becomes susceptible with a probability d 2 (0, 1). if such a system operates on a finite population, then eventually all agents will become susceptible and that would end the infection. if there is a small probability of a new mutation and infection in any given period, the system will be ergodic and always have some probability of future infection. to get a feeling for the long run outcomes in large societies, the literature has examined a steady state (i.e., a situation in which the system essentially remains constant) of a process that is idealized as operating on an infinite (continuous) population. formally, a steady-state is defined by having r(d) be constant over time for each d. working with an approximation at the limit (termed a "mean-field" approximation that in this case can be justified with a continuum of agents, but with quite a bit of technical detail), a steady-state condition can be derived to be for each d. (1ã�r(d))nyd is the rate at which agents of degree d who were susceptible become infected and r(d)d is the rate at which infected individuals of degree d recover. letting l â¼ n d , it follows that solving (5) and (8) simultaneously leads to a characterization of the steady-state y: this system always has a solution, and therefore a steady-state, where y â¼ 0 so there is no infection. it can also have other solutions under which y is positive (but always below 1 if l is finite). unless p takes very specific forms, it can be difficult to find steady states y > 0 analytically. special cases have been analyzed, such as the case of a power distribution, where p(d ) â¼ 2d ã�3 (e.g., see pastor-satorras and vespignani (2000, 2001) ). in that case, there is always a positive steady-state infection rate. more generally, lopez-pintado (2008) addresses the question of when it is that there will be a positive steady-state infection rate. to get some intuition for her results, let so that the equation y â¼ h(y) corresponds to steady states of the system. we can now extend the analysis of granovetter's (1978) model that we described above, with this richer model in which h(y) accounts for network attributes. while the fixed-point equation identifying granovetter's stable points allowed for rather arbitrary diffusion patterns (depending on the cost distribution f), the function h has additional structure to it that we can explore. in particular, suppose we examine the infection rate that would result if we start at a rate of y and then run the system on an infinite population for one period. noting that h(0) â¼ 0, it is clear that 0 is always a fixed-point and thus a steady-state. since h(1) < 1, and h is increasing and strictly concave in y (which is seen by examining its first and second derivatives), there can be at most one fixed-point besides 0. for there to be another fixed-point (steady-state) above y â¼ 0, it must be that h 0 (0) is above 1, or else, given the strict concavity, we would have h(y) < y for all positive y. moreover, in cases where h 0 (0) > 1, a small perturbation away from a 0 infection rate will lead to increased infection. in the terminology we have introduced above, 0 would be a tipping point. since higher infection rates lead to the possibility of positive infection, as do degree distributions with high variances (relative to mean). the idea behind having a high variance is that there will be some "hub nodes" with high degree, who can foster contagion. going back to our empirical insights, this analysis fits the observations that highlylinked individuals are more likely to get infected and experience speedier diffusion. whether the aggregate behavior exhibits the s-shape that is common in many realworld diffusion processes will depend on the particulars of h, much in the same way that we discussed how the s-shape in granovetter's model depends on the shape of the distribution of costs f in that model. here, things are slightly complicated since h is a function of y, which is the probability of infection of a neighbor, and not the overall probability of infection of the population. thus, one needs to further translate how various y's over time translate into population fractions that are infected. beyond the extant empirical studies, this analysis provides some intuitions behind what is needed for an infection to be possible. it does not, however, provide an idea of how extensive the infection spread will be and how that depends on network structure. while this does not boil down to as simple a comparison as (12), there is still much that can be deduced using (9), as shown by jackson and rogers (2007) . while one cannot always directly solve notice that lyd 2 hdiã°lyd ã¾ 1ã� is an increasing and convex function of d. therefore, the right hand side of the above equality can be ordered when comparing different degree distributions in the sense of stochastic dominance (we will return to these sorts of comparisons in some of the models we discuss below). the interesting conclusion regarding steady-state infection rates is that they depend on network structure in ways that are very different at low levels of the infection rate l compared to high levels. while the above models provide some ideas about how social structure impacts diffusion, they are limited to settings where, roughly speaking, the probability that a given individual adopts a behavior is simply proportional to the infection rate of neighbors. especially when it comes to situations in which opinions or technologies are adopted, purchasing decisions are made, etc., an individual's decision can depend in much more complicated ways on the behavior of his or her neighbors. such interaction naturally calls on game theory as a tool for modeling these richer interactions. we start with static models of interactions on networks that allow for a rather general impact of peers' actions on one's own optimal choices. the first model that explicitly examines games played on a network is the model of "graphical games" as introduced by kearns, littman, and singh (2001) , and analyzed by kakade, kearns, langford, and ortiz (2003) , among others. the underlying premise in the graphical games model is that agents' payoffs depend on their own actions and the actions of their direct neighbors, as determined by the network of connections. 20 formally, the payoff structure underlying a graphical game is as follows. the payoff to each player i when the profile of actions is x â¼ (x 1 , . . ., x n ) is where x n i g ã° ã� is the profile of actions taken by the neighbors of i in the network g. most of the empirical applications discussed earlier entailed agents responding to neighbors' actions in roughly one of two ways. in some contexts, such as those pertaining to the adoption of a new product or new agricultural grain, decisions to join the workforce, or to join a criminal network, agents conceivably gain more from a particular action the greater is the volume of peers who choose a similar action. that is, payoffs exhibit strategic complementarities. in other contexts, such as experimentation on a new drug, or contribution to a public good, when an agent's neighbors choose a particular action, the relative payoff the agent gains from choosing a similar action decreases, and there is strategic substitutability. the graphical games environment allows for the analysis of both types of setups, as the following example (taken from galeotti, goyal, jackson, vega-redondo, and yariv (2010)) illustrates. example 1 (payoffs depend on the sum of actions) player i's payoff function when he or she chooses x i and her k neighbors choose the profile (x 1 , . . ., x k ) is: where f(ã�) is nondecreasing and c(ã�) is a "cost" function associated with own effort (more general but much in the spirit of (2)). the parameter l 2 r determines the nature of the externality across players' actions. the shape and sign of lf determine the effects of neighbors' action choices on one's own optimal choice. in particular, the example yields strict strategic substitutes (complements) if, assuming differentiability, lf 00 is negative (positive). there are several papers that analyze graphical games for particular choices of f and l. to mention a few examples, the case where f is concave, l â¼ 1, and c(ã�) is increasing and linear corresponds to the case of information sharing as a local public good studied by bramoullã© and kranton (2007) , where actions are strategic substitutes. in contrast, if l â¼ 1, but f is convex (with c 00 > f 00 > 0), we obtain a model with strategic complements, as proposed by goyal and moraga-gonzalez (2001) to study collaboration among local monopolies. in fact, the formulation in (13) is general enough to accommodate numerous further examples in the literature such as human capital investment (calvã³-armengol and jackson (2009)), crime and other networks (ballester, calvã³-armengol, and zenou (2006) ), some coordination problems (ellison (1993) ), and the onset of social unrest (chwe (2000) ). the computer science literature (e.g., the literature following kearns, littman, and singh (2001) , and analyzed by kakade, kearns, langford, and ortiz (2003) ) has focused predominantly on the question of when an efficient (polynomial-time) algorithm can be provided to compute nash equilibria of graphical games. it has not had much to say about the properties of equilibria, which is important when thinking about applying such models to analyze diffusion in the presence of strategic interaction. in contrast, the economics literature has concentrated on characterizing equilibrium outcomes for particular applications, and deriving general comparative statics with respect to agents' positions in a network and with respect to the network architecture itself. information players hold regarding the underlying network (namely, whether they are fully informed of the entire set of connections in the population, or only of connections in some local neighborhood) ends up playing a crucial role in the scope of predictions generated by network game models. importantly, graphical games are ones in which agents have complete information regarding the networks in place. consequently, such models suffer from inherent multiplicity problems, as clearly illustrated in the following example. it is based on a variation of (13), which is similar to a model analyzed by bramoullã© and kranton (2007) . example 2 (multiplicity -complete information) suppose that in (13), we set l â¼ 1, choose x i 2 {0, 1}, and have and c(x i ) cx i , where 0 < c < 1. this game, often labeled the best-shot public goods game, may be viewed as a game of local public-good provision. each agent would choose the action 1 (say, experimenting with a new grain, or buying a product that can be shared with one's friends) if they were alone (or no one else experimented), but would prefer that one of their neighbors incur the cost c that the action 1 entails (when experimentation is observed publicly). effectively, an agent just needs at least one agent in his or her neighborhood to take action 1 to enjoy its full benefits, but prefers that it be someone else given that the action is costly and there is no additional benefit beyond one person taking the action. note that, since c < 1, in any (pure strategy) nash equilibrium, for any player i with k neighbors, it must be the case that one of the agents in the neighborhood chooses the action 1. that is, if the chosen profile is (x 1 , . . ., x k ), then x i ã¾ p k jâ¼1 x j ! 1. in fact, there is a very rich set of equilibria in this game. to see this, consider a star network and note that there exist two equilibria, one in which the center chooses 0 and the spokes choose 1, and a second equilibrium in which the spoke players choose 0 while the center chooses 1. figure 3 illustrates these two equilibria. in the first, depicted in the left panel of the figure, the center earns more than the spoke players, while in the second equilibrium (in the right panel) it is the other way round. even in the simplest network structures equilibrium multiplicity may arise and the relation between network architecture, equilibrium actions, and systematic patterns can be difficult to discover. while complete information regarding the structure of the social network imposed in graphical game models may be very sensible when the relevant network of agents is small, in large groups of agents (such as a country's electorate, the entire set of corn growers in the 50's, sites in the world-wide web, or academic economists), it is often the case that individuals have noisy perceptions of their network's architecture. as the discussion above stressed, complete information poses many challenges because of the widespread occurrence of equilibrium multiplicity that accompanies it. in contrast, when one looks at another benchmark, where agents know how many neighbors they will have but not who they will be, the equilibrium correspondence is much easier to deal with. moreover, this benchmark is an idealized model of settings in which agents make choices like learning a language or adopting a technology that they will use over a long time. in such contexts, agents have some idea of how many interactions they are likely to have in the future, but not exactly with whom the interactions will be. a network game is a modification of a graphical game in which agents can have private and incomplete information regarding the realized social network at place. we describe here the setup corresponding to that analyzed by galeotti, goyal, jackson, vega-redondo, and yariv (2010) and yariv (2005, 2007) , restricting attention to binary action games. 21 uncertainty is operationalized by assuming the network is determined according to some random process yielding our distribution over agents' degrees, p(d), which is common knowledge. each player i has d i interactions, but does not know how many interactions each neighbor has. thus, each player knows something about his or her local neighborhood (the number of direct neighbors), but only the distribution of links in the remaining population. consider now the following utility specification, a generalization of (2). agent i has a cost of choosing 1, denoted c i . costs are randomly and independently distributed across the society, according to a distribution f c . normalize the utility from the action 0 to 0 and let the benefit of agent i from action 1 be denoted by v(d i , x), where d i is i 0 s degree and she expects each of her neighbors to independently choose the action 1 with probability x. agent i's added payoff from adopting behavior 1 over sticking to the action 0 is then this captures how the number of neighbors that i has, as well as their propensity to choose the action 1, affects the benefits from adopting 1. in particular, i prefers to choose the action 1 if this is a simple cost-benefit analysis generalizing granovetter (1978) 's setup in that benefits can now depend on one's own degree (so that the underlying network is accounted for). let f(d, x) f c (v(d, x) ). in words, f(d, x) is the probability that a random agent of degree d chooses the action 1 when anticipating that each neighbor will choose 1 with an independent probability x. note that v(d, x) can encompass all sorts of social interactions. in particular, it allows for a simple generalization of granovetter's (1978) model to situations in which agents' payoffs depend on the expected number of neighbors adopting, dx. existence of symmetric bayesian equilibria follows standard arguments. in cases where v is nondecreasing in x for each d, it is a direct consequence of tarski's fixed-point theorem. in fact, in this case, there exists an equilibrium in pure strategies. in other cases, provided v is continuous in x for each d, a fixed-point can still be found by appealing to standard theorems (e.g., kakutani) and admitting mixed strategies. 22 homogeneous costs. suppose first that all individuals experience the same cost c > 0 of choosing the action 1 (much like in example 2 above). in that case, as long as v(d, x) is monotonic in d (nonincreasing or nondecreasing), equilibria are characterized by a threshold. indeed, suppose v(d, x) is increasing in d, then any equilibrium is characterized by a threshold d ã� such that all agents of degree d < d ã� choose the action 0 and all agents of degree d > d ã� choose the action 1 (and agents of degree d ã� may mix between the actions). in particular, notice that the type of multiplicity that appeared in example 2 no longer occurs (provided degree distributions are not trivial). it is now possible to look at comparative statics of equilibrium behavior and outcomes using stochastic dominance arguments on the network itself. for ease of exposition, we illustrate this in the case of nonatomic costs (see galeotti, goyal, jackson, vega-redondo, and yariv (2010) for the general analysis). heterogeneous costs. consider the case in which f c is a continuous function, with no atoms. in this case, a simple equation is sufficient to characterize equilibria. let x be the probability that a randomly chosen neighbor chooses the action 1. then f(d, x) is the probability that a random (best responding) neighbor of degree d chooses the action 1. we can now proceed in a way reminiscent of the analysis of the sis model. recall thatpã°dã� denoted the probability that a random neighbor is of degree d (see equation (4)). it must be that again, a fixed-point equation captures much of what occurs in the game. in fact, equation (15) characterizes equilibria in the sense that any symmetric 23 equilibrium results in an x that satisfies the equation, and any x that satisfies the equation corresponds to an equilibrium where type (d i , c i ) chooses 1 if and only if inequality (14) holds. given that equilibria can be described by their corresponding x, we often refer to some value of x as being an "equilibrium." consider a symmetric equilibrium and a corresponding probability of x for a random neighbor to choose action 1. if the payoff function v is increasing in degree d, then the expected payoff of an agent with degree d ã¾ 1 is ã�and agents with higher degrees choose 1 with weakly higher probabilities. indeed, an agent of degree d ã¾ 1 can imitate the decisions of an 22 in such a case, the best response correspondence (allowing mixed strategies) for any (d i , c i ) as dependent on x is upper hemi-continuous and convex-valued. taking expectations with respect to d i and c i , we also have a set of population best responses as dependent on x that is upper hemi-continuous and convex valued. 23 symmetry indicates that agents with the same degree and costs follow similar actions. agent of degree d and gain at least as high a payoff. thus, if v is increasing (or, in much the same way, decreasing) in d for each x, then any symmetric equilibrium entails agents with higher degrees choosing action 1 with weakly higher (lower) probability. furthermore, agents of higher degree have higher (lower) expected payoffs. much as in the analysis of the epidemiological models, the multiplicity of equilibria is determined by the properties of f, which, in turn, correspond to properties ofp and f. for instance, â�¢ if f(d, 0) > 0 for some d in the support of p, and f is concave in x for each d, then there exists at most one fixed-point, and â�¢ if f(d, 0) â¼ 0 for all d and f is strictly concave or strictly convex in x for each d, then there are at most two equilibria-one at 0, and possibly an additional one, depending on the slope of f(x) at x â¼ 0. 24 in general, as long as the graph of f(x) crosses the 45-degree line only once, there is a unique equilibrium (see figure 4 below). 25 the set of equilibria generated in such network games is divided into stable and unstable ones (those we have already termed in section 3.2 as tipping points). the simple characterization given by (15) allows for a variety of comparative statics on fundamentals pertaining to either type of equilibrium. in what follows, we show how these 24 as before, the slope needs to be greater than 1 for there to be an additional equilibrium in the case of strict concavity, while the case of strict convexity depends on the various values of f(d, 1) across d. 25 morris and shin (2003, 2005) consider uncertainty on payoffs rather than on an underlying network. in coordination games, they identify a class of payoff shocks that lead to a unique equilibrium. heterogeneity in degrees combined with uncertainty plays a similar role in restricting the set of equilibria. in a sense, the analysis described here is a generalization in that it allows studying the impact of changes in a variety of fundamentals on the set of stable and unstable equilibria, regardless of multiplicity, in a rather rich environment. moreover, the equilibrium structure can be tied to the network of underlying social interactions. x t figure 4 the effects of shifting f(x) pointwise. comparative statics tie directly to a simple strategic diffusion process. indeed, it turns out there is a very useful technical link between the static and dynamic analysis of strategic interactions on networks. an early contribution to the study of diffusion of strategic behavior allowing for general network architectures was by morris (2000) . 26 morris (2000) considered coordination games played on networks. his analysis pertained to identifying social structures conducive to contagion, where a small fraction of the population choosing one action leads to that action spreading across the entire population. the main insight from morris (2000) is that maximal contagion occurs when the society has certain sorts of cohesion properties, where there are no groups (among those not initially infected) that are too inward looking in terms of their connections. in order to identify the full set of stable of equilibria using the above formalization, consider a diffusion process governed by best responses in discrete time (following yariv (2005, 2007) ). at time t â¼ 0, a fraction x 0 of the population is exogenously and randomly assigned the action 1, and the rest of the population is assigned the action 0. at each time t > 0, each agent, including the agents assigned to action 1 at the outset, best responds to the distribution of agents choosing the action 1 in period tã�1, accounting for the number of neighbors they have and presuming that their neighbors will be a random draw from the population. let x t d denote the fraction of those agents with degree d who have adopted behavior 1 at time t, and let x t denote the link-weighted fraction of agents who have adopted the behavior at time t. that is, using the distribution of neighbors' degreespã°dã�, then, as deduced before from equation (14), at each date t, and therefore x t â¼ x dp ã°dã�fã°d; x tã�1 ã� â¼ fã°x tã�1 ã�: as we have discussed, any rest point of the system corresponds to a static (bayesian) equilibrium of the system. 26 one can find predecessors with regards to specific architectures, usually lattices or complete mixings, such as conway's (1970) "game of life," and various agent-based models that followed such as the "voter model" (e.g., see clifford and sudbury (1973) and holley and liggett (1975) ), as well as models of stochastic stability (e.g., kandori, mailath, robb (1993) , young (1993) , ellison (1993) ). if payoffs exhibit complementarities, then convergence of behavior from any starting point is monotone, either upwards or downwards. in particular, once an agent switches behaviors, the agent will not want to switch back at a later date. 27 thus, although these best responses are myopic, any eventual changes in behavior are equivalently forward-looking. figure 4 depicts a mapping f governing the dynamics. equilibria, and resting points of the diffusion process, correspond to intersections of f with the 45-degree line. the figure allows an immediate distinction between two classes of equilibria that we discussed informally up to now. formally, an equilibrium x is stable if there exists e 0 > 0 such that f(x ã� e) > x ã� e and f(x ã¾ e) < x ã¾ e for all e 0 > e > 0. an equilibrium x is unstable or a tipping point if there exists e 0 > 0 such that f(x ã� e) < x ã� e and f(x ã¾ e) > x ã¾ e for all e 0 > e > 0. in the figure, the equilibrium to the left is a tipping point, while the equilibrium to the right is stable. the composition of the equilibrium set hinges on the shape of the function f. furthermore, note that a point-wise shift of f (as in the figure, to a new function f) shifts tipping points to the left and stable points to the right, loosely speaking (as sufficient shifts may eliminate some equilibria altogether), making adoption more likely. this simple insight allows for a variety of comparative statics. for instance, consider an increase in the cost of adoption, manifested as a first order stochastic dominance (fosd) shift of the cost distribution f c to f c . it follows immediately that: fã°xã� â¼ x dp ã°dã�f c ã°vã°d; xã�ã� x dp ã°dã�f c ã°vã°d; xã�ã� â¼ fã°xã� and the increase in costs corresponds to an increase of the tipping points and decrease of the stable equilibria (one by one). intuitively, increasing the barrier to choosing the action 1 leads to a higher fraction of existing adopters necessary to get the action 1 to spread even more. this formulation also allows for an analysis that goes beyond graphical games regarding the social network itself, using stochastic dominance arguments (following jackson and rogers (2007) ) and yariv (2005, 2007) ). for instance, consider an increase in the expected degree of each random neighbor that an agent has. that is, supposep ' fosdp and, for illustration, assume that f(d, x) is nondecreasing in d for all x. then, by the definition of fosd, f 0 ã°xã� â¼ x dp 0 ã°dã�fã°d; xã� ! x dp ã°dã�fã°d; xã� â¼ fã°xã�; and, under p 0 , tipping points are lower and stable equilibria are higher. 27 if actions are strategic substitutes, convergence may not be guaranteed for all starting points. however, whenever convergence is achieved, the rest point is an equilibrium, and the analysis can therefore be useful for such games as well. similar analysis allows for comparative statics regarding the distribution of links, by simply looking at mean preserving spreads (mps) of the underlying degree distribution. 28 going back to the dynamic path of adoption, we can generalize the insights that we derived regarding the granovetter (1978) model. namely, whether adoption paths track an s-shaped curve now depends on the shape of f, and thereby on the shape of both the cost distribution f and agents' utilities. there is now a substantial and growing body of research studying the impacts of interactions that occur on a network of connections. this work builds on the empirical observations of peer influence and generates a rich set of individual and aggregate predictions. insights that have been shown consistently in real-world data pertain to the higher propensities of contagion (of a disease, an action, or behavior) in more highly connected individuals, the role of "opinion leaders" in diffusion, as well as an aggregate s-shape of many diffusion curves. the theoretical analyses open the door to many other results, e.g., those regarding comparative statics across networks, payoffs, and cost distributions (when different actions vary in costs). future experimental and field data will hopefully complement these theoretical insights. a shortcoming of some of the theoretical analyses described in this chapter is that the foundation for modeling the underlying network is rooted in simple forms of random graphs in which there is little heterogeneity among nodes other than their connectivity. this misses a central observation from the empirical literature that illustrates again and again the presence of homophily, people's tendency to associate with other individuals who are similar to themselves. moreover, there are empirical studies that are suggestive of how homophily might impact diffusion, providing for increased local connectivity but decreased diffusion on a more global scale (see rogers (1995) for some discussion). beyond the implications that homophily has for the connectivity structure of the network, it also has implications for the propensity of individuals to be affected by neighbors' behavior: for instance, people who are more likely to, say, be immune may be more likely to be connected to one another, and, similarly, people who are more likely to be susceptible to infection may be more likely to be connected to one another. 29 furthermore, background factors linked to homophily can also affect the payoffs individuals receive when making decisions in their social network. enriching the interaction structure in that direction is crucial for deriving more accurate diffusion predictions. this is an active area of current study (e.g., see , bramoullã© and rogers (2010) , currarini, jackson, and pin (2006 , and peski (2008)). ultimately, the formation of a network and the strategic interactions that occur amongst individuals is a two-way street. developing richer models of the endogenous formation of networks, together with endogenous interactions on those networks, is an interesting direction for future work, both empirical and theoretical. 30 creating social contagion through viral product design: theory and evidence from a randomized field experiment, mimeo distinguishing influence based contagions from homophily driven diffusion in dynamic networks a field study on matching with network externalities, mimeo similarity and polarization in groups, mimeo the mathematical theory of infectious diseases and its applications who's who in networks. wanted: the key player social capital in the workplace: evidence on its formation and consequences the diffusion of microfinance in rural india a new product growth model for consumer durables place of work and place of residence: informal hiring networks and labor market outcomes who gets the job referral? evidence from a social networks experiment, mimeo chains of affection: the structure of adolescent romantic and sexual networks strategic experimentation in networks diversity and popularity in social networks, mimeo. calvã³-armengol, a the mechanism through which this occurs can be rooted in background characteristics such as wealth, or more fundamental personal attributes such as risk aversion. risk averse individuals may connect to one another and be more prone to protect themselves against diseases by, e.g., getting immunized there are also some models that study co-evolving social relationships and play in games with neighbors the spread of obesity in a large social network over 32 years communication and coordination in social networks a model of spatial conflict medical innovation: a diffusion study the role of the airline transportation network in the prediction and predictability of global epidemics invasion threshold in heterogeneous metapopulation networks socio-economic distance and spatial patterns in unemployment learning about a new technology: pineapple in ghana an economic model of friendship: homophily, minorities and segregation identifying the roles of race-based choice and chance in high school friendship network formation evolution of conventions in endogenous social networks, mimeo the role of information and social interactions in retirement plan decisions: evidence from a randomized experiment short-run subsidies and long-run adoption of new health products: evidence from a field experiment learning, local interaction, and coordination local conventions on random graphs on the evolution of random graphs the market maven: a diffuser of marketplace information building social capital through microfinance learning by doing and learning from others: human capital and technical change in agriculture complex networks and local externalities: a strategic approach, mimeo non-market interactions. nber working paper number 8053 crime and social interactions the 1/d law of giving r&d networks network formation and social coordination the strength of weak ties threshold models of collective behavior economic action and social structure: the problems of embeddedness getting a job: a study in contacts and careers hybrid corn: an exploration of the economics of technological change ergodic theorems for weakly interacting infinite systems and the voter model citizens, politics and social communication. cambridge studies in public opinion and political psychology job information networks, neighborhood effects and inequality social structure, segregation, and economic behavior social and economic networks relating network structure to diffusion properties through stochastic dominance. the b.e on the formation of interaction networks in social coordination games social games: matching and the play of finitely repeated games diffusion on social networks diffusion of behavior and equilibrium properties in network games correlated equilibria in graphical games homophily, selection, and socialization in adolescent friendships learning, mutation, and long run equilibria in games measuring trust in peruvian shantytowns, mimeo personal influence: the part played by people in the flow of mass communication graphical models for game theory a contribution to the mathematical theory of epidemics economic networks in the laboratory: a survey the people's choice: how the voter makes up his mind in a presidential campaign directed altruism and enforced reciprocity in social networks the dynamics of viral marketing contagion in complex social networks models for innovation diffusion (quantitative applications in the social sciences endogenous inequality in integrated labor markets with two-sided search birds of a feather: homophily in social networks patterns of influence: a study of interpersonal influence and of communications behavior in a local community global games: theory and applications heterogeneity and uniqueness in interaction games asymmetric effects in physician prescription behavior: the role of opinion leaders, mimeo. pastor-satorras epidemic dynamics and endemic states in complex networks design innovation and fashion cycles complementarities, group formation, and preferences for similarity workers and wages in an urban labor market the diffusion of hybrid seed corn in two iowa communities local network effects and network structure les lois de l'imitation: etude sociologique. elibron classics, translated to english in the laws of imitation social interactions, local spillovers and unemployment identifying formal and informal inuence in technology adoption with network externalities medical innovation revisited: social contagion versus marketing effort the evolution of conventions innovation diffusion in heterogeneous populations: contagion, social influence, and social learning key: cord-024981-yfuuirnw authors: severin, paul n.; jacobson, phillip a. title: types of disasters date: 2020-05-14 journal: nursing management of pediatric disaster doi: 10.1007/978-3-030-43428-1_5 sha: doc_id: 24981 cord_uid: yfuuirnw disasters are increasing around the world. children are greatly impacted by both natural disasters (forces of nature) and man-made (intentional, accidental) disasters. their unique anatomical, physiological, behavioral, developmental, and psychological vulnerabilities must be considered when planning and preparing for disasters. the nurse or health care provider (hcp) must be able to rapidly identify acutely ill children during a disaster. whether it is during a natural or man-made event, the nurse or hcp must intervene effectively to improve survival and outcomes. it is extremely vital to understand the medical management of these children during disasters, especially the use of appropriate medical countermeasures such as medications, antidotes, supplies, and equipment. skeleton as a result of incomplete calcification and active bone growth centers. protected organs, such as the lungs and heart, may be injured due to overlying fractures. cervical spine injuries can also be pronounced, as in patients with head trauma. in fact, spinal cord injury may be present without any radiographic abnormalities of the spine. finally, vital signs will vary based on the pediatric patient's age. this may be a pitfall during rapid evaluation by any nurse or hcp not accustomed to the care of children. younger pediatric patients have higher metabolic rates and, therefore, higher respiratory rates and heart rates. this can be a distinct disadvantage versus older pediatric patients when encountering similar diseases. an example is inhaled toxins (e.g., nerve agents and lung-damaging agents). infants and children will suffer greater toxicity since they inhale at a faster rate due to higher metabolic demands and thus, distribute the toxin more rapidly to various end-organs. understanding respiratory differences is essential to the management of the acutely ill pediatric patient. the most common etiology for cardiorespiratory arrest in children is respiratory pathology, typically of the upper airway. most of the airway resistance in children occurs in the upper airway. nasal obstruction can lead to severe respiratory distress due to infants being obligate nose breathers. their relatively large tongue and small mouth can lead to airway obstruction quickly, especially when the neuromuscular tone is abnormal such as during sedation or encephalopathy. in infants, physiologic (i.e., copious secretions) and pathologic (i.e., edema, vomitus, blood, and foreign body) factors will exaggerate this obstruction. securing the airway in such events can be quite challenging. typically, the glottis is located more anterior and cephalad. appropriate visualization during laryngoscopy can be further hampered by the prominent occiput that causes neck flexion and, therefore, reduces the alignment of visual axes. the omega or horseshoeshaped epiglottis in young infants and children is quite susceptible to inflammation and swelling. as in epiglottitis, the glottis becomes strangulated in a circumferential manner leading to dangerous supraglottic obstruction. children also have a natural tendency to laryngospasm and bronchospasm. finally, due to weaker cartilage in infants, dynamic airway collapse can occur especially in states of increased resistance and high expiratory flow. along with altered pulmonary compensation and compliance, a child may rapidly progress to respiratory failure and possibly arrest. cardiovascular differences are critical in the pediatric patient. typical physiological responses tend to allow compensation with seemingly normal homeostasis. with tachycardia and elevated systemic vascular resistance, younger pediatric patients can maintain normal blood pressure despite decreased cardiac output and poor perfusion (compensated shock). since children have less blood and volume reserve, they progress to this state quickly. in pediatric patients with multiorgan injury or severe gastrointestinal losses, these compensatory mechanisms are pushed to their limits. the unaccustomed hcp may be lulled into complacency since the blood pressure is normal. all the while, the pediatric patient's organs are being poorly perfused. once these compensatory mechanisms are exhausted, the patient rapidly progresses to hypotension and, therefore, hypotensive shock. if not reversed expeditiously, this may lead to irreversible shock, ischemia, multiorgan dysfunction, and death. pediatric patients with altered mental status pose significant problems. the differential diagnosis will be very broad in the comatose patient based on development alone. for example, younger pediatric patients can present with nonconvulsive status epilepticus (ncse) instead of generalized convulsive status epilepticus (gcse). in fact, ncse is more common among younger pediatric patients than gcse, especially in those from 1 to 12 months of age. furthermore, many of them are previously well without preexisting diseases such as epilepsy. other disease states may include poisoning, inborn errors of metabolism, meningitis, and other etiologies of encephalopathy. using the modified pediatric glasgow coma scale (gcs) is the cornerstone when evaluating the young pediatric patient when they are preverbal. pupillary response, external ocular movements, and gross motor response may be challenging to evaluate in a developmentally young or delayed pediatric patient. pediatric traumatic brain injury is extremely devastating. whether considered accidental (motor vehicle crash) or nonaccidental (abusive head trauma), evaluation of the neurological status of the acutely injured pediatric patient can be problematic, especially the gcs. some prefer to use the avpu system (alert, responds to verbal, responds to pain, and unresponsive). due to the disproportionately larger head and weaker neck muscles, there is more risk of acceleration-deceleration injuries (fall from a significant height, vehicular ejection, and abusive head trauma). furthermore, the softer skull, dural structural differences, and vessel supply will place the pediatric patient at risk for brain injury and intracranial hemorrhage. finally, due to pediatric brain composition, the risk of diffuse axonal injury and cerebral edema is much higher. although spinal cord injury is rare in young pediatric patients, morbidity and mortality are significant. in pediatric patients less than 9 years of age, the most commonly seen injuries are in the atlas, axis, and upper cervical vertebrae. in young pediatric patients, spinal injuries tend to be anatomically higher (cervical) versus adolescents (thoracolumbar). furthermore, congenital abnormalities, such as atlantoaxial abnormalities (trisomy 21), may exaggerate the process. the clinical presentation of spinal cord injury varies in young pediatric patients due to ongoing development. laxity of ligaments, wedge-shaped vertebrae, and incomplete ossification centers contribute to specific patterns of injuries. finally, spinal cord injury without radiographic abnormality (sciwora) may result. because of the disproportionately larger head, weaker neck muscles, and elasticity of the spine, significant distraction and flexion injury of the spinal cord may occur without apparent ligament or bony disruption (hilmas et al. 2008; jacobson and severin 2012; severin 2011). motor skills develop from birth. gross and fine motor milestones are achieved in a predictable manner and must be assessed during each hcp encounter. cognitive development will follow a similar pattern of maturation. the development of these skills can often predict injuries and their extent. for example, consider a house fire. a young infant, preschooler, and adolescent are sleeping upstairs in house when a fire breaks out in the middle of the night. the smoke detectors begin to alarm. each child is awoken by the ensuing noise and chaos. based on the development, the adolescent will most likely make it out of the house alive. he will comprehend the threat, run down the stairs, and exit the house without delay. smoke inhalation may be minimal. if it is a middle adolescent, an attempt may be made to jump out of the window leading to multiple blunt trauma with or without traumatic brain injury. the preschooler most likely will be too scared and not understand how to escape. tragically, he may hide under a bed or in a closet. when the firefighters arrive and search the house, the preschooler may remain silent because of fear, especially of strangers in the house. he will most likely succumb to thermal injuries along with the effects of carbon monoxide and die. as far as the infant, he cannot walk, climb, crawl, or run. furthermore, he cannot scream for help or know how to escape. as the smoke engulfs the room, he will most likely suffer severe smoke inhalation injury including extensive carbon monoxide toxicity along with thermal injuries and die. this example also points out another important difference in pediatric patients: their dependence on caregivers. when considering neonates, for example, their entire existence depends on the caregiver, including feeding, changing of diapers, nurturing, and environmental safety. these dynamics are essential to the pediatric patient's health and survival, especially during a disaster. another aspect of development is the attainment of language skills. this, too, develops over time in a predictable fashion. one of the biggest challenges in pediatrics is the lack of the patient's ability to verbally convey complaints. as described above, verbal milestones vary among the different age ranges of the pediatric patient. hcps are often faced with a caregiver's subjective assessment of the problem. although it can be revealing and informative, this may not be available in an acute crisis situation. it will take the astute hcp to determine, for example, if an inconsolably crying infant is in pain from a corneal abrasion or something more life-threatening such as meningitis. this can also be a challenging task in a teenager, especially during middle adolescence. an hcp will have to determine, for example, if the seemingly lethargic middle adolescent is intoxicated with illicit drugs or has diabetic ketoacidosis. finally, the hcp will have to address developmental variances among their pediatric patients and any comorbid features. young pediatric patients can regress developmentally during any illness or injury. this is especially seen in patients with chronic medical conditions (cancer) or during prolonged hospitalization with rehabilitation (multisystem trauma). furthermore, those pediatric patients with developmental and intellectual disabilities, for example, will be difficult to evaluate based on the effects of their underlying pathology. these pediatric patients typically have unique variances in their physical exams (jacobson and severin 2012; severin 2011) . please refer to chap. 7 for more detailed information on pediatric development. pediatric patients will often reflect the emotional state of their caregiver. they take verbal and physical cues from their caregiver. at times, this may also occur in the presence of a nurse or hcp. the psychological impact of illness will vary greatly with the child's development and experience. children tend to have a greater vulnerability to post-traumatic stress disorder especially with disaster events. furthermore, they are highly prone to becoming psychiatric casualties despite the absence of physical injury to themselves. and as any pediatric hcp can tell you, the younger pediatric patients tend to also have greater levels of anxiety, especially while preparing for invasive procedures such as phlebotomy and intravenous line placement (hilmas et al. 2008; jacobson and severin 2012; severin 2011) . please refer to chap. 12 for more detailed content on mental health. the world health organization and the pan american health organization define a disaster as "an event that occurs in most cases suddenly and unexpectedly, causing severe disturbances to people or objects affected by it, resulting in the loss of life and harm to the health of the population, the destruction or loss of community property, and/or severe damage to the environment. such a situation leads to disruption in the normal pattern of life, resulting in misfortune, helplessness, and suffering, with adverse effects on the socioeconomic structure of a region or a country and/or modifications of the environment to such an extent that there is a need for assistance and immediate outside intervention" (lynch and berman 2009 ). types of disasters usually fall into two broad categories: natural and man-made. natural disasters are generally associated with weather and geological events, including extremes of temperature, floods, hurricanes, earthquakes, tsunamis, volcanic eruptions, landslides, and drought. naturally occurring epidemics, such as the 2009 h1n1, 2014 ebola, and 2019 novel coronavirus outbreaks, are often included in this category. man-made disasters are usually associated with a criminal attack such as an active shooter incident, or a terrorist attack using weapons such as explosive, biological, or chemical agents. however, man-made disasters can also refer to human-based technological incidents, such as a building or bridge collapse, or events related to the manufacture, transportation, storage, and use of hazardous materials, such as the 1986 chernobyl radiation leak and the 1984 bhopal toxic gas leak. even though disasters can be primarily placed into any of these two categories, they can often impact each other and compound the magnitude of any disaster incident (united states department of homeland security, office of inspector general 2009). a prime example is the march 2011 tohoku earthquake leading to a tsunami (natural) that triggered the fukushima daiichi nuclear disaster (man-made). disasters can also be characterized by the location of such an event. internal disasters are those incidents that occur within the health care facility or system. employees, physical plant, workflow and operations of the clinic, hospital, or system can be disrupted. external disasters are those incidents that occur outside of the health care facility or system. this impacts the community surrounding the facility, proximally or distally, but does not directly threaten the facility or its employees. as with natural and man-made disasters, internal and external disasters can impact each other. for example, an overflow of patients during a high census period may lead to the shutdown of the hospital to any new patients (internal disaster). this will place the hospital on bypass and possibly stress other hospitals in the community beyond their means (external disaster). a terrorist event, such as the release of sarin in a subway system during a busy morning commute, can lead to massive disruption in the community (external disaster). all the victims of the attack will seek medical care at nearby hospitals, possibly overwhelming the health care staff and depleting critical resources (internal disaster). characterization of disasters by geography (local, state, national, and international) can also be used. again, no matter the site of the incident, a disaster in one area could easily create a disaster in another geographical region. for example, a factory and its community could be ravaged by a hurricane (local disaster). if this is the only factory in the world to produce a certain medication, this could lead to critical shortages to hospitals all around the world (international disaster). the term "disaster preparedness" has been used over the years as a way to describe efforts to manage any disaster event. however, preparedness is only one aspect of the process. the use of the term disaster planning is more appropriate. it considers all aspects needed for an effective effort and is dependent on additional phases, not just preparedness. national preparedness efforts, including planning, are now informed by the presidential policy directive (ppd) 8 that was signed by the president in march 2011 and describes the nation's approach to preparedness (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013; united states department of homeland security 2018b). a recommended method for disaster preparedness efforts is the utilization of an "all-hazards" model of emergency management (adini et al. 2012; waugh 2000) . the four overlapping phases of the model include mitigation, preparedness, response, and recovery. the mitigation phase involves "activities designed to prevent or reduce losses from a disaster" (waugh 2000) . examples include land use planning in flood plains, structural integrity measures in earthquake zones, and deployment of security cameras. the preparedness phase includes the "planning of how to respond in an emergency or a disaster, and developing capabilities for more effective response" (waugh 2000) . examples include training programs for emergency responders, drills and exercises, early warning systems, contingency planning, and development of equipment and supply caches. up to this point, all planning efforts are proactive and not reactive. often times, a hazard analysis is conducted to delineate areas of strengths and identify potential risks. it helps in "the identification of hazards, assessment of the probability of a disaster, and the probable intensity and location; assessment of its potential impact on a community; the property, persons, and geographic areas that may be at risk; and the determination of agency priorities based on the probability level of a disaster and the potential losses" (waugh 2000) . after a disaster or emergency incident occurs, the response phase, or "immediate reaction to a disaster", (waugh 2000) begins. examples include mass evacuations, sandbagging buildings and other structures, providing emergency medical services, firefighting, and restoration of public order. in some situations, the response period may be a short (e.g., house fire), intermediate (e.g., bomb detonation), or extended (e.g., pandemic influenza) duration. after a period of time, the recovery phase follows. these are "activities that continue beyond the emergency period to restore lifelines" (waugh 2000) . examples include the provision of temporary shelter, restoration of utilities such as power, critical stress debriefing for responders, and victims, job assistance and small business loans, and debris clearance. recovery always seems to be the most unpredictable; it may take days to months to years. as demonstrated with recent hurricanes harvey, irma, and maria in 2017, the most affected regions are still in the phase of recovery and may be along a prolonged track as hurricane katrina in 2005. as mentioned, the early phases of planning (mitigation and preparedness) truly hinge upon the environment or community surrounding the health care site (e.g., clinic, hospital, or long-term care facility). identification of potential hazards and risks is a key step in disaster planning. using a hazard vulnerability assessment (hva) or a threat and hazard identification and risk assessment (thira) can provide a basis for mitigation and prevention tasks. an hva/thira emphasizes which types of natural or man-made disasters are likely to occur in a community (e.g., tornado, flood, chemical release, or terrorist event). they further highlight the impact those disasters may have on the community and any capabilities that are in place that may lessen the effects of the disaster (illinois emergency medical services for children 2018). a basic principle of the hva methodology is to determine the risk of such an event or attack occurring at a given hospital or hospital system. simply, the risk is a product of the probability of an event and the severity of such an event if it occurs (risk = probability ã� severity). however, there are many complexities in quantifying terrorism risk (waugh 2005; woo 2002) . it is important to note that in some circumstances, exposure may need to be included in the equation (risk = probability ã� severity ã� exposure), but usually for operational risk management applications (mitchell and decker 2004) . at any rate, issues to consider for the probability of an event occurring include, but are not limited to, geographic location and topography, proximity to hazards, degree of accessibility, known risks, historical data, and statistics of various manufacturer/vendor products. severity, on the other hand, is dependent on the gap between the magnitude of an event and mitigation for the given event (severity = magnitude -mitigation). magnitude varies upon the impact of the event to humans, property, and/or business. mitigation varies upon the development of internal and external readiness before a disaster strikes. as one can surmise, if the magnitude of the event outstrips the mitigation, the event is considered a threatening hazard. once the hva is completed, the health care site should immediately prioritize planning efforts for the top 5-10 hazards and develop plans accordingly. all other identified hazards must also be addressed to ensure a broad and robust disaster plan. it is important to realize that local and regional entities also perform comprehensive hvas. a concerted analysis among a hospital and key community stakeholders is optimal for a coordinated plan. an hva/thira contains both quantitative and qualitative components. specific tools have been developed through private and public organizations (e.g, fema) that can help in the analysis (united states department of homeland security, federal emergency management agency 2001). using these tools as a guide, the entity can determine what types of hazards have a high, medium, or low probability of occurring within specific geographic boundaries. typically, these tools do not have components specific to children or other at-risk populations. however, the tools can be adapted either directly by adding children to specific hazards or ensuring considerations specific to children are incorporated into the hva/thira calculations. the hva/thira should be reviewed and updated minimally on an annual basis to identify changing or external circumstances. this includes conducting a pediatric-specific disaster risk assessment to identify where children congregate and their risks (e.g., schools, popular field trip designations, summer camps, houses of worship, and juvenile justice facilities) (illinois emergency medical services for children 2018). of note, hva techniques have been utilized for pediatric-specific disaster plans. having a separate pediatric hva (phva) is crucial to a well-rounded and robust health care disaster plan. first, it demonstrates the extent of the pediatric population in the community. it is estimated that 25% of the population fits within the age range of pediatric patients. in some situations, it may be more. during the performance of a phva, it was demonstrated that 29% of the community was less than 19 years of age (jacobson and severin 2012) . second, a phva increases the situational awareness of those tasked to plan for disasters that involve children and adolescents. often times, children and adolescents are excluded from local and regional disaster plans. the unique vulnerabilities of pediatric patients will demand appropriate drills, exercises, equipment, medications, and expertise. thirdly, identifying pediatric risks in a community will help prioritize efforts of planning, especially in those hospitals not accustomed to caring for pediatric patients. finally, a phva helps to develop a framework for global pediatric disaster planning. this can extend beyond a local community and actually advance city, state, regional, and national disaster planning efforts. there has been a development of web-based tools to simplify and enhance the phva process (jacobson and severin 2012) . after an hva/thira has been completed, the results should be used to help direct and plan drills/exercises based on high impact and high probability threats. it is advised to conduct an hva/thira on an annual basis to assess specific threats unique to your organization's physical structure as well as the surrounding geographic environment. it will also provide insight into whether there is an improvement in previous planning efforts. completion of a population assessment that provides a demographic overview of the community with a breakdown of the childhood population is strongly recommended in conjunction with the hva/ thira. collaborating with other community partners, such as local health departments and emergency management agencies, can assist an organization with the conduction of a comprehensive hva/thira (illinois emergency medical services for children 2018). please see chap. 13 for further information on hospital planning. pediatric supplies, equipment, and medications will be scarce during a disaster. it will become more of an issue if the health care facility is not accustomed to caring for acutely ill pediatric patients. this will be further exacerbated by a massive surge of acutely ill pediatric patients, a widespread or prolonged disaster, and supply line disruptions. to protect the health security of children and families during a public health emergency, the assistant secretary for preparedness and response (aspr) manages and maintains the strategic national stockpile (sns), a cache of medical countermeasures for rapid deployment and use in response to a public health emergency or disaster (fagbuyi et al. 2016) . various pediatric-specific supplies and countermeasures are included in the sns. maintaining a supply of medications and medical supplies for specific health threats allows the stockpile to respond with the right product when a specific disease or agent is known. if a community experiences a large-scale public health incident in which the disease or agent is unknown, the first line of support from the stockpile is to send a broad-range of pharmaceuticals and medical supplies. place and martin 2012) . the emergency equipment and supply lists can easily be adapted for any pediatric disaster emergency (place and martin 2012) or incident requiring pediatric mass critical care (desmond et al. 2011) . ageappropriate nutrition, hygiene, bedding, and toys/distraction devices should also be available (illinois emergency medical services for children 2013) (tables 5.3 and 5.4). endotracheal tubes â�¢ uncuffed: 2.5 and 3.0 mm â�¢ cuffed or uncuffed: 3.5, 4.0, 4.5, 5.0, and 5.5 mm â�¢ cuffed: 6.0, 6.5, 7.0, 7.5, and 8.0 mm feeding tubes (5f and 8f) laryngoscope blades curved: 2 and 3; straight: 0, 1, 2, and 3 laryngoscope handle magill forceps (pediatric and adult) nasopharyngeal airways (infant, child, and adult) oropharyngeal airways (sizes 0-5) stylets for endotracheal tubes (pediatric and adult) suction catheters (infant, child, and adult) tracheostomy tubes (sizes 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, and 5.5 mm) yankauer suction tip bag-mask device (manual resuscitator), self-inflating (infant size: 450 ml; adult size: 1000 ml) clear oxygen masks (standard and nonrebreathing) for an infant, child, and adult masks to fit bag-mask device adaptor (neonatal, infant, child, and adult sizes) nasal cannulas (infant, child, and adult) nasogastric tubes (sump tubes): infant (8f), child (10f), and adult (14f-18f) laryngeal mask airway a vascular access arm boards (infant, child, and adult sizes) catheter over-the-needle device (14-24 gauge) intraosseous needles or device (pediatric and adult sizes) intravenous catheter-administration sets with calibrated chambers and extension tubing and/or infusion devices with ability to regulate rate and volume of infusate umbilical vein catheters (3.5f and 5.0f) b central venous catheters (4.0f-7.0f) intravenous solutions to include normal saline, dextrose 5% in normal saline, and dextrose 10% in water fracturemanagement devices extremity splints, including femur splints (pediatric and adult sizes) spine-stabilization method/devices appropriate for children of all ages c (continued) laryngeal mask airways could be shared with anesthesia but must be immediately accessible to the ed b feeding tubes (size 5f) may be used as umbilical venous catheters but are not ideal. a method for securing the umbilical catheter, such as an umbilical tie, should also be available c a spinal stabilization device is one that can stabilize the neck of an infant, child, or adolescent in a neutral position when a pediatric disaster victim presents acutely ill to the hospital, various emergency interventions will be needed to stabilize the patient. evaluation of the pediatric patient should include a primary survey (abcde), secondary survey (focused sample history and focused physical examination), and diagnostic assessments (laboratory, radiological, and other advanced tests). this will guide further therapeutic interventions. particular attention should be given to the identification of respiratory and/or circulatory derangements of the child, including airway obstruction, respiratory failure, shock, and cardiopulmonary failure. interventions will be based on physiologic derangements of the pediatric patient and determined by the scope of practice and protocols, such as standard resuscitation algorithms for neonatal (american academy of pediatrics and american heart association et al. 2016) and pediatric (american heart association 2016) victims. the hcp must be knowledgeable of various emergency medications (table 5 .1) used for children, the appropriate dosages and their mechanism of action, any potential side effects, and drug/drug interactions. other medications, such as antibiotics, antidotes, or countermeasures, may be needed as well. pharmacologic therapy should be initiated immediately based on clinical suspicion and not delayed due to pending laboratory tests (e.g., antibiotics for presumed infection/sepsis or antidotes for suspected nerve agents). dosages should be based on the patient's weight or a length-based weight system. (montello et al. 2006) or hard copy countermeasure manuals may be more practical, especially during a disaster incident when computer service or internet access may be unreliable. in 1988, the centre for research on the epidemiology of disasters (cred) launched the emergency events database (em-dat). em-dat was created with the initial support of the world health organization (who) and the belgian government. the main objective of the database is to serve the purposes of humanitarian action at national and international levels. the initiative aims to rationalize decision-making for disaster preparedness as well as provide an objective base for vulnerability assessment and priority setting. em-dat contains essential core data on the occurrence and effects of over 22,000 mass disasters in the world from 1900 to the present day. the database is compiled from various sources, including united nation agencies, nongovernmental organizations (ngos), insurance companies, research institutes, and press agencies (cred 2019). as described in the cred report entitled natural disasters 2017: lower mortality, higher cost, a disaster is entered into the database if at least one of the following criteria is fulfilled: 10 or more people reported killed; 100 or more people reported affected; declaration of a state of emergency; and/or call for international assistance (cred 2018). in economic losses, poverty and disasters 1998-2017: cred/unisdr report, the cred defines a disaster as "a situation or event which overwhelms local capacity, necessitating a request at national or international level for external assistance; an unforeseen and often sudden event that causes great damage, destruction and human suffering" (cred 2018). the cred em-dat classifies disasters according to the type of hazard that triggers them. the two main disaster groups are natural and technological disasters. there are six natural disaster subgroups. geophysical disasters originate from the solid earth and include earthquake (ground movement and tsunami), dry mass movement (rock fall and landslides), and volcanic activity (ash fall, lahar, pyroclastic flow, and lava flow). lahar is a hot or cold mixture of earthen material flowing on the slope of a volcano either during or between volcanic eruptions. meteorological disasters are caused by short-lived, micro-to meso-scale extreme weather and atmospheric conditions that last from minutes to days and include extreme temperatures (cold wave, heat wave, and severe winter conditions such as snow/ice or frost/ freeze), fog, and storms. storms can be extra-tropical, tropical, or convective. convective storms include derecho, hail, lightning/thunderstorm, rain, tornado, sand/dust storm, winter storm/blizzard, storm/surge, and wind. derecho is a widespread and usually fast-moving windstorm associated with convection/convective storm and includes downburst and straight-line winds. hydrological disasters are caused by the occurrence, movement, and distribution of surface/subsurface freshwater and saltwater and include floods, landslides (an avalanche of snow, debris, mudflow, and rockfall), and wave action (rogue wave and seiche). flood types can be coastal, riverine, flash, or ice jam. climatological disasters are caused by longlived, meso-to macro-scale atmospheric processes ranging from intraseasonal to multidecadal climate variability and include drought, glacial lake outburst, and wildfire (forest fire, land fire: brush, bush, or pasture). biological disasters are caused by the exposure to living organisms and their toxic substances or vectorborne diseases that they may carry and include epidemics (viral, bacterial, parasitic, fungal, and prion), insect infestation (grasshopper and locust), and animal accidents. extraterrestrial disasters are caused by asteroids, meteoroids, and comets as they pass near-earth, enter earth's atmosphere, and/or strike the earth, and by changes in the interplanetary conditions that affect the earth's magnetosphere, ionosphere, and thermosphere. types include impact (airbursts) and space weather (energetic particles, geomagnetic storm, and shockwave) events (cred 2019). there are three technological disaster subgroups. industrial accidents include chemical spills, collapse, explosion, fire, gas leak, poisoning, radiation, and oil spills. a chemical spill is an accidental release occurring during the production, transportation, or handling of hazardous chemical substances. transport accidents include disasters in the air (airplanes, helicopters, airships, and balloons), on the road (moving vehicles on roads or tracks), on the rail system (train), and on the water (sailing boats, ferries, cruise ships, and other boats). miscellaneous accidents vary from collapse to explosions to fires. collapse is an accident involving the collapse of a building or structure and can either involve industrial structures or domestic/nonindustrial structures (cred 2019). technological disasters are considered man-made, but as suggested by their subgroup, they are accidental and not intentional. the united nations office for disaster risk reduction (unisdr) and cred report, economic losses, poverty, and disasters 1998-2017, reviews global natural disasters during that time period, their economic impact, and the relationship with poverty. between 1998 and 2017, climate-related and geophysical disasters killed 1.3 million people and left a further 4.4 billion injured, homeless, displaced, or in need of emergency assistance. although the majority of fatalities were due to geophysical events, mostly earthquakes and tsunamis, 91% of all disasters was caused by floods, storms, droughts, heatwaves, and other extreme weather events. the financial impact was staggering. in 1998-2017, disaster-hit countries reported direct economic losses valued at us$ 2908 billion, of which climate-related disasters caused us$ 2245 billion or 77% of the total. this was up from 68% (us$ 895 billion) of losses (us$ 1313 billion) reported between 1978 and 1997. overall, reported losses from extreme weather events rose by 151% between these two 20-year periods. in absolute monetary terms, over the last 20-years, the usa recorded the biggest losses (us$ 945 billion), reflecting high asset values as well as frequent events. china, by comparison, suffered a significantly higher number of disasters than the usa (577 vs. 482) but lower total losses (us$ 492 billion) (cred 2018) (figs. 5.1, 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8 and 5.9 in 2018, 281 climate-related and geophysical incidents in the world were estimated with 10,733 deaths and over 60 million people impacted. indonesia recorded approximately half of the deaths with india accounting for half of those impacted by disasters. notable features of 2018 were intense seismic activity in indonesia, a series of disasters in japan, floods in india, and an eventful year for both volcanic activity and wildfires. however, an ongoing trend of lower death tolls from previous years continued into 2018 (centre for research on the epidemiology of disasters (cred) and united nations office for disaster risk reduction (unisdr) 2019) (tables 5.5, 5.6, 5.7, 5.8, 5.9, 5.10 and 5.11 there are no specific deviations when medically managing children after a natural disaster. according to sirbaugh and dirocco (2012) "small-scale mass casualty incidents occur daily in the united states. few present unusual challenges to the local medical systems other than in the number of patients that must be treated at one time. except in earthquakes, explosions, building collapses, and some types of terrorist attacks, the same holds true for large-scale disasters. sudden violent disaster mechanisms can produce major trauma cases, including patients needing field amputations or management of crush syndrome. for the most part, medicine after a disaster is much the same as it was before the disaster, with more minor injuries, more people with exacerbations of their chronic illnesses, and number of patients seeking what is ordinarily considered primary care. this is true for children and adults." it should be noted, however, that children have a predisposition to illness and injury after natural disasters. the hcp must be able to identify any health problems and treat the child effectively and efficiently while utilizing standard resuscitation protocols as indicated. traumatic injuries may be seen after any natural disaster. the injuries can range from minor scrapes and bruises to major blunt trauma or traumatic brain injury. children are at increased risk for injury since adults are distracted by recovery efforts and may not be able to supervise them closely. the environment may not be safe due to environmental hazards, such as collapsed buildings, sinkholes, and high water levels. dangerous equipment used during relief efforts may be present, such as heavy earth moving equipment, chainsaws, and power generators. hazardous chemicals, such as gasoline and other volatile hydrocarbons, may be readily accessible or taint the environment. without suitable shelter, children are also exposed to weather, animals, and insects (sirbaugh and dirocco 2012) . infectious diseases may also pose a problem to children after a natural disaster. infectious patterns will persist during a disaster based on the season and time of year. there may be outbreaks or epidemics of highly contagious infections (e.g., influenza, respiratory syncytial virus, streptococcus pyogenes) due to mass sheltering of children and families. poor nutrition or decreased availability of food may lower their resistance against infections. various water-borne or food-borne diseases may cause illnesses in children. poor hygiene and mass shelter environments may exacerbate these illnesses. immunized children should be protected against common preventable diseases after a natural disaster but still could be a problem in mass groups that are not completely or appropriately immunized. after the 2010 haiti earthquake, there were increased cases of diarrhea, cholera, measles, and tetanus in children months after the earthquake despite some level of vaccination (sirbaugh and dirocco 2012) . children are at risk for various environmental emergencies. austere environments will impact children greatly. heat exposure coupled with minimal access to drinkable water may lead to severe dehydration. exposure to the cold may lead to frostbite or hypothermia. children are at risk for carbon monoxide toxicity due to generator use or natural gas poisoning due to disrupted gas lines. there is always a risk for thermal injury due to the use of candles and other flame sources. exposure to animals (snakes) and insects (spiders) may increase the risk of envenomation. submersion injury and drowning incidents may escalate. this will be due to lack of supervision of children around storm drains, newly formed bodies of water, or rushing waters of storm diversion systems (sirbaugh and dirocco 2012) . mental health issues are often seen in children after natural disasters. even though a child may not be injured, they may become "psychiatric casualties" due to the horrific sights they have seen during or after the disaster. children and adolescents with behavioral or psychiatric problems may experience worsening symptoms and signs due to stress, trauma, disruption of routines, or availability of medications. this is often exacerbated if the parent, guardian, caregiver, or hcp is also having difficulty coping with the stress of the disaster. in general, the most common mental health problem in children is a post-traumatic stress disorder. however, separation anxiety, obsessive-compulsive symptoms, and severe stranger anxiety can also be seen in children after a traumatic event (sirbaugh and dirocco 2012) . see chap. 12 for more detailed information. terrorism impacts children and families all around the world (tables 5.12 and 5.13). after the events of 9/11, much attention has been given to the possibility of another mass casualty act of terrorism, especially with weapons of mass destruction, that include chemical, biological, nuclear, radiological, and explosive devices (cbnre), or other forms of violence such as active shooter incidents and mass shootings (jacobson and severin 2012) . since then, other incidents, both foreign and domestic, have involved children and complicates the concept of and the response to terrorism. johnston (2017) said it best in his review of terrorist and criminal attacks targeting children: "one of the more accepted defining characteristics of terrorism is that it targets noncombatants including men, women, and children. however, terrorist attacks specifically targeting children over other noncombatants are uncommon. this is for the same reason that most terrorists have historically avoided mass casualty terrorism: the shock value is so great that such attacks erode support for the terrorists' political objectives. the 9/11 attacks represent an increasing trend in mass casualty terrorism. at the same time, policymakers are examining this evolving threat, they must increasingly consider the threat of terrorist attacks targeting children." based on historical events, it is clear infants, toddlers, children, and adolescents have been victims of terrorism. this global trend of terrorists targeting children seems to be escalating (johnston 2017) . therefore, it is imperative to understand terrorism and ways it impacts the children and families served by the health care community. combs (2018) defines terrorism as "an act of violence perpetrated on innocent civilian noncombatants in order to evoke fear in an audience". however, she goes on to argue that to become an operational definition, there must also be the addition of a "political purpose" of the violent act. therefore, "terrorism, then, is an act composed of at least four crucial elements: 1) it is an act of violence, 2) it has a political motive or goal, 3) it is perpetrated against civilian noncombatants, and 4) it is staged to be played before an audience whose reaction of fear and terror is the desired result." (combs 2018) . there are different typologies of terrorism. at least five types of terror violence have been suggested by feliks gross: "mass terror is terror by a state, where the regime coerces the opposition in the population, whether organized or unorganized, sometimes in an institutionalized manner. dynastic assassination is an attack on a head of state or a ruling elite. random terror involves the placing of explosives where people gather (such as post offices, railroads, and cafes) to destroy whoever happens to be there. focused random terror restricts the placing of explosives, for example to where significant agents of oppression are likely to gather. finally, tactical terror is directed solely against the ruling government as a part of a 'broad revolutionary strategic plan'" (combs 2018 ). an additional typology offered is "lone wolf terror which involves someone who commits violent acts in support of some group, movement, or ideology, but who does stand alone, outside of any command structure and without material assistance from any group" (combs 2018) . martin (2017) reviews eight different terrorism typologies in the ever shifting, multifaceted world of modern terrorism. the new terrorism "is characterized by the threat of mass casualty attacks from dissident terrorist organizations, new and creative configurations, transnational religious solidarity, and redefined moral justifications for political violence" (martin 2017) . state terrorism is "committed by governments against perceived enemies and can be directed externally against adversaries in the international domain or internally against domestic enemies" (martin 2017) . dissident terrorism is "committed by nonstate movements and groups against governments, ethno-national groups, religious groups, and other perceived enemies" (martin 2017) . religious terrorism is "motivated by an absolute belief that an otherworldly power has sanctioned and commanded the application of terrorist violence for the greater glory of the faithâ�¦[it] is usually conducted in defense of what believers consider to be the one true faith" (martin 2017) . ideological terrorism is "motivated by political systems of belief (ideologies), which champion the self-perceived inherent rights of a particular group or interest in opposition to another group or interest. the system of belief incorporates theoretical and philosophical justifications for violently asserting the rights of the championed group or interest" (martin 2017) . international terrorism "spills over onto the world's stage. targets are selected because of their value as symbols of international interests, either within the home country or across state boundaries" (martin 2017) . criminal dissident terrorism "is solely profit-driven, and can be some combination of profit and politics. for instance, traditional organized criminals accrue profits to fund their criminal activity and for personal interests, while criminalpolitical enterprises acquire profits to sustain their movement" (martin 2017) . gender-selective terrorism "is directed against an enemy population's men or women because of their gender. systematic violence is directed against men because of the perceived threat posed by males as potential soldiers or sources of opposition. systematic violence is directed against women to destroy an enemy group's cultural identity or terrorize the group into submission" (martin 2017) . the all-hazards national planning scenarios are an integral component of dhs's capabilities-based approach to implementing homeland security presidential directive 8: national preparedness (hspd-8). the national planning scenarios are planning tools and are representative of the range of potential terrorist and natural disasters and the related impacts that face the nation. the federal interagency community has developed 15 all-hazards planning scenarios for use in national, federal, state, and local homeland security preparedness activities. the objective was to develop a minimum number of credible scenarios to establish the range of response requirements to facilitate disaster planning (dhs 2006) (table 5 .14). twelve of the scenarios represent terrorist attacks while three represent natural disasters or naturally occurring epidemics. this ratio reflects the fact that the nation has recurring experience with natural disasters but faces newfound dangers, including the increasing potential for use of weapons of mass destruction by terrorists. the scenarios form the basis for coordinated federal planning, training, exercises, and grant investments needed to prepare for all hazards. dhs employed the scenarios as the basis for a rigorous task analysis of prevention, protection, response, and recovery missions and identification of key tasks that supported the development of essential all-hazards capabilities (united states department of homeland security, federal emergency management agency 2019) (table 5 .15). each of the 15 scenarios follows the same outline to include a detailed scenario description, planning considerations, and implications. for each of the 12 terrorismrelated scenarios, fema national preparedness directorate (npd) partnered with dhs office of intelligence and analysis (i&a) and other intelligence community and law enforcement experts to develop and validate prevention prequels. the prequels provide an understanding of terrorists' motivation, capability, intent, tactics, techniques and procedures, and technical weapons data. the prequels also provide a credible adversary based on known threats to test the homeland security community's ability to understand and respond to indications and warnings of possible terrorist attacks (united states department of homeland security, federal emergency management agency 2019). a chemical agent of terrorism is defined as any chemical substance intended for use in military operations to kill, seriously injure, or incapacitate humans (or animals) through its toxicological effects. chemicals excluded from this list are riot-control agents, chemical herbicides, and smoke/flame materials. chemical agents are classified as toxic agents (producing injury or death) or incapacitating agents (producing temporary effects). toxic agents are further described as nerve agents (anticholinesterases), blood agents (cyanogens), blister agents (vesicants), and lung-damaging agents (choking agents). incapacitating agents include stimulants, depressants, psychedelics, and deliriants (banks 2014; departments of the army, the navy, and the air force, and commandant, marine corps 1995). nerve agents are organophosphate anticholinesterase compounds. they are used in various insecticide, industrial, and military applications. military-grade agents include tabun (ga), sarin (gb), soman (gd), cyclosarin (gf), venom x (vx), and the novichok series. these are all major military threats. the only known battlefield use of nerve agents was the iraq-iran war. however, other nerve agent incidents, such as the 1995 tokyo subway attack (sarin), the chemical attacks in syria (chlorine, sarin, mustard), and the attempted assassination of sergei skripal in salisbury, uk (novichok), support that civilian threats also exist. nerve agents are volatile chemicals and can be released in liquid or vapor form. however, the liquid form can become vapor depending upon its level of volatility (e.g, g-agents are more volatile than vx). the level of toxicity depends on the agent, concentration of the agent, physical form, route and length of exposure, and environmental factors (temperature and wind) (tables 5.16 and 5.17). nerve agents exert their effects by the inhibition of esterase enzymes. acetylcholinesterase inhibition prevents the hydrolysis of acetylcholine. the clinical result is a cholinergic crisis and subsequent overstimulation of muscarinic and nicotinic receptors throughout the body including the central nervous system. clinical muscarinic responses include sludge (salivation, lacrimation, urination, defecation, gastrointestinal distress, and emesis) and dumbels (diarrhea, urinary incontinence, miosis/muscle fasciculation, bronchorrhea/bronchospasm/bradycardia, emesis, lacrimation, and salivation). nicotinic responses vary by site. preganglionic sympathetic nerve stimulation produces mydriasis, tachycardia, hypertension, and pallor. however, stimulation at the neuromuscular junction leads to muscular fasciculation and cramping, weakness, paralysis, and diaphragmatic weakness. central nervous system presentations range from anxiety and restlessness to seizures, coma, and death (banks 2014; rotenberg and newmark 2003; rotenberg 2003b ). pediatric manifestations (table 5 .19) may vary from the classic clinical responses due to their unique vulnerabilities (hilmas et al. 2008 ): â�¢ children may manifest symptoms earliest and possibly more severe presentations. â�¢ could be hospitalized for similarly related illnesses and diseases. â�¢ smaller mass. â�¢ lower baseline cholinesterase activity. â�¢ tendency to bronchospasm. â�¢ pediatric airway and respiratory differences. â�¢ altered pulmonary compensation. â�¢ lower reserves of cardiovascular system and fluids. â�¢ isolated central nervous system signs (stupor, coma). â�¢ less miosis. â�¢ vulnerability to seizures and neurotransmitter imbalances (excitability). â�¢ immature metabolic systems. differential diagnoses include upper or lower airway obstruction, bronchiolitis, status asthmaticus, cardiogenic shock, acute gastroenteritis, seizures, and poisonings (carbon monoxide, organophosphates, and cyanide). diagnostic tests include acetylcholinesterase levels, red blood cell cholinesterase levels, and an arterial blood gas. treatment (tables 5.20 and 5.21) includes decontamination (reactive skin decontamination lotion â® [potassium 2,3-butanedione monoximate], soap and water, and 0.5% hypochlorite solution), supportive care, and administration of nerve agent antidotes (atropine, pralidoxime chloride, and diazepam). atropine is a competitive antagonist of acetylcholine muscarinic receptors and reverses peripheral muscarinic symptoms. it does not restore function at the neuromuscular junction nicotinic receptors. it does, however, treat early phases of convulsions. pralidoxime chloride separates the nerve agent from acetylcholinesterase and restores enzymatic function. it also binds free nerve agent. the major goal is to prevent "aging" of the enzyme (e.g., gd). diazepam provides treatment of nerve agent-induced seizures and prevents secondary neurologic injury. typically, associated seizures are refractory to other antiepileptic drugs. the antiseizure effect of diazepam is enhanced by atropine (banks 2014; cieslak and henretig 2016; messele et al. 2018) . potential medical countermeasures include trimedoxime (tmb4), hi-6 (an h-series oxime), obidoxime, "bioscavengers" (butyrylcholinesterase, carboxylesterase, organophosphorus acid anhydride hydrolase, and human serum paraoxonase), novel anticonvulsant drugs, n-methyl-d-aspartate (nmda) receptor antagonists (ketamine, dexanabinol), and common immunosuppressants such as cyclosporine a (jokanovic 2015; merrill et al. 2015 ; national institutes of health 2007; united states department of health and human services 2017). all patients should be observed closely for electroencephalographic changes and neuropsychiatric pathologies. polyneuropathy, reported after organophosphate insecticide poisoning, has not been reported in humans exposed to nerve agents and has been produced in animals only at unsurvivable doses. the intermediate syndrome has not been reported in humans after nerve agent exposure, nor has it been produced in animals. muscular necrosis has occurred in animals after high-dose nerve agent exposure but reversed within weeks; it has not been reported in humans (banks 2014). on march 4, 2018, sergei skripal, a former russian double agent, and his daughter, yulia skripal, were found unresponsive on a park bench in salisbury, uk. they were brought to a nearby hospital and treated for signs consistent with a cholinergic crisis due to a nerve agent exposure. analysis of the skripals found the presence of a secret nerve agent called novichok. further testing found high concentrations of the agent on the front-door handle of his home. one of the investigating police officers, detective sergeant nick bailey, unknowingly touched the door-handle and also became ill. all three survived due to rapid recognition of the nerve agent exposure by hospital personnel. four months later, two other people, dawn sturgess and charlie rowley, became ill with identical symptoms in the town of amesbury, 7 miles from salisbury. they were later confirmed to have high concentrations of novichok on their hands from a perfume bottle found in a recycling bin. both were immediately treated, but dawn sturgess later died. charlie rowley survived. it was believed the discarded perfume bottle contained novichok and was discarded by the assailants after the attempt on sergei skripal. on september 5, 2018, the uk government revealed that their investigation uncovered two suspects from closed circuit television (cctv) footage near the skripal's home. the suspects entered the uk on russian passports using the names alexander petrov and ruslan boshirov, stayed in a london hotel for 2 days, visited salisbury briefly, and then returned to moscow. minute traces of novichok were also found in the london hotel where they had stayed. the uk prime minister, teresa may, said that the suspects are thought to be officers from russia's military intelligence service the glavnoye razvedyvatel'noye upravleniye (gru), and that this showed that the poisoning was "not a rogue operation" and was "almost certainly" approved at a senior level of the russian state. the two suspects later appeared on russian tv denying the accusations and saying they were just "tourists" who had traveled all the way from moscow to salisbury just to see the "famous cathedral". however, cctv of the cathedral area found no evidence of the two men visiting the cathedral, although they were captured on cctv near the skripal's home. in a development in september 2018, one of the men was revealed as actually being a russian intelligence officer named colonel anatoliy chepiga and was a decorated veteran of russian campaigns in chechnya and ukraine. and later in october, the second man was named as dr. alexander mishkin, a naval medical doctor allegedly recruited by the gru (chai et al. 2018; may 2018) . novichok (ð�ð¾ð²ð¸ñ�ð¾ñ�: russian for "newcomer") is a highly potent nerve agent developed from the russian classified nerve agent program known as foliant. almost everything known about these agents is due to a russian defector, vil mirzayanov (2009) who was an analytical chemist at the russian state research institute of organic chemistry and technology (gosniiokht). he has described the details of the novichok program in his book "state secrets: an insider's chronicle of the russian chemical weapons program". the first three nerve agents of the novichok series developed in the program were substance-33, a-230, and a-232 (table 5 .18). they were synthesized as unitary agents, like vx, tabun, soman, and sarin. unitary means that the chemical structure was produced at its maximum potency. however, the novichok agents were developed as binary agents: maximum potency when two inert substances are combined together prior to deployment to create the active nerve agent (cieslak and henretig 2003) . very little is known about the chemistry of these weaponized organophosphate agents. however, they appear to be more potent than current nerve agents. for example, the ld 50 of novichok agents is reported 0.22 î¼g/kg similar to 2-(dimethylamino)ethyl n,n-dimethylphosphoramidofluoridate (vg), a novel fourth generation nerve agent. furthermore, novichok-5 is 8ã� more effective than vx and novichock-7 is 10ã� more effective than soman (cieslak and henretig 2003; hoenig 2007) . clinically, they behave like other organophosphates by binding to acetylcholinesterase preventing the breakdown of acetylcholine thereby leading to a cholinergic crisis. there appears to be a similar "aging" process as seen with other nerve agents. in addition, the novichok agents binding to peripheral sensory nerves distinguishes this class of organophosphates. prolonged or high-dose exposure results in debilitating peripheral neuropathy. exposure to these agents is fatal unless aggressively managed (cieslak and henretig 2003) . decontamination is essential to prevent ongoing exposure to the patient and medical personnel. clothing should be removed and quickly placed in a sealed bag (prevents ongoing exposure to the emission of vapors) followed by thorough washing with soap and water. application of dry bleach powder should be avoided as it may hydrolyze nerve agents into toxic metabolites that can produce ongoing cholinergic effects. supportive care is essential. antidote therapy should be given as usual for nerve agents, including atropine, diazepam, and pralidoxime chloride (united states department of health and human services, office of the assistant secretary for preparedness and response, national library of medicine 2019; united states department of health and human services, chemical hazards emergency medical management (chemm) 2019). of note, the toxicity of the novichok agents may not rely on anticholinesterase inhibition. some have suggested that reactive oximes like potassium 2,3-butanedione monoximate are preferred oximes for antidotal therapy (cieslak and henretig 2003) . cyanide is a naturally occurring chemical. it can be found in plants and seeds. it is also used in many industrial applications and is a common product of combustion of synthetic materials. typical cyanogens include hydrogen cyanide (ac) and cyanogen chloride (ck). low levels of cyanide are detoxified by a natural reaction in the human body using the rhodanese system. there is reversible metabolism with vitamin b12a to vitamin b12 (cyanocobalamin). an irreversible reaction occurs with sulfanes to produce thiocyanates and sulfates. the former is excreted via the urinary tract. when cyanide overwhelms this natural process, cyanide binds to (1990) a vx = venom x (cieslak and henretig 2003) cytochrome oxidase within the mitochondria and disrupts cellular respiration. cyanide has an affinity for fe+3 in the cytochrome a3 complex and oxidative phosphorylation is interrupted. cells can no longer use oxygen to produce atp and lactic acidosis ensues from resultant anaerobic metabolism. when inhaled, cyanide produces rapid onset of clinical signs. findings include transient tachypnea and kussmaul breathing (from hypoxia of carotid and aortic bodies), hypertension and tachycardia (from hypoxia of aortic body), and neurologic findings such as seizures, muscle rigidity (trismus), opisthotonus, and decerebrate posturing. other findings include cherry red flush, acute respiratory failure/ arrest, bradycardia, dissociative shock, and cardiac arrest. venous blood samples exhibit a bright red color. arterial blood gas may demonstrate a metabolic acidosis with an increased anion gap due to lactic acid (banks 2014; cieslak and henretig 2016; rotenberg 2003a) . pediatric manifestations (table 5 .19) may vary from the classic clinical responses due to their unique vulnerabilities (hilmas et al. 2008 ): â�¢ thinner integument leading to shorter time from exposure to symptom development. â�¢ higher vapor density (ck) and concentration accumulation in living zone of children, â�¢ higher minute ventilation and metabolism. â�¢ abdominal pain, nausea, restlessness, and giddiness are common early findings. â�¢ cyanosis mostly noted other than classic cherry red flushing of the skin. â�¢ resilient with recovery even when just using supportive measures alone. differential diagnoses include meningitis, encephalitis, gastroenteritis, ischemic stroke, methemoglobinemia, and poisonings (nerve agents, organophosphates, methanol, hydrogen sulfide, and carbon monoxide). diagnostic tests include arterial blood gas, lactic acid, and thiocyanate levels. treatment (tables 5.20 and 5.21) includes decontamination, supportive care, and administration of cyanide antidote kit (nitrites and thiosulfate). the nitrites facilitate the production of methemoglobinemia (fe+3) which attracts cyanide molecules forming cyanmethemoglobin. amyl nitrite pearls are crushed into gauze and placed over the mouth/nose or in a mask used for bag/mask ventilation. sodium nitrite is given parenterally and dosed according to the patient's estimated hemoglobin so as to prevent severe methemoglobinemia. since the formation of cyanmethemoglobin is a reversible reaction, and sodium thiosulfate is given to extract the cyanide. dosing is also dependent upon estimated hemoglobin. along with the naturally occurring rhodanese enzymatic system, the irreversible reaction forms thiocyanate. thiocyanate is water soluble and is excreted harmlessly via the kidneys (banks 2014; cieslak and henretig 2016). potential medical countermeasures (national institutes of health 2007; united states army medical research institute of infectious diseases (usamriid) 2014) include hydroxocobalamin, cobinamide (a cobalamin precursor), dicobalt edetate, cyanohydrin-forming compounds (alpha-ketoglutarate and pyruvate), s-substituted crystallized rhodanese, sulfur-containing drugs (n-acetylcysteine), and methemoglobin inducers (4-dimethylaminophenol and others). blistering agents, or vesicants, promote the production of blisters. typical examples include sulfur mustard (hd), nitrogen mustard (hn), and lewisite (l). these agents, especially sulfur mustard, are considered capable chemical weapons since illness may not occur until hours or days later. vesicants are alkylating agents that affect rapidly reproducing and poorly differentiated cells in the body. however, they can also produce cellular oxidative stress, deplete glutathione stores, and promote immature cognitive function unable to flee emergency immature coping mechanisms inability to discern threat, follow directions, and protect self high risk for developing ptsd bbb blood-brain barrier, bsa body surface area, cns central nervous system, ptsd post-traumatic stress disorder (hilmas et al. 2008) intense inflammatory responses. clinical findings are initially cutaneous (erythema, pruritus, yellow blisters, ulcers, and sloughing), respiratory (hoarseness, cough, voice changes, pneumonia, respiratory failure, acute lung injury, and acute respiratory distress syndrome), and ophthalmologic (pain, irritation, blepharospasm, photophobia, conjunctivitis, corneal ulceration, and globe perforation) in nature. after exposure through these primary portals of entry, other sites are affected, including the gastrointestinal tract (nausea, vomiting, and mucosal injury), the hematopoietic system (bone marrow suppression), the cardiovascular system (l), reproductive system (hd, hn) , and the central nervous system (lethargy, headache, malaise, and depression) (banks 2014; yu et al. 2003) . pediatric manifestations (table 5 .19) may vary from the classic clinical responses due to their unique vulnerabilities (hilmas et al. 2008 ): â�¢ thinner integument leading to shorter time from exposure to symptom development. â�¢ higher vapor density and concentration accumulation in the living zone. â�¢ higher minute ventilation and metabolism. â�¢ greater pulmonary injury. â�¢ ocular findings more frequent (less self-protection and more hand/eye contact). â�¢ gastrointestinal manifestations more prominent. â�¢ unable to escape and decontaminate. â�¢ unable to verbalize complaints (i.e., pain). treatment (tables 5.20 and 5.21) includes decontamination and supportive care. currently, there are no antidotes for mustard toxicity (cieslak and henretig 2016) . agents under investigation include antioxidants (vitamin e), anti-inflammatory drugs (corticosteroids), mustard scavengers (glutathione, n-acetylcysteine), and nitric oxide synthase inhibitors (l-nitroarginine methyl ester). other therapeutics under investigation include the use of british anti-lewisite (bal), reactive skin protectants, and ocular therapies (national institutes of health 2007; usamriid 2014). lung-damaging agents are toxic inhalants and potentially can affect the entire respiratory tract. typical examples include chlorine (cl 2 ), phosgene (carbonyl chloride), oxides of nitrogen, organofluoride polymers, hydrogen fluoride, and zinc oxide. since many of these chemicals are readily available and have multiple industrial applications, they are considered terrorist weapons of opportunity. toxicity is dependent upon agent particle size, solubility, and method of release. large particles produce injury in the nasopharynx (sneezing, pain, and erythema). midsize particles affect the central airways (painful swelling, cough, stridor, wheezing, and rhonchi). small particles cause injury at the level of the alveoli (dyspnea, chest tightness, and rales). highly soluble agents, such as chlorine, dissolve with mucosal moisture and immediately produce strong upper airway reactions. less soluble agents, such as phosgene, travel to the lower airway before dissolving and subsequently causing toxicity. it is important, however, to realize that very few lungdamaging agents affect only the upper or lower airway (e.g., cl 2 ). if the agent is aerosolized, solid or liquid droplets suspend in the air and distribute by size. if it is a gas or vapor release, there is uniform distribution throughout the lungs and toxicity will be based on solubility and reactivity of the agent (banks 2014; burklow et al. 2003; cieslak and henretig 2016) . pediatric manifestations (table 5 .19) may vary from the classic clinical responses due to their unique vulnerabilities (hilmas et al. 2008 ): â�¢ pediatric airway and respiratory tract issues (obligate nose breathers, relatively small mouth/large tongue, copious secretions, anterior/cephalad vocal cords, omega or horseshoe-shaped epiglottis, tendency of laryngospasm and bronchospasm, and anatomically small, "floppy" airways). â�¢ high vapor density and concentration accumulation in the living zone. â�¢ unable to verbalize or localize physical complaints. â�¢ rapid dehydration and shock secondary to pulmonary edema. â�¢ increased minute ventilation and metabolism. differential diagnoses include smoke inhalation injury, cardiogenic shock, heart failure, traumatic injury, asthma, bronchiolitis, and poisoning (cyanide). treatment (tables 5.20 and 5.21) includes decontamination and supportive care. currently, there are no antidotes for lung-damaging agent toxicity (cieslak and henretig 2016) . potential countermeasures include novel positive-pressure devices, drugs to prevent lung inflammation, and treatments for chemically induced pulmonary edema (beta agonists, dopamine, insulin, allopurinol, and ibuprofen). in addition, drugs are being investigated that act at complex molecular pathways of the lung the centers for disease control and prevention (cdc) has delineated bioterrorism agents and diseases into three categories based on priority. category a agents include organisms with the highest risk because the ease of dissemination or transmission from person-to-person, result in high mortality rates, have the potential for major public health impact, promote public panic and social disruption, and require special action of public health preparedness. these agents/diseases include smallpox (variola major), anthrax (bacillus anthracis), plague (yersinia pestis), viral hemorrhagic fevers (filoviruses [ebola, marburg] and arenaviruses [lassa, macupo]), botulinum toxin (from clostridium botulinum), and tularemia (francisella tularensis). category b agents, the second highest priority, include those that are moderately easy to disseminate, result in moderate morbidity and low mortality rates, and require specific enhancements of diagnostic capacity and enhanced disease surveillance. these agents/diseases include ricin toxin (ricinus communis), brucellosis (brucella species), epsilon toxin of clostridium perfringens, food safety threats (salmonella species, escherichia coli o157:h7, shigella), glanders (burkholderia mallei), meliodosis (burkholderia pseudomallei), psitticosis (chlamydia psittaci), typhus fever (rickettsia prowazekii), q fever (coxiella burnetii), staphylococcal enterotoxin b, trichothecenes mycotoxin, viral encephalitis (alphaviruses, such as eastern equine encephalitis, venezuelan equine encephalitis, and western equine encephalitis), and water safety threats (vibrio cholera, cryptosporidium parvum). category c agents have the next priority and include emerging pathogens that could be engineered for mass dissemination because of availability, ease of production and dissemination, and have the potential for high morbidity and mortality rates and major health impact. recognition of a biologic attack is essential. there are various epidemiologic clues to consider when determining whether the outbreak is natural or man-made (markenson et al. 2006; cieslak 2018; usamriid 2014) : â�¢ the appearance of a large outbreak of cases of a similar disease or syndrome, or especially in a discrete population. â�¢ many cases of unexplained diseases or deaths. â�¢ more severe disease than is usually expected for a specific pathogen or failure to respond to standard therapy. â�¢ unusual routes of exposure for a pathogen, such as the inhalational route for disease that normally occur through other exposures. â�¢ a disease case or cases that are unusual for a given geographic area or transmission season. â�¢ disease normally transmitted by a vector that is not present in the local area. â�¢ multiple simultaneous or serial epidemics of different diseases in the same population. â�¢ a single case of disease by an uncommon agent (smallpox, some viral hemorrhagic fevers, inhalational anthrax, pneumonic plague). â�¢ a disease that is unusual for an age group. â�¢ unusual strains or variants of organisms or antimicrobial resistance patterns different from those known to be circulating. â�¢ a similar or identical genetic type among agents isolated from distinct sources at different times and/or locations. â�¢ higher attack rates among those exposed in certain areas, such as inside a building if released indoors, or lower rates in those inside a sealed building if released outside. â�¢ outbreaks of the same disease occurring in noncontiguous areas. â�¢ zoonotic disease outbreaks. â�¢ intelligence of a potential attack, claims by a terrorist or aggressor of a release, and discovery of munitions, tampering, or other potential vehicle of spread (spray device, contaminated letter). one should know the cellular, physiological, and clinical manifestations of each biologic agent. furthermore, knowledge of distinct presentation patterns of children will be helpful to diagnosis. in any event, the ten steps in the management of biologic attack victims, pediatric, or otherwise, should be applied (cieslak and henretig 2003; cieslak 2018; usamriid smallpox is caused by the orthopoxvirus variola and was declared globally eradicated in 1980. the disease is highly communicable from person-to-person and remains a threat due to its potential for weaponization. the only stockpiles are at the cdc and at the russian state centre for research on virology and biotechnology. however, clandestine stockpiles in other parts of the world are unknown. since the cessation of smallpox vaccination, the general population has little or no immunity. the three clinical forms of smallpox include ordinary, flat, and hemorrhagic. another form, modified type, occurred in those previously vaccinated who were no longer protected. the asymptomatic incubation period is from 7 to 17 days (average 12 days) after exposure. a prodrome follows that lasts for 2-4 days and is marked by fever, malaise, and myalgia. lesions start on the buccal and pharyngeal mucosa. the rash then spreads in a centrifugal fashion, and the lesions are synchronous. initially, there are macules followed by papules, pustules, and scabs in 1-2 weeks. other clinical features include extensive fluid loss and hypovolemic shock, nausea, vomiting, diarrhea, bacterial superinfections, viral bronchitis and pneumonitis, corneal ulceration with or without keratitis, and encephalitis. death, if it occurs, is typically during the second week of clinical disease. variola minor caused a mortality of 1% in unvaccinated individuals. however, the variola major type caused death in 3% and 30% in those vaccinated and unvaccinated, respectively. flat (mostly children) and hemorrhagic (pregnant women and immunocompromised) types caused severe mortality in those populations infected. the differential diagnoses for smallpox include chickenpox (varicella), herpes, erythema multiforme with bullae, or allergic contact dermatitis. varicella typically has a longer incubation period (14-21 days) and minimal or no prodrome. furthermore, the rash distributes in a centripetal fashion and the progression is asynchronous (images 5.1 and 5.2). diagnosis of smallpox is mostly clinical (centers for disease control and prevention 2019a). if considered, contact public health immediately. laboratory confirmation (cdc or who) can be done by dna sequencing, polymerase chain reaction (pcr), restriction fragment-length polymorphism (rflp), real-time pcr, and microarrays. these are more sensitive and specific than the conventional virological and immunological approaches (goff et al. 2018) . generally, treatment is largely supportive (table 5 .23). fluid losses and hypovolemic shock must be addressed. also, due to electrolyte and protein loss, replacement therapy will be required. bacterial superinfections must be aggressively treated with appropriate antibiotics. biologic countermeasures and antivirals against smallpox are under investigation, including cidofovir, brincidovir (cmx-001), and tecovirimat (st-246). these agents have shown efficacy in orthopoxvirus animal models and have been used to treat disseminated vaccinia infection under emergency use. cidofovir has activity against poxviruses in animal studies (in vitro and in vivo) and some humans (eczema vaccinatum and molluscum contagiosum). brincidovir is an oral formulation of cidofovir with less nephrotoxicity and has recently been announced as an addition to the strategic national stockpile (sns) for patients with smallpox. tecovirimat is a potent and specific inhibitor of orthopoxvirus replication. a recent study found that treatment with tecovirimat resulted in 100% survival of cynomolgus macaques challenged with intravenous variola virus. the disease was milder in tecovirimat-treated survivors and viral shedding was reduced compared to placebo-treated survivors. prophylaxis comes in the form of the smallpox vaccine (vaccinia virus), acam2000 â® , which replaced wyeth dryvaxâ�¢ in 2007. safety profile of the two vaccines appears to be similar. side effects of vaccination range from low-grade fever and axillary lymphadenopathy to inadvertent inoculation of the virus to other body sites to generalized vaccinia and cardiac events (myopericarditis). rare, but typically fatal complications include progressive vaccinia, eczema vaccinatum, postvaccination encephalomyelitis, and fetal vaccinia. modified vaccinia ankara (mva) smallpox vaccine (bavarian nordic's imvamune â® ) is a live, highly attenuated, viral vaccine that is under development as a future nonreplicating smallpox vaccine (greenberg et al. 2016; kennedy and greenberg 2009 ). passive immunoprophylaxis exists in the form of vaccinia immune globulin (vig) and is used for primarily treating complications from smallpox vaccine. limited information suggests that vig may be of use in postexposure prophylaxis of smallpox if given the first week after exposure and with vaccination. monoclonal antibodies may represent another form of immunoprophylaxis. postexposure administration of human monoclonal antibodies has protected rabbits from a lethal dose of an orthopoxvirus. as mentioned, smallpox is highly communicable person-to-person (table 5 .25). contact precautions with full personal protective equipment (ppe) are required. airborne isolation with the use of an n-95 mask is needed for baseline protection. an n-95 mask or powered airpurifying respirator (papr) is recommended for protection during high risk procedures (beigel and sandrock 2009; goff et al. 2018; rotz et al. 2005; pittman et al. 2018 ; usamriid 2014). anthrax is caused by the aerobic, spore-forming, nonmotile, encapsulated gram-positive rod bacillus anthracis. it is a naturally occurring disease in herbivores. humans contract the illness by handling contaminated portions of infected animals, especially hides and wool. infection is introduced by scratches or abrasions on the skin. there is concern for potential aerosol dispersal leading to intentional infection through inhalation: it is fairly easy to obtain, capable of large quantity production, stable in aerosol form, and highly lethal. anthrax spores enter the body via skin, ingestion, or inhalation. the spores germinate inside macrophages and become vegetative bacteria. the vegetative form is released, replicates in the lymphatic system, and produces intense bacteremia. the production of virulence factors leads to overwhelming sepsis. the main virulence factors are encoded on two plasmids. one produces an antiphagocytic polypeptide capsule. the other contains genes for the synthesis of three proteins it secretes: protective antigen, edema factor, and lethal factor. the combination of protective antigen with lethal factor or edema factor forms binary cytotoxins, lethal toxin, and edema toxin. the anthrax capsule, lethal toxin, and edema toxin act in concert to drive the disease. three clinical syndromes occur with anthrax: cutaneous, gastrointestinal, and inhalational. cutaneous anthrax is the most common naturally occurring form. after an individual is exposed to infected material or the agent itself, there is a 1-12 day (average 7 days) incubation period. a painless or pruritic papule forms at the site of exposure. the papule enlarges and forms a central vesicle, which is followed by erosion into a coal-black but painless eschar. edema surrounds the area and regional lymphadenopathy may occur. gastrointestinal anthrax is rare. typically, it develops after ingestion of viable vegetative organisms found in undercooked meats of infected animals. the two forms of gastrointestinal anthrax, oropharyngeal and intestinal, have incubation periods of 1-6 days. the oropharyngeal form is marked by fever and severe pharyngitis followed by ulcers and pseudomembrane formation. other findings include dysphagia, regional lymphadenopathy, unilateral neck swelling, airway compromise, and sepsis. the intestinal form begins with fever, nausea, vomiting, and abdominal pain. bowel edema develops which leads to mesenteric lymphadenitis with necrosis, shock, and death. endemic inhalational anthrax (woolsorters' disease) is also extremely rare and is due to inhaling spores. therefore, any case of inhalational anthrax should be assumed to be due to intentional exposure until proven otherwise. the incubation period is 1-5 days but can be up to 43 days. there is a prodrome of 1-2 days consisting of fever, malaise, and cough. within 24 h, the disease rapidly progresses to respiratory failure, hemorrhagic mediastinitis (wide mediastinum), septic shock, multiorgan failure, and death. patients with inhalational anthrax may also have hemorrhagic meningitis. mortality is greater than 80% in 24-36 h despite aggressive treatment of inhalational anthrax. the differential diagnoses of ulceroglandular lesions include antiphospholipid antibody syndrome, brown recluse spider bite, coumadin/heparin necrosis, cutaneous leishmaniasis, cutaneous tuberculosis, ecthyma gangrenosum, glanders, leprosy, mucormycosis, orf, plague, rat bite fever, rickettsial pox, staphylococcal/ streptococcal ecthyma, tropical ulcer, tularemia, and typhus. the differential diagnoses of ulceroglandular syndromes include cat scratch fever, chancroid, glanders, herpes, lymphogranuloma venereum, melioidosis, plague, staphylococcal and streptococcal adenitis, tuberculosis, and tularemia. the differential diagnoses for inhalational anthrax include influenza and influenza-like illnesses from other causes. the differential diagnoses of mediastinal widening include normal variant, aneurysm, histoplasmosis, sarcoidosis, tuberculosis, and lymphoma. the diagnosis of anthrax is by culture and gram stain of the blood, sputum, pleural fluid, cerebrospinal fluid, or skin. specimens must be handled carefully, especially by lab personnel and those performing autopsies. elisa and pcr are available at some reference laboratories. the chest radiograph of inhalational anthrax shows the classic widening of the mediastinum. additional findings include hemorrhagic pleural effusions, air bronchograms, and/or consolidation (purcell et al. 2018 ). supportive treatment is indicated, including mechanical ventilation, pleural effusion drainage, fluid and electrolyte support, and vasopressor administration. for inhalational anthrax, antibiotic treatment is unlikely to be effective unless started before respiratory symptoms develop. treatment (table 5 .22) includes ciprofloxacin (or levofloxacin or doxycycline), clindamycin, and penicillin g. raxibacumab, a monoclonal antibody, was approved by the fda in 2012 for the treatment of inhalational anthrax in combination with recommended antibiotic regimens and prophylaxis for inhalational anthrax when other therapies are unavailable or inappropriate. it works by inhibiting anthrax antigen binding to cells and, therefore, prevents toxins from entering cells (kummerfeldt 2014) . the adult dose is 40 mg/kg given iv over 2 h and 15 min. the dose for children is weight based; â�¤15 kg: 80 mg/kg; >15-50 kg: 60 mg/kg; >50 kg: 40 mg/kg. premedication with diphenhydramine iv or po is recommended 1 h before the infusion. it can also be used as postexposure prophylaxis in high risk spore exposure cases (cieslak and henretig 2016; migone et al. 2009 ; the medical letter 2013). obiltoxaximab (anthim) is a recently approved monoclonal antibody treatment for inhalational anthrax in combination with recommended antibiotic regimens and prophylaxis for inhalational anthrax when other therapies are unavailable or inappropriate. adults and children >40 kg should receive a single obiltoxaximab dose of 16 mg/kg. the recommended dose is 24 mg/kg for children >15-40 kg and 32 mg/kg for those weighing â�¤15 kg. premedication with diphenhydramine is recommended to reduce risk of hypersensitivity reactions (the medical letter 2018). in patients with inhalational anthrax, intravenous anthrax immune globulin (anthrasil) should be considered in addition to appropriate antibiotic therapy (mytle et al. 2013 ; the medical letter 2016; usamriid 2014). postexposure prophylaxis includes ciprofloxacin (or levofloxacin or doxycycline) for 60 days plus administration of vaccine; since spores can persist in human in addition to appropriate antibiotic regimen, monoclonal antibody therapy (see text for dosing) and intravenous anthrax immune globulin should be administered for inhalational anthrax c levofloxacin or ofloxacin may be an acceptable alternative to ciprofloxacin d rifampin or clarithromycin may be acceptable alternatives to clindamycin as a drug that targets bacterial protein synthesis. if ciprofloxacin or another quinolone is employed, doxycycline may be used as a second agent because it also targets protein synthesis e ampicillin, imipenem, meropenem, or chloramphenicol may be acceptable alternatives to penicillin as drugs with good cns penetration f assuming the organism is sensitive, children may be switched to oral amoxicillin (40-80 mg/kg/d divided q8 h) to complete a 60-day course. the first 14 days of therapy of postexposure prophylaxis, however, should include ciprofloxacin or levofloxacin and/or doxycycline regardless of age. vaccination should also be provided; if not, antibiotic course will need to be longer g according to most experts, ciprofloxacin is the preferred agent for oral prophylaxis h ten days of therapy may be adequate for endemic cutaneous disease. a full 60-day course is recommended in the setting of terrorism, however, because of the possibility of concomitant inhalational exposure tissues for a long time, antibiotics must be given for a longer period if vaccine is not also given. the anthrax vaccine adsorbed (ava biothraxâ�¢) is derived from sterile culture fluid supernatant taken from an attenuated strain of bacillus anthracis and does not contain any live or dead organisms. the vaccine is given 0.5 ml intramuscularly at 0 and 4 weeks then at 6, 12, and 18 months followed by yearly boosters (pittman et al. 2018; usamriid 2014) . consult with cdc for current pediatric recommendations. anthrax is not contagious in the vegetative form during clinical illness (table 5 .25). contact with infected animals increases likelihood of spread. therefore, contact should be limited and the use of appropriate ppe in endemic areas is indicated (beigel and sandrock 2009; purcell et al. 2018; usamriid 2014) . plague is caused by yersinia pestis, a nonmotile, nonsporulating gram-negative bacterium. it is a zoonotic disease of rodents. it is typically found worldwide and is endemic in western and southwestern states. humans develop the disease after contact with infected rodents, or being bitten by their fleas. after a rodent population dies off, the fleas search for other sources of blood, namely humans. this is when large outbreaks of human plague occur. pneumonic plague is a very rare disease and when it is present in a patient, it may be highly suspicious for intentional dispersal of this deadly agent. three clinical syndromes occur with plague: bubonic plague (85%), septicemic plague (13%), and primary pneumonic plague (1-2%). bubonic plague occurs after an infected flea bites a human. after an incubation period of 2-8 days, there is onset of high fever, severe malaise, headache, myalgias, and nausea with vomiting. almost 50% have abdominal pain. around the same time, a characteristic bubo forms which is tender, erythematous, and edematous without fluctuation. buboes typically form in the femoral or inguinal lymph nodes, but other areas can be involved as well (axillary, intraabdominal). the spleen and liver can be tender and palpable. the disease disseminates without therapy. severe complications can ensue, including pneumonia, meningitis, sepsis, and multiorgan failure. pneumonia is particularly concerning since these patients are extremely contagious. mortality of untreated bubonic plague is 60%, but 5% with efficient and effective treatment. septicemic plague is characterized by acute fever followed by sepsis without bubo formation. the clinical syndrome is very similar to other forms of gram-negative sepsis: chills, malaise, tachycardia, tachypnea, hypotension, nausea, vomiting, and diarrhea. in addition to sepsis, disseminated intravascular coagulation can ensue leading to thrombosis, necrosis, gangrene, and the formation of black appendages. multiorgan failure can quickly follow. untreated septicemic plague is almost 100% fatal versus 30-50% in those treated. pneumonic plague is very rare and should be considered due to an intentional aerosol release until proven otherwise. the incubation period is relatively short at 1-3 days. sudden fever, cough, and respiratory failure quickly follow. this form produces a fulminant pneumonia with watery sputum that usually progresses to bloody. within a short period of time, septic shock and disseminated intravascular coagulation develop. ards and death may occur. mortality rate of pneumonic plague is very high but may respond to early treatment. plague meningitis is a rare complication of plague. it can occur in 6% of patients with septicemia and pneumonic forms and is more common in children. usually occurring a few weeks into the illness, it affects those receiving subtherapeutic doses of antibiotics or bacteriostatic antibiotics that do not cross the blood-brain barrier (tetracyclines). fever, meningismus, and other meningeal signs occur. plague meningitis is virtually indistinguishable from meningococcemia. the differential diagnoses of bubonic plague include tularemia, cat scratch fever, lymphogranuloma venereum, chancroid, scrub typhus, and other staphylococcal and streptococcal infections. the differential diagnoses of septicemic plague should include meningococcemia, other forms of gram-negative sepsis, and rickettsial diseases. the differential diagnosis of pneumonic plague is very broad. however, sudden appearance of previously healthy individuals with rapidly progressive gram-negative pneumonia with hemoptysis should strongly suggest pneumonic plague due to intentional release. diagnosis can be made clinically as previously described. demonstration of yersinia pestis in blood or sputum is paramount. methylene blue or wright's stain of exudates may reveal the classic safety-pin appearance of yersinia pestis. culture on sheep blood or macconkey agar demonstrates beaten-copper colonies (48 h) followed by fried-egg colonies (72 h). detection of yersinia pestis f1-antigen by specific immunoassay is available, but the result is available retrospectively. chest radiograph of patients will demonstrate patchy infiltrates (centers for disease control and prevention 2018a; worsham et al. 2018) . treatment includes mechanical ventilation strategies for ards, hemodynamic support (fluid and vasopressor administration), and antimicrobial agents (table 5 .23). gentamicin or streptomycin is the preferred antimicrobial treatment. alternatives include doxycycline or ciprofloxacin or levofloxacin or chloramphenicol. in cases of meningitis, chloramphenicol is recommended due to its ability to effectively cross the blood-brain barrier. streptomycin is in limited supply and is available for compassionate use. it should be avoided in pregnant women. postexposure prophylaxis includes doxycycline or ciprofloxacin. no licensed plague vaccine is currently in production. a previous licensed vaccine was used in the past. it only offered protection against bubonic plague but not aerosolized yersinia pestis. the plague bacterium secretes several virulence factors (fraction 1 (f1) and v (virulence) proteins) that as subunit proteins are immunogenic and possess protective properties. recently, an f1-v antigen (fusion protein) vaccine developed by usamriid provided 100% protection in monkeys against high-dose aerosol challenge. there is no passive immunoprophylaxis (i.e., immune globulin) available for pre-or postexposure of plague (usamriid 2014). use of standard precautions for patients with bubonic and septicemic plague is indicated. suspected pneumonic plague will require strict isolation with respiratory droplet precautions for at least 48 h after initiation of effective antimicrobial therapy, or until sputum cultures are negative in confirmed cases. an n-95 respirator should be used for baseline protection (table 5 .25). it is also recommended to use an n-95 respirator or papr for high risk procedures (beigel and sandrock 2009; ; centers for disease control and prevention 2017; centers for disease control and prevention 2018b; pittman et al. 2018; usamriid 2014) . in a mass casualty setting, parenteral therapy might not be possible. in such cases, oral therapy (with analogous agents) may need to be used b ofloxacin (and possibly other quinolones) may be acceptable alternatives to ciprofloxacin or levofloxacin; however, they are not approved for use in children c concentration should be maintained between 5 and 20 î¼g/ml. some experts have recommended that chloramphenicol be used to treat patients with plague meningitis, because chloramphenicol penetrates the blood-brain barrier. use in children younger than 2 may be associated with adverse reactions but might be warranted for serious infections d ribavirin is recommended for arenavirus or bunyavirus infections and may be indicated for a viral hemorrhagic fever of an unknown etiology although not fda approved for these indications. for intravenous therapy use a loading dose: 30 kg iv once (max dose, 2 g), then 16 mg/kg iv q6 h for 4 days (max dose, 1 g), and then 8 mg/kg iv q8 h for 6 days (max dose, 500 mg). in a mass casualty setting, it may be necessary to use oral therapy. for oral therapy, use a loading dose of 30 mg/kg po once, then 15 mg/kg/day po in 2 divided doses for 10 days viral hemorrhagic fever has a variety of causative agents. however, the syndromes they produce are characterized by fever and bleeding diathesis. the etiologies include rna viruses from four distinct families: arenaviridae, bunyaviridae, filoviridae, and flaviviridae. the filoviridae (includes ebola and marburg) and arenaviridae (includes lassa fever and new world viruses) are category a agents. based on multiple identified characteristics, there is strong concern for the weaponization potential of the viral hemorrhagic fevers. specifically, there has been demonstration of high contagiousness in aerosolized primate models. there are five identified ebola species, but only four are known to cause disease in humans. the natural reservoir host of ebola virus remains unknown. however, on the basis of evidence and the nature of similar viruses, researchers believe that the virus is animal-borne and that bats are the most likely reservoir. four of the five virus strains occur in an animal host native to africa. marburg virus has a single species. geographic distribution of ebola and marburg is africa (fitzgerald et al. 2016 ). both diseases are very similar clinically. incubation period is typically 5-10 days with a range of 2-16 days. symptoms may include fever, chills, headache, myalgia, nausea, and vomiting. there is rapid progression to prostration, stupor, and hypotension. the onset of a maculopapular rash on the arms and trunk is classic. disseminated intravascular coagulation and thrombocytopenia develops with conjunctival injection, petechiae, hemorrhage, and soft tissue bleeding. there is a possible central nervous system and hepatic involvement. bleeding, uncompensated shock, and multiorgan failure are seen. high viral load early in course is associated with poor prognosis. death usually occurs during the second week of illness. mortality rate of marburg is 25-85% and for ebola 50-90%. in a retrospective cohort study of children during the 2014/2105 ebola outbreak in liberia and sierra leone (all less than 18 years with a median age of 7 years with one-third less than 5 years of age), the most common features upon presentation were fever, weakness, anorexia, and diarrhea. about 20% were initially afebrile. bleeding was rare upon initial presentation. the overall case fatality rate was 57%. factors associated with death included children less than 5 years of age, bleeding at any time during hospitalization, and high viral load (smit et al. 2017) . in another retrospective cohort study of children at two ebola centers in sierra leone in 2014 (all less than 5 years of age), presenting symptoms included weakness, fever, anorexia, diarrhea, and cough. about 25% were afebrile on presentation. the case fatality rate was higher in children less than 2 years (76%) versus 2-5 years of age (46%) and 9 times more likely to die if child had a higher viral load. signs associated with death included fever, emesis, and diarrhea. interestingly, hiccups, bleeding, and confusion were only observed in children who died (shah et al. 2016) . lassa virus and new world viruses (junin, machupo, sabia, and guanarito) are transmitted from person-to-person. the vector in nature is the rodent. the incubation period is from 5 to 16 days. the geographical distribution is west africa and south america, respectively. the south american hemorrhagic fevers are quite similar but differ from lassa fever. the onset of the south american viruses is insidious and results in high fever and constitutional symptoms. petechiae or vesicular enanthem with conjunctival injection is common. these fevers are associated with neurologic disease (hyporeflexia, gait abnormalities, and cerebellar dysfunction). seizures portend a poor prognosis. mortality ranges from 15% to over 30%. on the contrary, lassa viruses are mild. less than 10% of infections result in severe disease. signs include chest pain, sore throat, and proteinuria. hemorrhagic disease is uncommon. other features include neurologic disease such as encephalitis, meningitis, cerebellar disease, and cranial nerve viii deafness (common feature). mortality can be as high as 25%. differential diagnoses include malaria, meningococcemia, hemolytic uremic syndrome, thrombotic thrombocytopenic purpura, and typhoid fever. diagnosis is through detection of the viral antigen testing by elisa or viral isolation by culture at the cdc. no specific therapy is present and generally involves supportive care, especially mechanical ventilation strategies for ards, hemodynamic support, and renal replacement therapy. for the arenaviridae and bunyaviridae groups, ribavirin may be indicated ( (pittman et al. 2018) . there is no current vaccine for ebola that is licensed by the fda. an experimental vaccine called rvsv-zebov was found to be highly protective against ebola virus in a trial conducted by the world health organization (who) and other international partners in guinea in 2015. fda licensure for the vaccine is expected in 2019. until then, 300,000 doses have been committed for an emergency use stockpile under the appropriate regulatory mechanism in the event and an outbreak occurs before fda approval is received (centers for disease control and prevention 2019b; henao-restrepo et al. 2015) . another ebola vaccine candidate, the recombinant adenovirus type-5 ebola vaccine, was evaluated in a phase 2 trial in sierra leone in 2015. an immune response was stimulated by this vaccine within 28 days of vaccination and strict contact precautions (hand hygiene, double gloves, gowns, shoe and leg coverings, and face shield or goggles) and droplet precautions (private room or cohorting, surgical mask within 3 ft) are mandatory for viral hemorrhagic fevers. airborne precautions (negative-pressure isolation room with 6-12 air exchanges per h) should also be instituted to the maximum extent possible and especially for procedures that induce aerosols (e.g., bronchoscopy). at a minimum, a fit-tested, hepa filter-equipped respirator (e.g., an n-95 mask) should be used, but a battery-powered papr or a positive pressure-supplied air respirator should be considered for personnel sharing an enclosed space with, or coming within 6 ft of, the patient. multiple patients should be cohorted in a separate ward or building with a dedicated airhandling system when feasible (table 5 .25). environmental decontamination is accomplished with hypochlorite or phenolic disinfectants (beigel and sandrock 2009; radoshitzky et al. 2018; usamriid 2014; won and carbone 2005) . francisella tularensis, a small aerobic, nonmotile gram-negative coccobacillus, causes tularemia (rabbit fever). clinical disease is caused by two isolates, biovars jellison type a and b. this organism can be stabilized for weaponization and delivered in a wet or dry form. the incubation period is usually 3-6 days (range 1-21 days). initial symptoms are nonspecific and mimic the flu-like symptoms or other upper respiratory tract infections. there is acute onset of fever with chills, myalgias, cough, fatigue, and sore throat. the two clinical forms of tularemia are typhoidal and ulceroglandular diseases. typhoidal tularemia (5-15%) occurs after inhalational exposure and sometimes intradermal or gastrointestinal exposures. there is abrupt onset of fever, headache, malaise, myalgias, and prostration. it presents without lymphadenopathy. nausea, vomiting, and abdominal pain are sometimes present. untreated, there is a 35% mortality rate in naturally acquired cases (vs. 1-3% in those treated). it is higher if pneumonia is present. this form would be most likely seen during an aerosol release of the agent. ulceroglandular tularemia (75-85%) occurs through skin or mucus membrane inoculation. there is abrupt onset of fever, chills, headache, cough, and myalgias along with a painful papule at the site of exposure. the papule becomes a painful ulcer with tender regional lymph nodes. skin ulcers have heaped up edges. in 5-10%, there is focal lymphadenopathy without an apparent ulcer. lymph nodes may become fluctuant and drain when receiving antibiotics. without treatment, they may persist for months or even years. in some cases (1-2%), the primary entry port is the eye leading to oculoglandular tularemia. patients have unilateral, painful, and purulent conjunctivitis with local lymphadenopathy. chemosis, periorbital edema, and small nodular granulomatous lesions or ulceration may be found. oropharyngeal tularemia with pharyngitis may occur in 25% of patients. findings include exudative pharyngitis/tonsillitis, ulceration, and painful cervical lymphadenopathy. the differential diagnosis is antibiotic unresponsive pharyngitis, infectious mononucleosis, and viral pharyngitis. pulmonary involvement (47-94%) is seen in naturally occurring disease. it ranges from mild to fulminant. various processes include pneumonia, bronchiolitis, cavitary lesions, bronchopleural fistulas, and chronic granulomatous diseases. left untreated, 60% will die. differential diagnoses include those for typhoidal (typhoid fever, rickettsia, and malaria) or pneumonic (plague, mycoplasma, influenza, q-fever, and staphylococcal enterotoxin b) tularemia. diagnosis should be considered when there is a cluster of nonspecific, febrile, systemically ill patients who rapidly progress to fulminant pneumonitis. tularemia can be diagnosed by recovering the organism from sputum (pcr or dfa) or serology at a state health laboratory. chest radiograph is nonspecific with possible hilar adenopathy. treatment is streptomycin or gentamicin (table 5 .23). alternatives include doxycycline, ciprofloxacin, or chloramphenicol. a live-attenuated vaccine (ndbr 101) exists and typically used for laboratory personnel working with francisella tularensis. there is no passive immunoprophylaxis. ciprofloxacin or doxycycline can be given as pre-and postexposure prophylaxis (beigel and sandrock 2009; hepburn et al. 2018; pittman et al. 2018; usamriid 2014) . botulinum neurotoxins (bont) are produced from the spore-forming, gram-positive, obligate anaerobe clostridium botulinum. it is the most potent toxin known to man. a lethal dose is 1 ng per kilogram. it is 100,000 times more toxic than sarin (gb). there are seven serotypes of botulinum toxin (a through g). a new serotype (h) has been tentatively identified in a case of infant botulism but has not been fully investigated. most common are serotypes a, b, and e. the toxin acts on the presynaptic nerve terminal of the neuromuscular junction and cholinergic autonomic synapses. this disrupts neurotransmission and leads to clinical findings. there are three forms of botulism: foodborne, wound, and intestinal (infant or adult intestinal). botulinum toxin can also be released as an act of bioterrorism via ingestion or aerosol forms. incubation can be from 12 h after exposure to several days later. clinical findings of botulism include cranial nerve palsies such as ptosis, diplopia, and dysphagia. this is followed by symmetric descending flaccid paralysis. however, the victim remains afebrile, alert, and oriented. death is typically due to respiratory failure. prolonged respiratory support is often required (1-3 months). differential diagnoses include guillain-barre syndrome, myasthenia gravis, tick paralysis, stroke, other intoxications (nerve gas, organophosphates), inflammatory myopathy, congenital and hereditary myopathies, and hypothyroidism. diagnosis is mostly clinical. laboratory confirmation can be obtained by bioassay of patient's serum. other assays include immunoassays for bacterial antigen, pcr for bacterial dna, and reverse transcriptase-pcr for mrna to detect active synthesis of toxin. cerebrospinal fluid demonstrates normal protein (unlike guillain-barre syndrome). emg reveals augmentation of muscle action potential with repetitive nerve stimulation at 20-30 hz. treatment (table 5 .23) is mainly supportive including intubation and ventilator support. tracheostomy may be required due to prolonged respiratory weakness and failure. antibiotics do not play a role in treatment. botulism antitoxin heptavalent [a, b, c, d, e, f, g]-equine (bat) was approved by the fda in 2013. bat was developed at usamriid as one of two equine-derived heptavalent bont antitoxins. bat is approved to treat individuals with symptoms of botulism following a known or suspected exposure. it has the potential to cause hypersensitivity reactions in those sensitive to equine proteins. the safety of bat in pregnant and lactating women is unknown. evidence regarding safety and efficacy in the pediatric population is limited. in 2003, the fda approved botulinum immune globulin intravenous (babybig), a human botulism immune globulin derived from pooled plasma of adults immunized with pentavalent botulinum toxoid. it is indicated for the treatment of infants with botulism from toxin serotypes a and b. immediately after clinical diagnosis of botulism, adults (including pregnant women) and children should receive a single intravenous infusion of antitoxin (bat or, for infants with botulism from serotypes a or b, babybig) to prevent further disease progression. the administration of antitoxin should not be delayed for laboratory testing to confirm the diagnosis. the pentavalent toxoid vaccine (previously for protection against a, b, c, d, and e; but not f or g) is no longer available as of 2011. no replacement vaccine is currently available. standard isolation precautions (table 5 .25) should be followed (beigel and sandrock 2009; dembek et al. 2018; pittman et al. 2018; timmons and carbone 2005; usamriid 2014 ). ricin is a potent cytotoxin derived from the castor bean plant ricinus communis. it is related in structure and function to shiga toxins and shiga-like toxin of shigella dysenteriae and escherichia coli, respectively. it consists of two glycoprotein subunits, a and b, connected by a disulfide bond. the b-chain allows the toxin to bind to cell receptors and gain entrance into the cell. once ricin enters the cell, the disulfide chemical linkage is broken. the free a chain then acts as an enzyme and inactivates ribosomes thereby disrupting normal cell function. cells are incapable of survival and soon die. ricin has a high terrorist potential due to it characteristics: readily available, ease of extraction, and notoriety (maman and yehezkelli 2005) . three modes of exposure exist: oral, inhalation, and injection. four to eight hours after inhalation exposure, the victim develops fever, chest tightness, cough, dyspnea, nausea, and arthralgias. airway necrosis and pulmonary capillary leak ensues within 18-24 h. this is followed quickly by severe respiratory distress, ards, and death due to hypoxemia within 36-72 h. injection may cause minimal pulmonary vascular leak. pain at the site and local lymphadenopathy may occur. however, it may be followed by nausea, vomiting, and gastrointestinal hemorrhage. ingestion leads to necrosis of the gastrointestinal mucosa, hemorrhage, and organ necrosis (spleen, liver, and kidney). diagnosis is suspected when multiple cases of acute lung injury occur in a geographic cluster. serum and respiratory secretions can be checked for antigen using elisa. pulmonary intoxication is managed by mechanical ventilation. gastrointestinal toxicity is managed by gastric lavage and use of cathartics. activated charcoal has little value due to the size of ricin molecules. supportive care is indicated for injection exposure. in general, treatment is largely supportive, especially for pulmonary edema that can result from the capillary leak. there is no vaccine available or prophylactic antitoxin for human use. however, there are two ricin vaccines in the development that focus on the ricin toxin a (rta) chain subunit. a mutant recombinant rta chain vaccine, rivax, has been shown to be safe and immunogenic in humans. the other vaccine is another recombinant rta chain vaccine, rvec . it has shown effectiveness in animal models by producing protective immunity against aerosol challenge with ricin in animal models. standard precautions are advised for health care workers (pittman et al. 2018; roxas-duncan et al. 2018; traub 2005 ; usamriid 2014). recent events which include the nuclear reactor meltdown at fukushima and international tension between nuclear powers, spark concern over potential devastation from nuclear catastrophes. there are numerous examples of radiation disasters in history. sixty-six thousand people were killed in hiroshima and thirty-nine thousand people were killed in nagasaki from nuclear bombs detonated over these cities in 1945 (avalon project-documents in law, history and diplomacy n.d.). many other people suffered from long-term consequences of radiation poisoning. in 1986, 21,000 square kilometers of land in russia, ukraine, and belarus were contaminated with radiation from a meltdown at a nuclear power plant in chernobyl, ukraine. one hundred and thirty-five thousand people were permanently evacuated from their homes (likhtarev et al. 2002) . long-term health consequences included many children who developed thyroid cancer several years later. many of these children died. a tsunami pummeled the east coast of japan in march of 2011. the power outage that ensued at the fukushima power plant led to a failure of the cooling system of the fuel rods, leading to a meltdown of four of the reactors at the plant. a massive quantity of radiation was released into the atmosphere, forcing people to evacuate their homes indefinitely. creative thinking and heroic actions by the tokyo fire department prevented entire populations of cities from being poisoned with radiation. terrorism experts are concerned that terrorist organizations will produce and detonate a radiological dispersion device (rdd), sometimes referred to as a dirty bomb. this is a conventional explosive, loaded with radioactive material which would be dispersed upon detonation. this would likely involve only one radioisotope. fewer people would be exposed and a smaller area would be contaminated than what would transpire with the detonation of a nuclear weapon. spreading fear and panic would be the primary purpose of such a device (mettler jr and voelz 2002) . radiation is the emission and propagation of energy through space or through a medium in the form of waves. radiation can be ionizing or nonionizing depending on the amount of energy released. most radiation that people encounter is low energy and, therefore, nonionizing with no biological effects. ionizing radiation emits enough energy to strip electrons from an atom, which provokes cellular changes and thereby, results in biological effects. radiation emitted from nuclear decay is always ionizing (radiation emergency assistance center/training site (react/s-cdc) 2006). atomic nuclei are held together by a very powerful binding energy despite positively charged protons repelling each other. this energy is released from unstable nuclei in the form of electromagnetic waves or particles. when ionizing radiation reaches biological tissue, chemical bonds are disrupted, free radicals are produced, and dna is broken. electromagnetic waves are of two types, x-rays and gamma rays. x-rays are relatively low energy and less penetrating. gamma rays have a shorter wavelength and contain relatively higher energy, making them more penetrating of biological tissue. ionizing radiation in the form of particles consists of alpha particles, beta particles, and neutrons. alpha particles are the largest of the forms of particulate radiation. they are composed of two neutrons and two protons. they do not easily penetrate solid surfaces, including clothes and skin. however, they can cause severe damage to an organism if internalized. in 2006, in the united kingdom, alexander litvienko, an ex kgb agent was poisoned with a radioactive element called polonium (mcphee and leikin 2009). a small amount of polonium was sprinkled into his food. polonium releases alpha particles when it decays. it was relatively safe for the assassin to carry this element with him because of the relatively poor ability of alpha particles to penetrate clothing and skin. once it is ingested, however, alpha particles have profound biological effects. mr. litvienko became very ill, and ultimately died. beta particles are high energy electrons discharged from the nucleus and are highly penetrating. neutrons emitted from a nucleus are also highly penetrating. in general, neutrons are only released by the detonation of a nuclear weapon. ionizing radiation of any form cannot be detected by our senses. it is not smelled, felt by touch, tasted, or seen. it is possible to be exposed to a lethal dose of radiation without realizing it. in goiania, brazil, in 1987, children found a canister of radioactive cesium ( 137 cs) that had been looted from a medical center and left in the street. the children liked the appearance of the substance but were not able to sense any abnormalities or danger with it. they began to rub it on their bodies because they liked the way it made them glow in the dark. the children all became ill. ultimately, 250 people were exposed to this radioisotope. it took 10 days before physicians recognized that the people had radiation poisoning. four people died of acute radiation syndrome. four factors determine the severity of exposure to ionizing radiation: time, distance, dose, and shielding. time is the time of exposure to the radiation source. distance is the distance from the radiation source. based on the inverse square law, exposure is reduced exponentially with increasing distance from the radiation source. dose is measured by the amount of energy released by the source and is numerically described by how many disintegrations per second occur, in curies (ci) or becquerels (bq). shielding is the efficacy of the barrier to the radioactive source. lead is well-known to be a very effective shield to x-rays. in a radiation exposure, injury to skin from trauma or burns may cause a greater degree of contamination because of loss of the shielding of the skin. there are four important principles for the nurse or hcp to understand with regard to exposure to ionizing radiation: external exposure, external contamination, internal contamination, and incorporation. external contamination occurs when radioactive material is carried on a person after exposure. this person can then contaminate others. removing contaminated clothing eliminates 90% of the toxin. others are then less vulnerable to exposure. internal contamination is when a radioactive substance enters the body through inhalation, ingestion, or translocation through open skin. incorporation is internalization of the toxin into body organs. incorporation is dependent on the chemical and not the radiological properties of the radioactive toxin. radioactive iodine, 131 i, is taken up by the thyroid gland because iodine enters the gland as part of normal physiology (advanced hazmat life support (ahls) 2003). ionizing radiation can damage chromosomes directly and indirectly, causing ravaging biological effects. indirect damage comes from the production of h + and oh â�� . free radical formation upsets biochemical processes and causes inflammation. these effects can take anywhere from seconds to hours to be expressed. clinical changes can take from hours to years to be realized (zajtchuk et al. 1989 ). immediately after a major radiation exposure, the clinical matters of most concern are those related to trauma from blast and thermal injuries. these injuries may be life-threatening and must be addressed first. after thermal and traumatic injuries are addressed, attention should be paid to the severity of radiation exposure. severe exposure can cause acute radiation syndrome. "the acute radiation syndrome is a broad term used to describe a range of signs and symptoms that reflect severe damage to specific organ systems and that can lead to death within hours or up to several months after exposure" (national council on radiation protection (ncrp) and measurements 2001; national council on radiation protection (ncrp) and measurements 2009). the mechanism of cell death from toxic radiation exposure is related to the inhibition of mitosis. organs with the most rapidly dividing cells are the most susceptible. the gastrointestinal and the hematopoietic are the organ systems most notably affected. the organs of pediatric patient have a higher mitotic index, in general, to those of adults and are more vulnerable to injury from radiation poisoning. the time of onset and the severity of acute radiation syndrome are controlled by the total radiation dose, the dose rate, percent of total body exposed, and associated thermal and traumatic injuries. there is a 50% death rate (ld 50 ) within 60 days for people exposed to a dose of radiation of 2.5-4.0 gy. the ld 50 is lower for the pediatric population. the acute radiation syndrome is composed of four phases: prodromal, latent, manifest illness, and death or recovery. inflammatory mediator release during the prodromal phase causes damage to cell membranes. this phase occurs during the first 48 h after exposure to radiation. nausea and vomiting and fever can occur during this time. if these symptoms occur during the first 2 h after exposure, there is a poor prognosis. the onset of the latent phase is usually in the first 4 days post exposure but can ensue anytime during the first 21 days thereafter. all cell lines of the hematopoietic system are affected. lymphocytes and platelets, the most rapidly dividing cells of the bone marrow, are most severely affected. the illness phase manifests after 30 days since radiation exposure. infection, impaired wound healing, anemia, and bleeding occur during this time of illness. the hematopoietic, gastrointestinal, central nervous, and integumentary are the organ systems affected. there is a marked reduction of cells from all cell lines of the bone marrow. there is a direct correlation with the drop in absolute lymphocyte count with the dose of radiation received. the absolute lymphocyte count is commonly used to estimate the dose of radiation received. the gastrointestinal (gi) epithelial lining, one of the most rapidly dividing cell lines of the body is the second most vulnerable to radiation poisoning. the radiation dose required to affect the gi system is 8 gy. vomiting, diarrhea, and a capillary leak syndrome for gi tract are common manifestations. hypovolemia and electrolyte instability ensue. translocation of bacteria into the bloodstream, combined with the diminished immunity caused by the decimation of the hematopoietic system, place victims at high risk for septic shock. another organ system affected by the acute radiation syndrome is the central nervous system. this requires a large dose of at least 30 gy. manifestations include cerebral edema, disorientation, hyperthermia, seizures, and coma. acute radiation syndrome that involves the central nervous system is always fatal. the integumentary system is frequently affected by the acute radiation syndrome, especially if the skin is in direct contact with a radioisotope. epilation, erythema, dry desquamation, wet desquamation, and necrosis occur respectively with increasing severity associated with increasing doses of radiation. radiation burns can be distinguished from thermal or chemical burns by their delayed onset. it can take days to weeks for radiation burns to affect victims. thermal and chemical burns cause signs and symptoms more acutely. hospitals that anticipate victims of radiation should prepare areas of triage with decontamination supplies and techniques ready to be deployed. an emergency department (ed) should be divided into "clean" and "dirty" areas. the dirty area is created for the purpose of decontamination to prevent the spread of radioisotopes. all health care personnel should wear ppe including surgical scrubs and gowns, face shields, shoe covers, caps, and two pairs of gloves. the inner pair of gloves is taped to the sleeves of the gown. each health care worker should be monitored for the exposure of the radiation and its dose with a dosimeter worn underneath the gown. the radiation safety officer of the hospital should take a leadership role in health care worker protection and decontamination procedures. consultation from the radiation emergency, assistance center (react/ts) is imperative. react/ts is a subsidiary of the u.s. department of energy. its contact information is as follows: phone number during business hours is 865-576-3131. the phone number is 865-576-1005 after business hours. the react/ts website is http://orise.orau. gov/reac/ts/. as victims arrive, triage protocols of mass casualty scenarios should be implemented. it should be noted that radiation exposure is not "immediately" lifethreatening. initial clinical management should focus on the abcde (airway, breathing, circulation, disability, and exposure) of basic trauma protocol. the "d" in the above acronym can also be a symbol for decontamination. after airway, breathing, and circulation are addressed, initial phase of decontamination entails careful removal of potentially contaminated clothing. caution should be exercised to remove the clothing gently, while rolling garments outward to prevent the release of dust of radioactive material that could contaminate people in the treatment area. further decontamination procedures take place after initial stabilization. skin decontamination procedures are identical to those of toxic chemical exposure with the following exceptions: â�¢ ppe are slightly different as described above. â�¢ gentle skin rubbing is done to prevent provocation of an inflammatory response and further absorption of the radioactive toxin. â�¢ only soap and water are used. rubbing alcohol and bleach should be avoided. it is advisable to shampoo the hair first, because it is usually the site of the highest level of contamination of the body, and runoff onto the body can then be cleansed during skin decontamination (radiation event medical management (remm) of the u.s. dept. of health and human services n.d.). it should be noted that health care workers are not at risk for contamination if they wear proper ppe during the resuscitation and decontamination process. the lack of knowledge of this point may lead to reluctance to treat patients and increase morbidity and mortality for victims. "no hcp has ever received a significant dose of radiation from handling, treating, and managing patients with radiation injuries and/or contamination."(react/s-cdc 2006). when initial resuscitation and decontamination have been completed, attention should be paid to ongoing support of ventilation, oxygenation, the management of fluid and electrolytes, and treatment of traumatic and burn injuries. infection control procedures are important due to the impending immunocompromised state of the victims. it is important to ascertain the details of the catastrophic event. data on the nature and size of the exposure and the types of radioactive agents involved are vital for ongoing management and decontamination. after the details of the nature of the exposure are uncovered, diagnostic tests should be done, including serial cbc and cytogenetic analysis of lymphocytes, otherwise known as cytogenic dosimetry (react/s-cdc 2006). measurements of change in lymphocyte counts and cytogenetic dosimetry are sensitive markers for the dose of radiation received by a victim. measurements of internal decontamination are done by the sampling and analysis of nasal and throat swabs, stool, and 24 h urine. wound samples and irrigation fluid should also be sampled. after initial stabilization, external decontamination, and diagnostic testing, internal decontamination is performed. external decontamination involves removal of clothes and cleaning the skin and hair. internal decontamination removes radioisotopes that are internalized via inhalation, ingestion, and entry into open wounds. because ionizing radiation is being released inside the body, internal decontamination must be performed promptly after initial resuscitation. since radioisotopes behave identically to their nonradioactive counterparts, antidotes are chosen based on the chemical, and not the radiological properties of the element. basic strategies of internal decontamination include chelation, competitive inhibition, enhanced gastrointestinal elimination, and enhanced renal elimination. specific agents are used for chelation of different radioisotopes. dtpa (diethyenetriaminepentaacetic acid) is administered for the elimination of heavy metals such as americium, californium, curium, and plutonium. dtpa comes in two forms, calcium dtpa (ca-dtpa) and zinc-dtpa (zn-dtpa). ca-dtpa is ten times more effective than zn-dtpa. for adults and adolescents, administration is as follows: â�¢ 1 g of ca-dtpa iv initially in the first 24 h, followed by 1 g zn-dtpa iv daily for maintenance. â�¢ for children less than 12 years of age administer: â�¢ fourteen mg/kg ca-dtpa iv initially, followed by fourteen mg/kg of zn-dtpa iv daily thereafter (national council on radiation protection (ncrp) and measurements 2009). â�¢ the initial dose of dtpa may be administered via inhalation to adolescents and adults if the contamination occurred via inhalation. this method of administration is not approved for pediatric use. chelation with dimercaprol (bal) is used to eliminate polonium. bal is a highly toxic drug and should be administered with caution. the dose is 2.5 mg per kg im four times a day for 2 days, then twice a day on the third day and once a day for 5-10 days, thereafter (national council on radiation protection (ncrp) and measurements 2009). alkalinization of the urine is renal protective during administration. a less toxic alternative to bal, dimercaptosuccinic acid (dmsa), otherwise known as chemet â® is also available. the dose of dmsa is ten mg per kg po every 8 h for 5 days. the same dose is given every 12 h for 14 days, thereafter (national council on radiation protection (ncrp) and measurements 2009). another mechanism for internal decontamination is competitive inhibition. the radioisotope, 131 i, is released during a meltdown of a reactor at a nuclear power plant. potassium iodide (ki) is widely recognized as a competitive inhibitor to its radioactive counterpart, 131 i, from being incorporated into the thyroid gland. ki blocks 90% of 131 i uptake into the thyroid gland if ki is given within the first hour of exposure. it will block 50% of incorporation if given within 5 h of exposure. its protective effect lasts for 24 h. with administration of this drug, thyroid function should be monitored closely. dosing guidelines (table 5 .26) are included in the table below (u.s. food and drug administration n.d.). gastrointestinal elimination is another mechanism of internal decontamination (table 5.27 ). an ion exchanger, prussian blue, (ferric ferrocyanide), binds elements that circulate through the enterohepatic cycle. since it is not absorbed through the gastrointestinal tract, prussian blue carries the toxins into the stool. it is highly effective in the elimination of 137 cs or thallium and was used during the 137 cs incident in goiania, brazil. the dosing of prussian blue is as follows: â�¢ infants: 0.2-0.3 mg per kg po three times a day (not fda approved). â�¢ children 2-12 years of age: 1 g po three times a day. â�¢ children â�¥12 years of age: 3 g po three times a day. â�¢ prussian blue is administered for at least 30 days, and can be adjusted based on the degree of poisoning (national council on radiation protection (ncrp) and measurements 2009). urinary elimination is another useful method of internal decontamination. tritium can be eliminated with excess fluid administration. uranium is eliminated by alkalinizing the urine to a ph of 8-9. sodium bicarbonate is given at a dose of 1 meq/kg iv every 4-6 h and is titrated to effect. if renal injury occurs, dialysis may be required. the basic approach to treating acute radiation syndrome is supportive therapy. gi losses from gastrointestinal difficulties are treated with iv fluids and electrolyte replacement. 5-ht3 antagonists can be used to suppress vomiting and benzodiazepines for anxiety. a patient suffering from acute radiation syndrome may be severely immunocompromised and requires a room with positive pressure isolation. colony stimulating agents for granulocytes and erythrocytes can be used for bone anemia and leukopenia. bone marrow transplant may be required for severe cases. a patient with skin contamination with radiation should be decontaminated with soap and water. a geiger counter can be helpful to identify areas of contamination. scrubbing is performed in a concentric matter, beginning at the outer layers of contamination and moving into the center since the area of greatest contamination is in the center. in this way, the area of contamination remains contained. attention should be paid to good nutrition and pain control. burn and plastic surgery service should also be consulted. more details on decontamination can be found in chap. 9. the psychological impact of a radiation catastrophe on the pediatric victims is likely to be devastating (american academy of pediatrics (aap) 2003). sleep disturbances, social withdrawal, altered play, chronic fear and anxiety, and developmental regression can occur. a correlation between the parent's psychological response and that of the child would occur as with other types of disaster. mental health professionals should be consulted in the event of this type of situation. please refer to chap. 12 for more information. a lot of concern has been expressed over the possibility of terrorist attacks involving explosive devices in recent years (depalma et al. 2005) . explosive devices are relatively simple to manufacture and easy to detonate. they can injure and kill many people and spread fear over large populations. victims of bomb blasts sustain more body regions injured, have more body injury severity scores, and require more surgeries than victims of nonexplosive trauma incidents. victims of explosives also have a higher mortality (kluger et al. 2004) . these observations are also true of pediatric victims (daniel-aharonson et al. 2003) . many factors influence the number of people injured and the severity of the injuries in an explosion. the magnitude of the explosion and its proximity to people and the number of people in the area affect the severity and number of injuries. other factors include the collapse of building or structure from the blast, promptness of the rescue operation, and the caliber and proximity of medical resources in the vicinity. victims who experience explosions in closed spaces are especially vulnerable to more severe injuries. twenty-nine case reports of injuries from terrorist bombings were reviewed (arnold et al. 2004a) . the investigators compared the injury severity of victims of explosions who sustained injuries from structural collapse, closed space explosions without structural collapse, and open space explosions. the mortality rate for these victims was 25%, 8%, and 4%, respectively. hospitalization rates were 25%, 36%, and 15%, respectively. ed visits were 48%, 36%, and 15%, respectively. victims of closed space explosions without structural collapse experienced greater hospitalizations rates than those involved in a structural collapse, because many of the victims involved in the structural collapse experienced immediate death. an explosion is defined as a rapid chemical conversion of a liquid or solid into a gas with energy release. substances that are chemically predisposed to explosion, called explosives, are characterized as low or high order, depending on the speed and magnitude of energy release. low-order explosives release energy at a relatively slow pace and explosions from these substances tend not to produce large air pressure changes or a "blast." the energy release is caused by combustion, producing heat. the involved material "goes up in flames." gunpowder, liquid fuel, and molotov cocktails are examples of low-order explosives (centers for disease control and prevention 2010). explosions from high-order materials cause a blast with a pressure wave in addition to causing the release of heat and light. the blast pressure wave causes compression of the surrounding medium which is physically transformed in all directions from the exact point of explosion. when an explosion occurs on land, air is the surrounding medium compressed. in bodies of water, the surrounding medium is water. the degree of medium compression and the distance that the energy wave travels is determined by the magnitude of the explosion. the power of the blast is measured in pounds per square inch (psi). the pressure blast wave has distinctive characteristics. the amplitude of the wave reaches its highest point immediately after the blast. the blast wave then rapidly decays as it travels through space. as the blast wave propagates, and compresses the surrounding medium, it leaves a vacuum because of displaced molecules in the surrounding medium and a negative phase of the wave ensues. in a land explosion, air molecules are displaced by the initial positive pressure, after which a negative pressure occurs in the vacated space. a wave that propagates through a confined space rebounds off of the wall and reverberates. it may interact with victims in the confined space many times, causing more severe injuries (stuhmiller et al. 1991) (fig. 5.10 ). four kinds of injury occur in high energy explosions. primary blast injuries occur directly from the pressure wave of the blast. secondary injuries occur from being struck by flying objects from the blast. these injuries can be blunt or penetrating. tertiary injuries occur when victims are displaced from a location and strike other objects or surfaces. all other injuries related to the blast are called quarternary. they include burns, inhalational injuries, toxic exposures, and traumatic injuries from structural collapse. primary injuries from blast waves affect bodily tissues with a tissue gas interface. when a pressure wave enters the body, tissue of gas filled organs compress slower than the air inside the tissue, causing stress in the tissue, possibly damaging it. this baseline positive phase originally described by friedlander, a blast wave consists of a short, high-amplitude overpressure peak followed by a longer depression phase. injury potential depends on the wave's amplitude as well as the slopes of its increase and decrease in pressure. x-axis refers to time and y-axis refers to pressure. (jacobson and severin 2012) also known as the "spalling effect." as the negative pressure phase of the blast wave propagates through, it causes more stress on the tissue and further damage. in addition to damaging tissues with an air tissue interface, pressure blasts can cause injury to the brain and can lead to limb detachments. despite the fact that primary blast injuries can be ravaging, they are less common than other types of injury from blasts. the tympanic membranes, lungs, and gastrointestinal tract are the most common organs sustaining injury from pressure waves. the tympanic membrane is the most vulnerable of these three organ systems (depalma et al. 2005; garth 1997) . five psi, which is considered a weak blast, will rupture 50% of tympanic membranes. to put this in perspective, c4, a commonly used explosive generates a pressure of four million psi. otoscopy can reveal ruptured tympanic membranes. neuropraxia, deafness, tinnitus, and vertigo are symptoms that can be experienced. severe blast injuries of the ear can result in damage to the organ of corti, resulting in permanent hearing loss. the second most common organ injured from a blast wave is the lung. fifteen psi are required to cause injury to this organ. lung injuries are more likely to occur from a blast within a closed space, or when victims sustain burns (burns commonly cause acute lung injury from release of inflammatory mediators). direct alveolar damage, blood vessel with bleeding, and inflammation are the three different manifestations of lung injury from blasts. alveolar damage can cause pneumothorax and pulmonary interstitial emphysema. when air dissects along the bronchovascular sheath, pneumomediastinum, pneumopericardium, and subcutaneous emphysema can occur. air that enters the pulmonary venous system can result in a systemic arterial air embolism, and possibly, a stroke. inflammation of the lungs from direct pressure damage to the tissue, cause acute lung injury and possibly, disseminated intravascular coagulation. clinical signs of lung injury include tachypnea, chest pain, hypoxia, rales, and dyspnea. if there is vascular disruption, hemoptysis can occur. air leaks from alveolar injury can result in diminished breath sounds, subcutaneous crepitance, increased resonance, and tracheal deviation. hemodynamic compromise will occur with tracheal deviation. alveolar damage, leading to air in the pulmonary venous system, can lead to a systemic arterial air embolism. air in the coronary arteries can lead to coronary ischemia with st and/or t waves changes on ecg. air embolism to cerebral arteries leads to cerebral vascular accidents (strokes) with focal neurological deficits. other manifestations of systemic air embolism include mottling of the skin, demarcated tongue blanching, and/or air in the retinal vessels (the most common sign of arterial air embolus). rapid death after initial survival is most often caused by arterial air embolus. initiation of positive pressure ventilation may trigger this event (ho and ling 1999) . a lung injury from a blast can also precipitate a vagal reflex resulting in bradycardia and hypotension. it is postulated that this occurs from the stimulation of c fibers in the lungs (guy et al. 1998 ). the gastrointestinal system is the third most common organ system affected by primary blast injury. physical stress and/or mesenteric infarct leads to weakening of the bowel wall with possible rupture. hemorrhage can also occur (paran et al. 1996; sharpnack et al. 1991) . the most common site of injury is the colon. injury to the bowel can be delayed and occur up to several days after the inciting incident. solid organs are spared because of their homogeneity and lack of air tissue interface. brain injury is becoming increasingly recognized as a result of primary blast. shearing injuries of the brain occur as a result of wave reverberation in the skull. hippocampal injury causing cognitive impairment has been shown in animal studies (cernak 2017; cernak et al. 2001; singer et al. 2005) . observations in humans have revealed electroencephalographic abnormalities and attention deficit disorder (born 2005) . human autopsies have revealed punctate hemorrhages and disintegration of nissl substance in victims who sustained blast injury without direct head trauma (guy et al. 1998) . research involving yucatan minipigs revealed that the brain sustains neuronal loss in the hippocampus after being subjected to primary blast injury. brain injury also occurred from the inflammation that ensued post blast (goodrich et al. 2016) . novel therapeutic approaches may be on the horizon for treatment of traumatic brain injury, including that caused by primary blast. intranasal insulin administered to rats subjected to traumatic brain injury resulted in enhanced neuronal glucose uptake and utilization, and subsequently improved motor function and memory. decreased neuroinflammation and preservation of the hippocampus were also noted (brabazon et al. 2017) . in a different investigation, a neuroprotective nucleotide, guanosine, was administered to rats subjected to traumatic brain injury. the treatment group of rats had better locomotor and cognitive outcomes than did the placebo group. programmed cell death and inflammation were also attenuated in the treatment group (gerbatin et al. 2017) . the leading cause of death from blast is from flying objects striking victims (secondary blast injury). eyes are particularly vulnerable. injuries resulting from displacement of the victims who strike objects are known as tertiary injuries. lighter weight children are particularly susceptible to this type of injury. burns, toxic exposures, and crush injuries constitute quaternary injuries. crush injuries commonly occur in explosions with structural collapse. the "crush syndrome" can occur when a trapped limb sustains prolonged compromise to the circulation, leading to rhabdomyolysis. tissue destruction and inflammatory response then occur. lifethreatening electrolyte abnormalities including hyperkalemia, renal failure, hyperuricemia, metabolic acidosis, acute respiratory distress syndrome, disseminated intravascular coagulation, and shock can result from crush syndrome (gonzalez 2005) . the crush syndrome is commonly seen in natural disasters that result in a lot of structural collapse. structural collapse and fires can cause the release of toxic materials such as carbon monoxide and cyanide. knowledge of the details of a blast can greatly enhance the ability of nurses and hcps to care for victims of a blast in a hospital setting. knowledge of whether a blast occurred in a closed or open space, whether structural collapse occurred, or if a victim was rescued from a collapsed area are details that can alert nurses and hcps as to what kind of injuries that they may anticipate. if toxic substances are released with a blast, nurses and hcps can prepare for decontamination techniques and antidote therapies. it would be advantageous for a hospital to be aware of the number of victims that are arriving for care. a mass casualty incident will stress the resources of the institution. hospital personnel should take stock of the resources that are available. the number of available ventilators and o-blood are examples of finite resources that should be considered. advanced trauma life support (atls) principles should be applied to all blast injury victims. abcd of initial resuscitation is applied. the "d" stands for disability as well as decontamination. decontamination techniques should be deployed if there is uncertainty about toxic exposure as described elsewhere in this chapter. on completion of abcd of initial resuscitation a secondary survey is performed, as described by atls protocol. attention should be paid to potential injuries that occur with blast injuries. ruptured tympanic membranes should alert the nurse or hcp of problems from primary blast injury. impaled objects should remain in place and removed in the operating room by surgical staff so that bleeding may be controlled. a thoracoscopy tube should be placed with an open three point seal over a wound on the side of the chest with an open pneumothorax. a hemothorax is also treated with a thoracoscopy tube. an autotransfusion setup can be applied to recirculate the blood from the pleural cavity of a hemothorax (wightman and gladish 2001 ) that would help preserve donor blood for other victims. for severe respiratory distress and/or impending respiratory failure, endotracheal intubation should be performed and positive pressure ventilation should be instituted. because lung tissue could be weakened from primary blast injury, caution should be exercised because of a high risk of pneumothorax, hemorrhage, or arterial air embolus. gentle application of positive pressure ventilation should be applied to avoid these complications. if only one lung is injured unilateral lung ventilation can be considered for larger children and adults. this technique is not suitable for babies and small children. supplemental oxygen with an fio 2 of 100% should be administered to patients suspected of having an arterial air embolus. hyperbaric oxygen therapy could even be considered to help accelerate the removal of air from the arteries. placement of the patient in the left lateral recumbent position may reduce the likelihood of the air lodging in the coronary arteries. victims of blast injuries should be treated identically to those of other types of trauma after initial resuscitation is completed. if primary blast injury occurred, frequent chest and abdominal x-rays should be performed in consideration of the possibility of lung or gastrointestinal injuries. limbs with open fractures should be immobilized and covered with sterile dressings. systemic, broad spectrum antibiotics should be administered to patients with open limb injuries. eyes that sustained chemical injury should be irrigated with water for an hour. all injured eyes should be covered. most ruptured tympanic membranes will heal spontaneously. victims with tympanic membrane injury should be advised to avoid swimming for some time. topical antibiotics are prescribed if dirt or debris is seen in the ear canal. oral prednisone is prescribed for hearing loss. victims with crush injuries should be treated with large volumes of iv fluids to treat inflammatory shock and possibly rhabdomyolysis. electrolytes should be monitored carefully as these patients are at risk for hyperkalemia, hyperphosphatemia, hyperuricemia, hypocalcemia, and acidosis. smoke inhalation, burns, and toxic exposures should be treated according to guidelines of burn, trauma, and toxicology protocols. mass casualty incidents (i.e. mass shootings, active shooter events, bombings, and other multifatality crimes) often attract extensive media coverage as well as the attention of policy makers. many agencies and organizations record and publish data on these incidents. the measurement and reporting does vary based on the absence of a common definition. however, it is clearly evident that mass casualty incidents (mcis) continue to increase in both number and scope (federal bureau of investigation 2017; office for victims of crime, office of justice programs, u.s. department of justice 2019). in the u.s., mass shootings are the most common and most closely tracked. the congressional research service (crs) defines mass shootings as events where more than four people are killed with a firearm "within one event, and in one or more locations in close proximity." congress uses the term mass killings and describes these events as "three or more killings in a single incident." the federal bureau of investigation (fbi) uses the term active shooter, which it defines as "an individual actively engaged in killing or attempting to kill people in a populated area." it is important to realize that nongovernmental ( ranking third of all locations for 2016 and 2017, seven of the 50 incidents occurred in educational environments resulting in five killed and 19 wounded. two incidents occurred in elementary schools, resulting in two killed (including a firstgrade student) and eight wounded (one teacher shot, three students shot, and four wounded from shrapnel). one incident occurred in a junior/senior high school, resulting in none killed and four wounded (two from shrapnel, all students). four incidents occurred at high schools (one outside a school during prom), resulting in three killed (all students) and seven wounded (all students). fortunately, no incident occurred at institutions of higher learning during 2016 or 2017 (advanced law enforcement rapid response training (alerrt) center, texas state university and federal bureau of investigation, u.s. department of justice 2018). notably, two of the 50 incidents occurred in houses of worship, resulting in 27 killed and 27 wounded. one of these incidents occurred at the first baptist church in sutherland springs, texas, and had the third highest number of casualties (26 killed and 20 wounded) in 2017. the dead included 10 women, 7 men, 8 children (7 girls and 1 boy), and an unborn child (goldman et al. 2017) . a summary report has also been developed for all 250 active shooter incidents from 2000 to 2017, including incidents per year (fig. 5.11 ), casualties per year (fig. 5.12) , and location ( fig. 5.13 ) categories (federal bureau of investigation 2017; federal bureau of investigation 2018). overall, there was an increase in number of active shooter incidents and casualties per year. location categories with number of incidents and statistics of their contribution were provided: areas of educational environments account for a large portion of locations for active shooter incidents, ranking only second to commercial areas. of the 37 incidents (14.8%) occurring at schools, one took place at a nursery (pre-k) school and one incident occurred during a school board meeting that was being hosted on school property but no students were involved (neither perpetrator or victim). the remainder (35 incidents) were perpetrated by or against students, faculty, and/or staff at k-12 schools (federal bureau of investigation 2018). finally, 15 active shooter incidents (6%) did occur at institutions of higher learning. as a reminder, no incident occurred at institutions of higher learning during 2016 or 2017. table 5 .31 provides a detailed summary of educational environment incidents from 2000 to 2017. since the beginning of 2018, other tragic active shooter attacks have occurred in the u.s. and greatly impacted children and adolescents. two of these such events have occurred in educational environments (united states secret service national threat assessment center 2018). on february 14, 2018, a gunman opened fire at marjory stoneman douglas high school. fourteen students and three staff members were killed while fourteen others were injured (follman et al. 2019) . twelve victims died inside the building, three died just outside the building on school premises, and two died in the hospital. the shooter was a former student of the school. another active shooter event occurred on may 18, 2018 at santa fe high school in santa fe, texas. the shooter killed ten individuals including eight students and two teachers while injuring 13 others. the shooter was an enrolled student at the school (follman et al. 2019) . based on the statistics of active shooter incidents, casualties, and locations, it is vital to prepare schools and plan for such events. national preparedness efforts, including planning, are now informed by the presidential policy directive (ppd) 8 that was signed by the president in march 2011 and describes the nation's approach to preparedness. this directive represents an evolution in our collective understanding of national preparedness based on the lessons learned from terrorist attacks, hurricanes, school incidents, and other experiences. ppd-8 defines preparedness around five mission areas and can be applied to school active shooter incidents. on march 21, 2005, at 2:49 p.m., jeffery james weise, 16, armed with a shotgun and two handguns, began shooting at red lake high school in red lake, minnesota. before the incident at the school, the shooter fatally shot his grandfather, who was a police officer, and another individual at their home. he then took his grandfather's police equipment, including guns and body armor, to the school. a total of nine people were killed, including an unarmed security guard, a teacher, and five students; six students were wounded. the shooter committed suicide during an exchange of gunfire with police campbell county comprehensive high school (education) on november 8, 2005, at 2:14 p.m., kenneth s. bartley, 14, armed with a handgun, began shooting in campbell county comprehensive high school in jacksboro, tennessee. before the shooting, he had been called to the office when administrators received a report that he had a gun. when confronted, he shot and killed an assistant principal and wounded the principal and another assistant principal. the shooter was restrained by students and administrators until police arrived and took him into custody pine middle school (education) on march 14, 2006, at 9:00 a.m., james scott newman, 14, armed with a handgun, began shooting outside the cafeteria at pine middle school in reno, nevada. no one was killed; two were wounded. the shooter was restrained by a teacher until police arrived and took him into custody essex elementary school and two residences (education) on august 24, 2006, at 1:55 p.m., christopher williams, 26, armed with a handgun, shot at various locations in essex, vermont. he began by fatally shooting his ex-girlfriend's mother at her home and then drove to essex elementary school, where his ex-girlfriend was a teacher. he did not find her, but as he searched, he killed one teacher and wounded another. he then fled to a friend's home, where he wounded one person. a total of two people were killed; two were wounded. the shooter also shot himself twice but survived and was apprehended when police arrived at the scene orange high school and residence (education) on august 30, 2006, at 1:00 p.m., alvaro castillo, 19, armed with two pipe bombs, two rifles, a shotgun, and a smoke grenade, began shooting a rifle from his vehicle at his former high school, orange high school in hillsborough, north carolina. he had fatally shot his father in his home that morning. one person was killed; two were wounded. the shooter was apprehended by police weston high school (education) on september 29, 2006, at 8:00 a.m., eric jordan hainstock, 15, armed with a handgun and a rifle, began shooting in weston high school in cazenovia, wisconsin. one person was killed; no one was wounded. the shooter was restrained by school employees until police arrived and took him into custody west nickel mines school (education) on october 2, 2006, at 10:30 a.m., charles carl roberts, iv, 32, armed with a rifle, a shotgun, and a handgun, began shooting at the west nickel mines school in bart township, pennsylvania. after the shooter entered the building, he ordered all males and adults out of the room. after a 20-min standoff, he began firing. the shooter committed suicide as the police began to breach the school through a window. five people were killed; five were wounded on april 2, 2012, at 10:30 a.m., su nam ko, aka one l. goh, 43, armed with a handgun, began shooting inside oikos university in oakland, california. he then killed a woman to steal her car. seven people were killed; three were wounded. the shooter was arrested by police later that day on august 27, 2012, at 10:45 a.m., robert wayne gladden jr., 15, armed with a shotgun, shot a classmate in the cafeteria of perry hall high school in baltimore, maryland. the shooter had an altercation with another student before the shooting began. he left the cafeteria and returned with a gun. no one was killed; one person was wounded. the shooter was restrained by a guidance counselor before being taken into custody by the school's resource officer sandy hook elementary school and residence (education) on december 14, 2012, at 9:30 a.m., adam lanza, 20, armed with two handguns and a rifle, shot through the secured front door to enter sandy hook elementary school in newtown, connecticut. he killed 20 students and six adults, and wounded two adults inside the school. prior to the shooting, the shooter killed his mother at their home. in total, 27 people were killed; two were wounded. the shooter committed suicide after police arrived taft union high school (education) on january 10, 2013, at 8:59 a.m., bryan oliver, 16, armed with a shotgun, allegedly began shooting in a science class at taft union high school in taft, california. no one was killed; two people were wounded. an administrator persuaded the shooter to put the gun down before police arrived and took him into custody new river community college, satellite campus (education) on april 12, 2013, at 1:55 p.m., neil allen macinnis, 22, armed with a shotgun, began shooting in the new river community college satellite campus in the new river valley mall in christiansburg, virginia. no one was killed; two were wounded. the shooter was apprehended by police after being detained by an off-duty mall security officer as he attempted to flee santa monica college and residence (education) on june 7, 2013, at 11:52 a.m., john zawahri, 23, armed with a handgun, fatally shot his father and brother in their home in santa monica, california. he then carjacked a vehicle and forced the driver to take him to the santa monica college campus. he allowed the driver to leave her vehicle unharmed but continued shooting until he was killed in an exchange of gunfire with police. five people were killed; four were wounded sparks middle school (education) on october 21, 2013, at 7:16 a.m., jose reyes, 12, armed with a handgun, began shooting outside sparks middle school in sparks, nevada. a teacher was killed when he confronted the shooter; two people were wounded. the shooter committed suicide before police arrived arapahoe high school (education) on december 13, 2013, at 12:30 p.m., karl halverson pierson, 18, armed with a shotgun, machete, and three molotov cocktails, began shooting in the hallways of arapahoe high school in centennial, colorado. as he moved through the school and into the library, he fired one additional round and lit a molotov cocktail, throwing it into a bookcase and causing minor damage. one person was killed; no one was wounded. the shooter committed suicide as a school resource officer approached him berrendo middle school (education) on january 14, 2014, at 7:30 a.m., mason andrew campbell, 12, armed with a shotgun, began shooting in berrendo middle school in roswell, new mexico. a teacher at the school confronted and ordered him to place his gun on the ground. the shooter complied. no one was killed; 3 were wounded: 2 students and an unarmed security guard. the shooter was taken into custody (continued) on june 5, 2014, at 3:25 p.m., aaron rey ybarra, 26, armed with a shotgun, allegedly began shooting in otto miller hall at seattle pacific university in seattle, washington. he was confronted and pepper sprayed by a student as he was reloading. one person was killed; 3 were wounded. students restrained the shooter until law enforcement arrived reynolds high school (education) on june 10, 2014, at 8:05 a.m., jared michael padgett, 15, armed with a handgun and a rifle, began shooting inside the boy's locker room at reynolds high school in portland, oregon. one student was killed; 1 teacher was wounded. the shooter committed suicide in a bathroom stall after law enforcement arrived marysville-pilchuck high school (education) on october 24, 2014, at 10:39 a.m., jaylen ray fryberg, 15, armed with a handgun, began shooting in the cafeteria of marysville-pilchuck high school in marysville, washington. four students were killed, including the shooter's cousin; 3 students were wounded, including one who injured himself while fleeing the scene. the shooter, when confronted by a teacher, committed suicide before law enforcement arrived florida state university (education) on november 20, 2014, at 12:00 a.m., myron may, 31, armed with a handgun, began shooting in strozier library at florida state university in tallahassee, florida. he was an alumnus of the university. no one was killed; 3 were wounded. the shooter was killed during an exchange of gunfire with campus law enforcement. umpqua community college (education) on october 1, 2015, at 10:38 a.m., christopher sean harper-mercer, 26, armed with several handguns and a rifle, began shooting classmates in a classroom on the campus of umpqua community college in roseburg, oregon. nine people were killed; 7 were wounded. the shooter committed suicide after being wounded during an exchange of gunfire with law enforcement. madison junior/ senior high school (education) on february 29, 2016, at 11:30 a.m., james austin hancock, 14, armed with a handgun, allegedly began shooting in the cafeteria of madison junior/senior high school in middletown, ohio. he shot two students before fleeing the building. no one was killed; four students were wounded (two from shrapnel). the shooter was apprehended near the school by law enforcement officers antigo high school (education) on april 23, 2016, at 11:02 p.m., jakob edward wagner, 18, armed with a rifle, began shooting outside a prom being held at his former school, antigo high school in antigo, wisconsin. two law enforcement officers, who were on the premises, heard the shots and responded immediately. no one was killed; two students were wounded. the shooter was wounded in an exchange of gunfire with law enforcement officers and later died at the hospital townville elementary school (education) on september 28, 2016, at 1:45 p.m., jesse dewitt osborne, 14, armed with a handgun, allegedly began shooting at the townville elementary school playground in townville, south carolina. prior to the shooting, the shooter, a former student, killed his father at their home. two people were killed, including one student; three were wounded, one teacher and two students. a volunteer firefighter, who possessed a valid firearms permit, restrained the shooter until law enforcement officers arrived and apprehended him on january 20, 2017, at 7:36 a.m., ely ray serna, 17, armed with a shotgun, allegedly began shooting inside west liberty salem high school, in west liberty, ohio, where he was a student. after assembling the weapon in a bathroom, the shooter shot a student who entered, then shot at a teacher who heard the commotion. the shooter shot classroom door windows before returning to the bathroom and surrendering to school administrators. no one was killed; two students were wounded. school staff members subdued the shooter until law enforcement arrived and took the shooter into custody freeman high school (education) on september 13, 2017, at 10:00 a.m., caleb sharpe, 15, armed with a rifle and a pistol, allegedly began shooting at freeman high school in rockford, washington, where he was a student. one student was killed; three students were wounded. a school employee confronted the shooter, ordered him to the ground, and held him there until law enforcement arrived and took him into custody rancho tehama elementary school and multiple locations in tehama county, california (education) on november 14, 2017, at 7:53 a.m., kevin janson neal, 44, armed with a rifle and two handguns, began shooting at his neighbors, the first in a series of shootings occurring in rancho tehama reserve, tehama county, california. after killing three neighbors, he stole a car and began firing randomly at vehicles and pedestrians as he drove around the community. after deliberately bumping into another car, the shooter fired into the car and wounded the driver and three passengers. the shooter then drove into the gate of a nearby elementary school. he was prevented from entering the school due to a lockdown, so he fired at the windows and doors of the building, wounding five children. upon fleeing the school, the shooter continued to shoot at people as he drove around rancho tehama reserve. law enforcement pursued the shooter; they rammed his vehicle, forced him off the road, and exchanged gunfire. the shooter's wife's body was later discovered at the shooter's home; the shooter apparently had shot and killed her the previous day. in total, five people were killed; 14 were wounded, eight from gunshot injuries (including one student) and six from shrapnel injuries (including four students). the shooter committed suicide after being shot and wounded by law enforcement during the pursuit aztec high school (education) on december 7, 2017, at approximately 8:00 a.m., william edward atchison, 21, armed with a handgun, began shooting inside aztec high school in aztec, new mexico. the shooter was a former student. two students were killed; no one was wounded. the shooter committed suicide at the scene, before police arrived a in a study of active shooter incidents in the united states between 2000 and 2013, the fbi identified 11 locations where the public was most at risk during an incident. these location categories include commercial areas (divided into business open to pedestrian traffic, businesses closed to pedestrian traffic, and malls), education environments (divided into schools [prekindergarten through 12th grade] and institutions of higher learning), open spaces, government properties (divided into military and other government properties), residences, houses of worship, and health care facilities. in 2018, the fbi added a new location category, other location, to capture incidents that occurred in venues not included in the 11 previously identified locations (federal bureau of investigation 2017). this table only includes educational environments. an entire list of all incidents from 2000 to 2017 at all locations can be found at https://www.fbi.gov/file-repository/activeshooter-incidents2000 .pdf/view (federal bureau of investigation 2018 prevention means the capabilities necessary to avoid, deter, or stop an imminent crime or threatened/actual mass casualty incident. prevention is the action schools take to prevent a threatened or actual incident from occurring. protection means the capabilities to secure schools against acts of violence and man-made or natural disasters. protection focuses on ongoing actions that protect students, teachers, staff, visitors, districts, networks, and property from a threat or hazard. mitigation means the capabilities necessary to eliminate or reduce the loss of life and property damage by lessening the impact of an event or emergency at the school. it also means reducing the likelihood that threats and hazards will happen. response means the school's or school district's capabilities necessary to stabilize an emergency once it has already happened or is certain to happen in an unpreventable way, establish a safe and secure environment, save lives and property, and facilitate the transition to recovery. recovery means the capabilities necessary to assist schools affected by an event or emergency in restoring the learning environment. it also means teaming with community partners to restore educational programming, the physical environment, business operations, and social, emotional, and behavioral health. the majority of prevention, protection, and mitigation activities generally occur before an incident, although these three mission areas do have ongoing activities that can occur throughout an active shooter incident. response activities occur during an incident, and recovery activities can begin during an incident and occur after an incident (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013; united states department of homeland security 2018b; united states department of homeland security 2018). in the k-12 school security guide, the u.s. department of homeland security (dhs) focuses on prevention and protection since the activities and measures associated with them occur prior to an incident (2018). effective preventative and protective actions decrease the probability that schools (or other facilities) will encounter incidents of gun violence or should an incident occur, it reduces the impact of that incident. the guide emphasizes that the level of security at a facility will be based on hazards relevant to the facility, people, or groups associated with it. it also warns that as new or different threats become apparent, the perception of the relative security changes and insecurity should drive change to reflect the level of confidence of the people of groups associated with the facility. the dhs utilizes a hometown security approach that emphasizes the process of connect, plan, train, and report (cptr) with the objective to realize effective, collaborative outcomes (united states department of homeland security 2018b). the initial phase is connect and occurs by a school or district reaching out and developing relationships in the community, including local law enforcement. having these relationships before an incident or event can help speed up the response when something happens. each school must begin with identification or development of a security team, group, or organization. this phase also emphasizes outreach, collaboration, and building of a coalition. there should be coalition members from within a school and may include district/school administrators, teachers, aides, facility operations personnel, human resources, administrative, counseling, and student groups. external groups directly related to the school might include boards of education, parent organizations, mental health groups/agencies, and teacher and bus driver unions. external groups indirectly related to the school include all responder organizations such as police and fire departments, sheriff's office, emergency medical services, emergency management, and the local dhs protective security advisor (psa). other tangential groups such as volunteer organizations, utility providers, and facilities in close geographic proximity should also be considered. core and advisory members of the coalition are established. a coalition champion is also identified and is the person who owns the majority of the responsibility for achieving a school's security goals. the champion organizes the coalition as it grows and matures (united states department of homeland security 2018b). the next phase is plan. this will bring the coalition together. the guide for developing high quality school emergency operations plans (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013) is an excellent resource for the coalition. a school security survey for gun violence can be completed and the coalition or user can quickly and effectively determine a facility's security proficiency (united states department of homeland security. 2018). specific portions of or topics within a school plan should be assigned to individuals, committees, or working groups most qualified to address them. the planning process must be sustainable. the amount of time spent in the planning phase should be commensurate with the amount of effort expended on the other phases (united states department of homeland security 2018b). the next phase of the process is to train on the plan developed by the coalition. determining who is responsible for what and how it should be done is the basic function of planning. in fact, telling various members of the team what is expected of them and when to do that activity is the function of training. it is vital to utilize the curricula development expertise possessed by the k-12 community. school administrators should take advantage of this skill set and find creative ways to address difficult topics, such as gun violence. it should be carried out in an effective and nontraumatic way. presenting the training in pieces or steps allows for a more comprehensive learning experience. it is important to validate training through exercises and drills, all of which should include the students. the training event should be followed by the completion and implementation of an after-action improvement plan with adjustment of the cptr as indicated (united states department of homeland security 2018b). the final phase in the process is report. the reporting phase is arguably the most important of all the phases. reporting principles underlie the other three phases and have profound prevention and protection impacts by driving forward information. dhs models the reporting phase using the "if you see something, say something â® " campaign (u.s. dhs, 2018) and the nationwide suspicious activity reporting (sar) initiative (nationwide suspicious activity reporting initiative (nsi) 2019). "if you see something, say something â® " focuses on empowering anyone who sees suspicious activity to do something about it by contacting local law enforcement, or if an emergency to call 9-1-1 (united states department of homeland security 2018a). this is a compelling capability when well organized and managed. a good plan for reporting, especially for a k-12 school, involves training staff and students on what is considered suspicious. there are many methods in which schools can employ to facilitate this, such as dedicated telephone numbers, websites for anonymous reporting, email or text messaging, and mobile phone applications. conducting simple drills for reporters and receivers keeps skills sharp and reinforces the importance of the effort with the goal to save lives. if the plan includes sharing all suspicious activity calls with the local fusion center then the probability of higher fidelity reporting increases (united states department of homeland security 2018b). when making changes to a school's plans, procedures, and protective measures, it is imperative the needs of individuals with special health care needs be addressed throughout the process. planning, training, and execution should always consider accessible alert systems for those who are deaf or hard of hearing; students, faculty, and staff who have visual impairments or are blind; individuals with limited mobility; alternative notification measures; people with temporary disabilities; visitors; people with limited english proficiency; sign cards with text-and picture-based emergency messages/symbols; and involving people with disabilities in all planning (united states department of homeland security, interagency security committee 2015). it is important to understand that no "profile" exists for an active shooter (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013). however, research indicates there may be signs or indicators. o'toole (2000) presents an in depth, systematic procedure for school shooter threat assessment and intervention. the model was designed to be used by educators, mental health professionals, and law enforcement agencies. its fundamental building blocks are the threat assessment standards, which provide a framework for evaluating a spoken, written, and symbolic threat, and the fourpronged assessment approach which provides a logical, methodical process to examine the threatener and assess the risk that the threat will be carried out. schools should learn the signs of a potentially volatile situation that may develop into an active shooter situation and proactively seek ways to prevent an incident with internal resources, or additional external assistance (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013). potential warning signs of a school shooter may include increasingly erratic, unsafe, or aggressive behaviors; hostile feelings of injustice or perceived wrongdoing; drug and alcohol abuse; marginalization or distancing from friends and colleagues; changes in performance at work or school; sudden and dramatic changes in home life or in personality; pending civil or criminal litigation; and observable grievances with threats and plans of retribution (united states department of homeland security 2018b). at a minimum, schools should establish and enforce policies that prohibit, limit, or determine unacceptable behaviors and consequences of weapons possession/use, drug possession/use, alcohol/tobacco possession/use, bullying/harassment, hazing, cyber-bullying/harassment/stalking, sexual assault/misconduct/harassment, bias crimes, social media abuse, and any criminal acts (united states department of homeland security 2018b). in addition to policies and positive school climates, school districts and administrators should establish dedicated teams to evaluate threats, such as a threat assessment team (tat). the team should include mental health professionals (e.g., forensic psychologist, clinical psychologist, and school psychologist) to contribute to the threat assessment process (united states department of homeland security 2018b). it is the responsibility of the tat to investigate and analyze communications and behaviors to make a determination on whether or not an individual poses a threat to him/herself or others (united states department of education, office of elementary and secondary education, office of safe and healthy students 2013). as well as tats, some schools have even opted to establish social media monitoring teams which look for keywords that may indicate bullying or other concerning statements. if a school opts to create such a team, it should work very closely with the tat to ensure that applicable privacy, civil rights and civil liberties, other federal, state and local laws, and information sharing protocols are followed. please refer to chap. 14 for further information. after an active shooter incident, field triage (e.g., jumpstart) must commence and the patient must be evaluated by an experienced emergency medicine or trauma surgeon, preferably by a pediatric specialist in those disciplines. if an active shooter incident is coupled with detonation of an explosive device, the child must be screened and decontaminated for radiation exposure ("dirty bomb"). triage tags are extremely helpful when multiple victims present in a short period of time. medical response to an active shooter event will focus on control of external hemorrhage along with circulatory stabilization. operative emergencies will be common and receive the highest priority. severe extremity injuries may be controlled with tourniquet application or other forms of hemorrhage control. re-evaluation is paramount to prevent ischemia to distal regions. however, thoracic or abdominal (truncal) injuries will need immediate surgical exploration and intervention. penetrating trauma will cause more vascular injuries than blunt trauma, and vascular surgical trays may be in short supply at a hospital. major procedure or surgical trays may become short in supply based on the increased operative demand. resuscitative blood transfusion therapy may utilize a massive blood transfusion protocol. since whole blood may be short in supply, some will simply use the 1:1:1 rule (administer one unit of packed cells: one unit of fresh frozen plasma: one unit of platelets). a unit for children may be substituted as an aliquot based on size of the patient (e.g., administer 10 ml/kg of packed cells: 10 ml/kg of fresh frozen plasma: 10 ml/kg of platelets). calcium must also be replaced when there is a large volume transfusion. due to extensive blood product utilization, there may be a heavy impact on institutional or regional blood supplies. plans should be in place to address these problems, including the implementation of allocation of scarce resources. mental health support and staff debriefs are essential and should be included after an active shooter event (hick et al. 2016 ). in conclusion, all forms of disasters, whether man-made or natural, impact infants, children, and adolescents throughout the world. effective and efficient interventions remain the cornerstone of sustaining a child's well-being while reducing untoward complications due to all forms of disasters. having a deep understanding of pediatric physiology and pathophysiology is crucial to all levels of disaster diagnostics and therapeutics. all nurses and hcps have an obligation to understand these principles and deliver excellent, compassionate care to the pediatric disaster victim. advanced law enforcement rapid response training evidence-based support for the all-hazards approach to emergency preparedness ahls advanced hazmat life support provider manual active shooter incidents in the united states in 2016 and 2017 radiation disasters and children-committee on environmental health apls: the pediatric emergency medicine resource textbook of neonatal resuscitation american college of emergency physicians, pediatric committee, and emergency nurses association pediatric committee american college of emergency physicians, national association of ems physicians, pediatric equipment guidelines committee emergency medical services for children (emsc) partnership for children stakeholder group handbook of emergency cardiovascular care for healthcare providers pediatric advanced life support provider manual mass casualty terrorist bombings: a comparison of outcomes by bombing type committee on pediatric emergency medicine and task force on terrorism. the pediatrician and disaster preparedness mass casualty terrorist bombings: a comparison of outcomes by bombing type assistant secretary for preparedness and response, united states department of health and human services avalon project-the atomic bombings of hiroshima and nagasaki fort sam houston, tx: borden institute, us army research institute of chemical defense (usamricd), office of the surgeon general intentional and natural outbreaks of infectious disease preparedness for acts of nuclear, biological, and chemical terrorism a study of active shooter incidents blast trauma: the fourth weapon of mass destruction intranasal insulin treatment of an experimental model of moderate traumatic brain injury industrial chemicals: terrorist weapons of opportunity medical response to terrorism: preparedness and clinical practice bioterrorism agents/diseases centers for disease control and prevention. emergency preparedness and response: information on specific types of emergencies blast injuries: a primer for clinicians centers for disease control and prevention. plague: resources for clinicians economic losses, poverty and disasters natural disasters 2017: lower mortality, higher cost general classification of disasters centre for research on the epidemiology of disasters (cred) and united nations office for disaster risk reduction (unisdr) understanding blast-induced neurotrauma: how far have we come? ultrastructural and functional characteristics of blast injury-induced neurotrauma novichok agents: a historical, current, and toxicological perspective medical management of potential biological casualties: a stepwise approach biologic and chemical terrorism an idea whose time has come? in: terrorism in the twenty-first century epidemiology of terror-related versus nonterror-related traumatic injury in children botulinum toxin departments of the army, the navy, and the air force, and commandant, marine corps. fm 8-825 treatment of chemical agent casualties and conventional military chemical injuries task force for pediatric emergency mass critical care. supplies and equipment for pediatric emergency mass critical care american academy of pediatrics disaster preparedness advisory council. medical countermeasures for children in public health emergencies, disasters, or terrorism quick look: 250 active shooter incidents in the united states from federal bureau of investigation. active shooter incidents in the united states from ebola virus disease in children mass shootings scientific foundations of trauma guanosine protects against traumatic brain injury-induced functional impairments and neuronal loss by modulating excitotoxicity, mitochondrial dysfunction, and inflammation smallpox and related othopoxviruses texas church shooting video shows gunman's methodical attack, official says. the new york times crush syndrome neuronal and glial changes in the brain resulting from explosive blast in an experimental model a randomized, double-blind, placebo-controlled phase ii trial investigating the safety and immunogenicity of modified vaccinia ankara smallpox vaccine (mva-bnâ®) in 58-80-year-old subjects physiologic responses to primary blast efficacy and effectiveness of an rvsv-vectored vaccine expressing ebola surface glycoprotein: interim results from the guinea ring vaccination cluster-randomised trial committee on drugs. preparing for pediatric emergencies: drugs to consider health and medical response to active shooter and bombing events: a discussion paper medical management of chemical toxicity in pediatrics systemic air embolism after lung trauma nerve agents. in: compendium of chemical warfare agents caring for non-injured and non-ill children in a disaster: a guide for non-medical professionals and volunteers pediatric disaster preparedness guidelines for hospitals current concepts of pediatric critical care. mount prospect, il: society of critical care medicine terrorist and criminal attacks targeting children pyridinium oximes in the treatment of poisoning with organophosphorus compounds imvamuneâ®: modified vaccinia ankara strain as an attenuated smallpox vaccine the special injury pattern in terrorist bombings raxibacumab: potential role in the treatment of inhalational anthrax chernobyl accident: retrospective and prospective estimates of external dose of the population of ukraine disasters and their effects on the population: key concepts. in: berman s, editor. pediatric education in disasters manual bioterrorism and infectious agents: a new dilemma for the 21st century national center for disaster preparedness. mailman school of public health, columbia university developing next-generation countermeasures for homeland security threat prevention novichok: the notorious nerve agent. molecule of the month death by polonium-210: lessons learned from the murder of former soviet spy alexander litvinenko physiologically based pharmacokinetic/pharmacodynamic modeling of countermeasures to nerve agents degradation of pesticides with rsdlâ® (reactive skin decontamination lotion kit: lc-ms investigation) major radiation exposure-what to expect and how to respond raxibacumab for the treatment of inhalation anthrax state secrets: an insider's chronicle of the russian chemical weapons program applying risk-based decision making methods and tools to u.s navy antiterrorism capabilities dosing cards for treatment of children exposed to weapons of mass destruction author: bethesda, md; 2001. national council on radiation protection (ncrp) & measurements. management of persons contaminated with radionuclides: handbook. (report no 161) national strategy for information sharing office for victims of crime the school shooter: a threat assessment perspective. quantico, va: critical incident response group (cirg), national center for the analysis of violent crime perforation of the terminal ileum induced by blast injury: delayed diagnosis or delayed perforation? medical countermeasures apls: the pediatric emergency medicine resource medical aspects of biological warfare medical aspects of biological warfare emergency management of radiation accident victims. reac/ts-cdc course. lecture conducted from oak ridge institute for science and education of health and human services. decontamination procedures-radiation event medical management & about decontamination of children video cyanide as a weapon of terror diagnosis and management of nerve agent exposure nerve agent attacks on children: diagnosis and management medical response to terrorism: preparedness and clinical practice active shooter incidents in the united states pediatric acute care: a guide for interprofessional practice inpatient signs and symptoms and factors associated with death in children aged 5 years and younger admitted to two ebola management centres in sierre leone, 2014: a retrospective cohort study the pathology of primary blast injury conventional terrorism and critical care apls: the pediatric emergency medicine resource characteristics and outcomes of pediatric patients with ebola virus disease admitted to treatment units in liberia and sierra leone: a retrospective cohort study the physics and mechanisms of primary blast injury guidance potassium iodide as a thyroid blocking agent in radiation emergencies. n.d.. w.fda.gov/downloads/drugs/guidancecomplianceregulatory information/guidances/ucm080542 medical response to terrorism: preparedness and clinical practice medical response to terrorism: preparedness and clinical practice usamriid's medical management of biological casualties handbook guide for developing high-quality school emergency operations plans report of the children's hhs interagency leadership on disasters (child) working group: update on department activities assistant secretary for preparedness and response biomedical advanced research and development authority united states department of health and human services, office of the assistant secretary for preparedness and response, national library of medicine. chemical hazards emergency medical management (chemm) united states department of health and human services, chemical hazards emergency medical management (chemm) public health emergency countermeasures enterprise (phemce) strategy and implementation plan united states department of homeland security planning and response to an active shooter: an interagency security committee policy and best practices guide united states department of homeland security. k-12 school security: a guide for preventing and protecting against gun violence united states department of homeland security. k-12 school security survey understanding your risks: identifying hazards and estimating losses fema fact sheet: national planning scenarios united states department of homeland security. national planning scenarios. version 21.3 final draft fema's progress in all-hazards mitigation enhancing school safety using a threat assessment model: an operational guide for preventing targeted school violence living with hazards, dealing with disasters: an introduction to emergency management terrorism and the all-hazards model explosions and blast injuries medical response to terrorism: preparedness and clinical practice quantitative terrorism risk assessment medical aspects of biological warfare vesicant agents and children medical consequences of nuclear warfare (textbooks of military medicine) safety and immunogenicity of a recombinant adenovirus type-5 vector-based ebola vaccine in healthy adults in sierra leone: a single-centre, randomised, double-blind, placebo-controlled, phase 2 trial key: cord-320172-qw47pf9r authors: greaves, peter title: vii digestive system 1 date: 2000-12-31 journal: histopathology of preclinical toxicity studies doi: 10.1016/b978-044450514-9/50007-3 sha: doc_id: 320172 cord_uid: qw47pf9r publisher summary this chapter deals with the digestive system. the major and minor salivary glands and their secretions also represent and integral part of the protective mechanism of the oral cavity, and derangement of saliva production may lead to loss of integrity of the oral mucosa. drug-induced abnormalities of taste sensation are also well-described phenomena occurring in man although human studies are necessary for the detection of these effects. inflammation of the oral cavity may involve the buccal mucosa, the gingiva (gingivitis), the tongue (glossitis), and the peridontal tissues (peridontitis). therapeutic agents can induce inflammatory lesions in the tongue. moreover, a protective layer of mucus, a visco-elastic material containing high molecular weight glycoproteins produced by the major and minor salivary glands, covers the stratified squamous mucosa of the oral cavity. salivary secretions also possess digestive enzyme activity although in herbivores and carnivores, it is usually low in contrast to high digestive enzyme activity in omnivorous species. the oral mucosa can be damaged by excessive local trauma from foreign materials, hard fragments in food and damaged teeth may produce ulceration of the mucosa with subsequent infection. however, the oral mucosa may show manifestations of local or systemic disease or derangement produced by therapeutic agents. excessive contact by therapeutic agents such as aspirin, potassium supplements and corticosteroids have been reported to produce local ulceration in the mouth (zentler-monro and northfield, 1979) . the increased use of mouthwashes over the last 20 years has resulted in a number of reported adverse effects to the buccal mucosa in people (gargari and kabani, 1995) . systemic disorders produced by anticoagulants or chemotherapeutic drugs may also be evident by bleeding or ulceration in the oral cavity (goepp, 1982) . buccal ulceration is also described as part of a generalised hypersensitivity reaction to drugs (zentler-monro and northfield, 1979) . the major and minor salivary glands and their secretions also represent and integral part of the protective mechanism of the oral cavity and derangement of saliva production may lead to loss of integrity of the oral mucosa. drugs that effect motor co-ordination can give rise to drooling and disruption of cricopharyngeal co-ordination (wyllie et al., 1986) . drug-induced abnormalities of taste sensation are also well-described phenomena occurring in man although human studies are necessary for the detection of these effects. indeed, many alterations in the oral mucosa are those that are more readily detected by careful clinical and macroscopic observation rather than exhaustive histopathological examination of the buccal mucosa in laboratory animals -provided the basic toxicity profile of a novel agent is adequately assessed in the usual preclinical studies. oral irritation studies are used in the testing of products for use in the oral cavity, mainly for surgical, dental and hygiene purposes but also therapeutic agents administered by the sublingual route. this route may be selected for substances that are broken down in the stomach or show a rapid first pass effect. as it is technically not feasible to perform full preclinical toxicity studies by the sublingual route, conventional oral or parenteral routes are preferred for systemic toxicity studies on such compounds. the choice of the best route will to a large extent be dictated by pharmacokinetic considerations. however, it is necessary to assess local irritancy potential to oral mucosa using a laboratory animal model. test species for oral irritation studies are usually rats, hamsters (cheek pouch), guinea pigs, dogs or primates using gross and histopathological assessment. a similar scheme to that employed in the histological assessment of skin irritancy is appropriate. inflammation of the oral cavity (stomatitis) may involve the buccal mucosa, the gingiva (gingivitis), the tongue (glossitis) and the peridontal tissues (peridontitis). although inflammatory lesions are found sporadically in untreated laboratory rodents, dogs and primates, stomatitis can be induced by systemic administration of high doses of therapeutic agents. anticancer and antimitotic agents are particularly liable to induce stomatitis. a notable example is bleomycin that is capable of producing stomatitis as part of its general effect on squamous cells (thompson et al., 1972) . in humans, the adverse effects on therapeutic ionising radiation on the salivary glands may also give rise to inflammatory changes in the oral cavity (fox, 1998) diuretics and other agents, which are capable of producing severe electrolyte disturbances and uraemia at excessive doses, can also produce stomatitis when then are administered in high doses to laboratory animals (garthoff et al., 1982) . these lesions may be analogous to the well-described association of ulcerative stomatitis and uraemia in man and laboratory animals (boyd, 1978; barker and van dreumel, 1985) . the dog appears very sensitive to the ulcerogenic effects of uraemia in the oral cavity, although as there is a poor correlation between actual levels of blood urea and stomatitis, other biochemical factors are undoubtedly involved. compounds, which effect pigmentation of the skin, can produce similar changes in pigmented oral mucosa. a number of drugs including chlorpromazine, quinacrine, chloroquine, amodiaquine and pyrimethamine cause pigmentation of the oral mucosa in man notably over the hard palate. chloroquine and pyrimethamine have also been shown to significantly increase numbers of active melano-cytes within the palatal mucosa of pigmented da rats when treated orally for 12 weeks (savage et al., 1986) . melanocytes in treated rats were shown to be enlarged and packed with melanin pigment and to possess extensive arborisation of cell processes between squamous cells. an experimental inhibitor of platelet aggregation, which produced pigment loss in the dark hair of long-evans rats and the skin of beagle dogs, also induced pallor of the normally pigmented oral mucous membranes in dogs (gracon et al., 1982; walsh and gough, 1989) . apart from loss of pigment, the histology of the mucous membranes and skin was normal. the tongue is conveniently sectioned for histological study, although reliance is often placed on careful visual inspection, because the usefulness of systematic histological examination of the tongue in routine preclinical safety studies has not been clearly established. a few lesions occur which are fairly specific to the tongue. amyloid may become deposited in the muscular and connective tissue of the tongue in amyloid-prone species, particularly mice (dunn, 1967) . mice, especially dba and dba/2ncrj strains, are liable to develop calcification in the lingual muscle spontaneously, even at a young age (imaoka et al., 1986) . calcified lesions are seen in the longitudinal muscle under the dorsolateral epithelium and the central part of the tongue, which, when severe, are associated with inflammation, granulation tissue, polypoid change, hyperplasia of the overlying squamous epithelium and ulceration. the histogenesis of this lesion is uncertain. in the dba/2ncrj mice, mineralisation of the tongue is associated with myocardial and aortic mineralisation (doi et al., 1985) . therapeutic agents can induce inflammatory lesions in the tongue. an example is provided by the investigational anticancer immunotoxin, zd0490, a mouse monoclonal antibody (c242) against colorectal carcinoma antigen conjugated to recombinant ricin a-chain. when administered to wistar-derived rats, this agent produced myocyte necrosis and inflammation specifically located below the ventral subepithelial surface of the tongue (westwood et al., 1996) . as the changes were different to the low grade myositis seen elsewhere in treated animals, these authors speculated that the changes in the tongue may have been related to the particular receptor profile of this area targeted by the monoclonal antibody. in common with other changes induced in the digestive tract of rats and cynomolgus monkeys by the administration of recombinant human epidermal growth factor, the tongue showed squamous epithelial hyperplasia characterised by a uniform increase in the thickness of the squamous epithelium in both species (breider et al., 1996; reindel et al., 1996) . at high doses, the squamous epithelium of the tongue of the primates was twice the thickness of control mucosa associated with elongation of rete pegs. teeth are usually only inspected by naked eye in conventional toxicity studies and this is appropriate for the assessment of a mature dentition. however, there has been an increasing awareness of dental lesions in toxicity studies, particularly as the teeth are visualised when the maxilla is examined histologically in inhalation studies. study of the rodent dentition in inhalation studies has shown that spontaneous lesions of the dentition are quite common. in one laboratory, malformations (dental dysplasia) of the maxillary incisors were observed in 3% of female and 9% of male cd-1 mice and 14.5% female and 10.5% sprague-dawley rats in 24 and 18 month inhalation studies respectively (losco, 1995) . the rat incisor and its pathology has been the subject of an excellent review (kuijpers et al., 1996) . unlike in humans, the rodent incisor continues to grow and differentiate throughout life and is renewed every 40-50 days. located at the centre of the tooth is the vascular pulp. this is surrounded by proliferating ondotoblasts which form predentin which when calcified becomes dentin. surrounding ameloblasts when induced by the presence of dentin produce the overlying enamel layer. it is these active cellular layers, which can be modified or damaged by xenobiotics, vitamin deficiencies, calcium, phosphate or magnesium deficiency, parathyroidectomy, hypophysectomy, hyperparathyroidism, adrenal insufficiency and fluorosis (kuijpers et al., 1996) . although in humans the mature dentition is no longer growing, in children the dentition is in a growth phase that starts in utero and lasts into the second decade. as increasing numbers of children survive malignant disease, damage to the mature dentition can occur as a result of cytotoxic therapy during childhood. clinical study of the teeth of children treated for malignancy have shown increased incidence of enamel hypoplasia and missing teeth (welbury et al., 1984) . histological examination of teeth from children treated with vincristine or combination chemotherapy for malignant disease has demonstrated prominent incremental lines in dentine correlating with the number of times the intravenous cytotoxic agents were administered (macleod et al., 1987) . it has also been shown that vincristine, a drug which interferes with the assembly of microtubules and reduces secretory activity in a number of cells including osteoblasts and chondroblasts, also effects dentine formation in the rat incisor (stene and koppang, 1976) . two weeks following a single intravenous dose of vincristine to young adult rats, a faint incremental line in the dentine was observed, probably a reflection of a direct effect of the drug on the dentinogenic tissue at the time of injection. at higher doses, focal niche-like or punched out defects in dentine were observed, expression of more severe injury to highly sensitive dentinogenic populations at the time of injection (stene and koppang, 1976 ). the precise mechanism of damage is not fully understood although decreased secretion of dentine matrix by odontoblasts has been demonstrated. calcification appears unaltered (stene and koppang, 1980) . administration of the alkylating anticancer agent cyclophosphamide or a sin-gle exposure to ionising radiation, produces localised niche-like or punched out defects in the rat incisor, rather than the more diffuse changes induced by vincristine. this presumably reflects more localised injury to a sensitive subpopulation of dentinogenic cells (koppang, 1973; vahlsing et al., 1997) . anticonvulsant drugs also produce changes in the dentition of man and experimental animals. in humans, reported alterations include tooth root resorption, small teeth, delayed shedding of deciduous teeth and retarded eruption of permanent teeth, features similar to those found in hypoparathyroid or pseudohypoparathyroid conditions (robinson et al., 1983) . tooth root alterations were also reported in a study in which young male wistar rats were treated with diphenylhyantoin for 1 month. treated rats showed evidence of molar root resorption lacunae that penetrated the cementum and involved the dentine. the lacunae contained a dense infiltrate of cells contiguous with similar cells in the surrounding periodontal ligament. robinson and harvey (1989) showed that the changes were similar to those occurring in parathyroidectomized rats but not those in rats made hypocalcaemic with a calcium deficient diet. they suggested that the changes induced by diphenylhydantoin in rats were similar to those in pseudohypoparathyroidism in which resistance of tooth roots to resorption is reduced. discoloration of teeth and bone is a well-described side effect of tetracycline administration and it has also been reported in patients treated with the semisynthetic derivative, minocycline (cale et al., 1988) . interestingly the ameloblastic epithelium of the enamel forming tissues of growing incisors in wistar rats treated with high doses of human recombinant epidermal growth factor showed hyperplasia characterised by pseudostratification, increased nuclear-cytoplasmic ratio and increased cytoplasmic eosinophilia (breider et al., 1996) . this finding is consistent with the presence of epidermal growth factor receptors in the cells of the enamel organ (martineau et al., 1991) . periodontitis is a common and important disease in man and animals although overt cases are not usually seen in toxicity studies. however, periodontitis of a degree sufficient to disrupt chronic rat toxicity and carcinogenicity studies has been reported. robinson (1985) described periodontitis in alpk/ap rats in which there were erosive granulomatous cavities adjacent to molar teeth with fistulas opening into the nasal cavity. these changes were associated with penetrating food fibres in the gingival sulcus and it was suggested that the presence of long pointed food fibres in the powdered diet was the main reason for occurrence of periodontitis. peridontitis in rodents also results from the effects of dental pathology such as fractures, malformation or malposition of incisors (losco, 1995) . drug-induced overgrowth of the gingival tissues is a well-described phenomenon in both humans and laboratory animals including dogs, cats, and rats. in man, these changes have been associated with diphenylhydantoin (phenytoin) (beghi et al., 1986) nifedipine, calcium channel blockers (ledermann et al., 1984) , cyclosporin a (barthold, 1987) and valproic acid (syrjamen and syrjamen, 1979) . cyclosporin a, diphenylhydantoin and calcium channel blockers have been associated with similar changes in laboratory animals (do'nascimento et al., 1985; latimer et al., 1986; waner et al., 1988) . in most instances there is swelling of the gingiva by firm nodular overgrowths around the teeth. histologically, these overgrowths are characterised by marked acanthosis of the squamous epithelium overlying connective tissue that is infiltrated by large numbers of chronic inflammatory cells. fibrovascular proliferation may be marked. in patients treated with cyclosporin, myxomatous degeneration is described in association with dense infiltration of plasma cells and lymphocytes (barthold, 1987) . secondary acute inflammation in association with food debris and hair shafts is described in dogs treated with oxodipine (waner et al., 1988) . the forces behind these changes are unclear. studies of changed induced by nifedipine and hydantoin have shown increases in extracellular ground substance and increased numbers of fibroblasts containing sulphated acid mucopolysaccharides (kantor and hassel, 1983; lucas et al., 1985) . these drugs may alter fibroblastic proliferative and synthetic activity, possibly by selection of a subpopulation of fibroblasts (hassel et al., 1976) . it has also been suggested that an underlying mechanism in phenytoin-induced gingival hyperplasia involves the decrease in salivary iga that develops in some patients (beghi et al., 1986) . study of cyclosporin a-induced changes have suggested that impairment of t lymphocyte function may permit overgrowth of oral bacteria and bacterial products which may influence fibroblast function (barthold, 1987) . a spontaneous form of gingival hyperplasia has been described in non-human primates (macaca mulata). this is characterised by an enlargement of the marginal and alveolar gingiva by connective tissue consisting of relatively poorly cellular bundles of collagen fibres. the lesions show little inflammatory alterations and the overlying squamous epithelium shows mild hyperkeratosis only (schiã¸dt et al., 1994) . this pathology is similar to hereditary gingival fibromatosis in humans. sessile or pedunculated squamous papillomas and infiltrating squamous carcinomas are occasionally found in the oral cavity of most laboratory animals including rodents (odashima, 1979; emminger and mohr, 1982; leiniger and jokinen, 1994; takahashi and okamiya, 1996; mohr, 1997) , rabbits (sundberg et al., 1985; sundberg and everitt, 1986) , and beagle dogs (watrach et al., 1970) . the microscopic structure of these neoplasms in rodents resembles those occurring in squamous epithelium in other sites. although a number of agents induce squamous neoplasms in the oral cavity, spontaneous squamous carcinomas are generally uncommon spontaneous lesions in laboratory animals. however, some strains of rodent may develop squamous neoplasms more commonly. for instance, in life time studies ad libitum fed brown-norway rats, 21% of males and 32% females developed oral squamous cell carcinomas although only 9% and 10% in food-restricted animals respectively (thurman et al., 1997) . it was suggested that certain pedigrees possessed a genetic predisposition to these neoplasms. papillomas occurring in rabbits and dogs are of note because they can occur in quite young animals, apparently as a result of infection with viruses of the papilloma group. viral inclusions may be seen in histological sections. the implications of papilloma viruses in laboratory species are that the progression of virally induced papillomas to malignant squamous carcinomas can be potentiated by nonviral factors including application of xenobiotics (howley et al., 1986) . in rabbits, the prevalence of oral papillomas varies considerably but they are quite common in some laboratory strains. they are overlooked because of their small size and a distribution limited to the ventral surface of the tongue (sundberg et al., 1985) . microscopically, they are typical squamous papillomas composed of irregular acanthotic squamous epithelium and a fibrovascular stalk of variable size. squamous cells at the margins of papillomas at the junction with normal mucosa, often show large, oval nuclei, marginated chromatin and central, basophilic, intranuclear inclusions, which electron microscopic examination shows to contain viral particles. oral papillomas in dogs develop as multiple growths, regressing spontaneously after a few months. they are also caused by a virus of the papilloma group, which possesses a high degree of specificity for the mucosa of the oral cavity and adjacent skin (watrach et al., 1970) . histologically, they are composed of proliferative masses of epithelial cells, keratinised on the surface and resting on an irregular connective tissue stroma or pedicule. large vesicular cells with basophilic intranuclear inclusions are also found in the granular cell layer, identifiable as virus arrays by electron microscopy (cheville and olson, 1964) . malignant change has been described in these canine lesions and this can occur in young beagle dogs (watrach et al., 1970) . although many types of papilloma viruses have been identified in both man and animals (pfister, 1984) , common antigenic determinants exist between viruses in different species. this immunological cross-reactivity can be exploited in the immunocytochemical localisation of papilloma viruses in epithelial lesions of many animal species. papilloma virus antigen has been demonstrated in oral papilloma of dogs and rabbits using antisera to bovine papilloma virus type i (sundberg et al., 1984) . cells positive for virus and viral inclusions are located in the upper layers of the epithelium, especially within cells of the granular layer. spontaneously developing odontogenic tumours are rare in rodents but they have been induced in laboratory animals given carcinogens such as nitrosoureas or exposed to ionising radiation (gã¶ssner and luz, 1994) . a range of tumours originating from dental tissues with epithelial, mesenchymal or mixed appearances has been reported in rodents (kuijpers et al., 1996) . the classification of odontogenic tumours is complex and confusing. they range from benign anomalies and cystic structures through to malignant neoplasms. the ameloblastoma comprises cords, nests, anastomosing strands or islands of ondontogenic epithelial cells within a fibrous stroma. the tumour cells resemble ameloblasts with the cords of spindle shaped cells similar to the stellate epithelium bounded by a peripheral layer of cuboidal or columnar cells resembling the inner enamel epithelium. other tumours of the odontogenic epithelium show induction of the mesenchymal elements or develop a complete sequence of odontogenic epithelium, odontogenic mesenchyme and dental hard tissues including dentine, enamel and cementum. in the rat these have been classified as odontoma characterised by the presence of all dental hard tissues and odontogenic fibroma composed of undifferentiated or primitive mesenchymal cells of developing dental tissue (mohr, 1997) . odontogenic tumours developing in fischer rats treated with aflatoxin were located in the upper jaw associated with the incisor teeth and were composed of proliferating fibroblast-like cells within which ovoid calcified bodies resembling cementum were seen (cullen et al., 1987) . occasional inclusions of solid epithelial nests were also seen. no metastatic deposits were found although the neoplasms were locally aggressive. in addition, squamous tumours and neoplasms of mesenchymal origin typical of other organs, bones and soft tissues are found in this region. although salivary glands may not represent vital organs in the same sense as the kidneys or heart, severe derangement of their secretions can alter both the quality and quantity of saliva. depending on the particular glands and cells affected, dry mouth, mucositis, and dental caries may develop . the severe oral complications of irreversible salivary damage and dysfunction, which can occur patients with head and neck cancer as a consequence of local irradiation, may have a significant impact on the efficacy of therapy, quality of life and survival (fox, 1998) . a protective layer of mucus, a visco-elastic material containing high molecular weight glycoproteins produced by the major and minor salivary glands, covers the stratified squamous mucosa of the oral cavity. these mucins usually contain more than 50% carbohydrate in the form of neutral and acidic oligosaccharide chains, o-glycosidically linked to threonine or serine. mucins possess several roles including mechanical flushing of the oral cavity, protection and lubrication of soft and hard tissues, modulation of oral microbial flora, buffering activity, regulation of calcium/phosphate equilibrium, digestion and extracellular post translation processing of molecules present in saliva . the heterogeneity of salivary glycoproteins suggests that they act as a defence against pathogenic microorganisms by competing with microbial binding sites of similar structure on the surface of cells lining the digestive tract (schulte, 1987) . minor salivary glands may also play an important part in the local immunosurveillance of the oral cavity for their ducts are anatomically closely associated with lymphoid tissue (nair and schroeder, 1986; nair et al., 1987) . salivary secretions also possess digestive enzyme activity although in herbivores and carnivores, it is usually low in contrast to high digestive enzyme activity in omnivorous species (junqueira et al., 1973) . the phylogenetic association of the salivary glands with the thyroid gland is evident functionally because salivary glands are capable of concentrating iodide in their secretions, although this is not under control of thyroid stimulating hormone (ingbar, 1985) . it has been shown that thyroxine accelerates the differentiation of the granular convoluted tubule cells and the appearance of epidermal growth factor in the submandibular gland of the neonatal mouse (chabot et al., 1987) . the structure of the salivary glands differs among laboratory species, between different glands in the same species and between sexes. it is usually considered that there are three major salivary glands, the parotid, the sublingual and the submandibular (submaxillary) glands. minor salivary glands are scattered in other locations throughout the mouth and oropharynx. in dogs and other carnivores, the zygomatic (infra-orbital) gland, located just below the zygomatic arch and the buccal (molar) gland are also often referred to as major salivary glands. microscopically, salivary glands are composed of secretory glands or 'endpieces' attached to a connecting system of intralobular and extralobular (secretory) ducts. secretory endpieces may be acinar or tubulo-acinar in nature. the secretory cells have been subdivided into serous, mucous, seromucous and special serous types. controversy remains about the precise nature of the secretory cells found in the various salivary glands of different species and this makes critical interspecies comparisons difficult (see detailed discussion of this problem by pinkstaff, 1980 ). the duct system is less complex. this comprises an intercalated duct which leads from the secretory endpiece into a striated (secretory or intralobular) duct, so termed because their lining cells are striated by delicate eosinophilic cytoplasmic rods. the striated ducts converge into interlobular ducts and a main excretory duct system. in rats, mice and hamsters, an overall similarity in gross and microscopic anatomy of the various salivary glands exists although there are histochemical differences (munhoz, 1971; glucksmann and cherry, 1973; dawe, 1979; pinkstaff, 1980; emmiger and mohr, 1982) . moreover, it has been demonstrated that salivary glands in rodents as well as a number of other species show morphological and histochemical sexual dimorphism (pinkstaff, 1998) . the sublingual gland in rats, mice and hamsters is composed principally of mucous acini, with indistinct serous demilunes. acini open into fairly long intercalated ducts lined by flat or cuboidal cells devoid of granules. the parotid gland is composed of serous-type secretory cells containing zymogen granules and prominent hyperchromatic basal cytoplasmic poles. the submandibular gland is anatomically the most complex salivary gland in rodents. secretory endpieces are composed of small or moderately sized cells with foamy cytoplasm and basophilic basal poles. the most striking feature is the presence of an additional duct segment interposed between the intercalated and striated ducts. this segment is lined by cylindrical epithelium with basal nuclei and eosinophilic cytoplasm containing secretory granules. this duct segment is termed the granular duct or granular convoluted tubule. these granular cells are of special interest because they contain a large number of heterologous biologically active peptides including nerve growth factor, epidermal growth factor, renin, and kallikrinins (barka, 1980; mori et al., 1983 ). the precise physiological role of many of these peptides in salivary gland remains uncertain. epidermal growth factor was originally isolated from the mouse salivary gland. it initiates premature eyelid opening and incisor eruption when injected into the neonatal mouse (cohen, 1962) . study of the mouse submandibular gland has shown that both epidermal growth factor and nerve growth factor are released into saliva following the administration of phenylephrine, sympathomimetic amine acting mainly on î±receptors and isoprenaline (isoproterenol), a î²-adrenergic agent (murphy et al., 1980) . immunohistochemical study also demonstrates that epidermal growth factor becomes depleted in mouse salivary tissue following administration of phenylephrine and similar agents (tsukitani and mori, 1987) . phenylephrine has been shown to cause marked secretory activity accompanied by loss of granules from granular cells, as well as loss of immune reactive carbonic anhydrase, an enzyme which participates both in membrane transport of bicarbonate ions into saliva and glandular secretion (noda, 1986) . morphological studies have shown that both acinar and granular tubular cells participate in this response to adrenergic agents (murphy et al., 1980) . this is in contrast to the effects of pilocarpine, a cholinergic agent, which elicits the secretion of saliva deficient in serous proteins with little or none of the growth factors, as its effects are more limited to acinar cells. glycoprotein secretion of rodent salivary glands has stimulated histochemical study using both conventional mucin histochemical techniques and labelled lectins which possess affinity for specific sugars or sugar sequences (tables 1 and 2 , pages 340 and 342). studies of rat, mouse and hamster salivary glands using batteries of labelled lectins have shown a greater heterogeneity of oligosaccharides in salivary glands than seen by classical histochemical techniques. there are considerable species differences and variations between murine strains and sexes of the same strain as well as heterogeneity among morphologically similar cells within one gland (schulte and spicer, 1983, 1984; schulte, 1987) . the results of histochemical studies are in excellent agreement with studies using biochemical methods but suggest a significant influence of genetic and hormonal factors on the synthesis of salivary glycoproteins. less attention has been paid to the structure and cytochemistry of the dog salivary glands. there appears to be little variation between the structure of salivary tissues between beagles and other strains although variation with age has been reported (reifel and travill, 1972; nagoyo and tandler, 1986) . munhoz (1971) has described the histochemical features of the dog parotid gland. the dog parotid is of seromucinous type secreting both acidic and neutral mucosubstances, in contrast to the more neutral mucosubstances secreted by rodent glands. the salivary glands of non-human primates are similar to those in man. they possess parotid glands of serous or seromucous type, submandibular glands with both serous and mucous acini and sublingual glands of mainly mucous type. the salivary glands of the non-human primate react to adverse stimuli such as ionising radiation in a similar manner to human salivary tissue . focal chronic inflammation of the salivary glands occurs sporadically in untreated rats, mice, hamsters, dogs or primates employed in toxicology although severity and prevalence is variable. sialoadenitis as a result of a corona virus, the sialodacryoadenitis virus, is a well-known and fairly ubiquitous condition in rats, first described by innes and stanton (1961) . the condition is characterised histologically by oedema and congestion of submandibular and parotid salivary glands as well as extra-orbital lachrymal and harderian glands. it is accompanied by inflammation of variable severity and chronicity in both glandular and connective tissue as well as degeneration and necrosis of duct epithelium (fig. 44) . the regenerative hyperplasia of the duct epithelium may be quite intense about a week after infection but all changes regress after about 2 weeks and glands are essentially normal after 3 or 4 weeks (carthew and slinger, 1981; percy and wojcinski, 1986) . there may be a delay in the appearance of inflammatory cells and the onset of repair in rats immunosuppressed with cyclophosphamide (hanna et al., 1984) . depletion of salivary gland epidermal growth factor also occurs during the infection (percy et al., 1988) . suppurative infections in the neck region of the rat such as those produced by klebsiella aerogenes also cause acute and chronic inflammation of salivary glands with fibrosis and glandular proliferation of salivary tissue (arseculeratne et al., 1981) . sialadenitis occurs spontaneously in autoimmune-prone strains of mice such as the nzb/nzw and sl/ni strains and it has been reported in ageing female, but not male bdf1 mice (hayashi et al., 1988) . the non-obese diabetic mouse known for its spontaneous insulin-dependent diabetes mellitus also develops immune mediated damage to submandibular glands (fujino-kurihara et al., 1985; tã¶rnwall et al., 1999) . in ageing bdf1 females the submandibular gland was shown to be involved by a destructive inflammatory process characterised by an intense infiltration by small and medium sized lymphocytes, associated with mild inflammation in other organs such as the parotid and sublingual glands, pancreas and kidney. immunocytochemistry showed that most of the lymphocytes were t cells (thy-1.2 and lyt-1 positive) of the helper/inducer subset (l3t4 or cd4 positive) and less than 10% were of suppresser/cytotoxic (lyt-2 or cd8 fig. 44 . section from salivary tissue from a sprague-dawley rat during an infection with the sialodacryoadenitis virus showing intense ductular inflammation. (he, ã�25.) positive) type (see haemopoietic and lymphatic systems, chapter iii). circulating anti-salivary duct antibody of igg was also detected in afflicted mice. it was suggested that helper/inducer t cells played a key role in the production of this change, unlike induced autoimmune sialoadenitis in which cytoxic t-cell subsets may directly destroy glandular tissue. it has been suggested that this process in ageing females is related to the decline in the number of splenic lyt-2 cells in mice with advancing age (hayashi et al., 1988) . these cells are believe to be the most susceptible to ageing (see haemopoietic and lymphatic systems, chapter iii) in the non-obese diabetic strain of mouse derived from jcl-icr mice, a periductal chronic inflammatory infiltrate is found in the submandibular gland at about the same time that immune-mediated insulitis is most marked. this suggests that there is an extension of the autoimmune process to salivary tissue (fujino-kurihara et al., 1985) . it is probable that helper/inducer cd4 t cells are essential components of this infiltrate and a number of cytokines and their receptors such as ip-10 (interferon-î³ inducible protein 10) and rantes (regulated upon activation normal t cell expressed and secreted) may have an important role (tã¶rnwall et al., 1999) . an autoimmune type of sialoadenitis can also be experimentally induced certain strains of mice. crj:cd1 mice, thymectomized at 3 days, a time point at which lyt-2 positive cells (cd8 suppresser t lymphocytes) can be maximally reduced, followed by immunisation at 28 and 42 days after birth with homogenates of salivary gland and complete freund's adjuvent, develop a distinctive sialoadenitis in the submandibular and to some extent the parotid glands (hayashi et al., 1985) . this sialoadenitis is characterised by degenerative changes in salivary glandular tissue associated with an extensive and intense infiltrate of small and medium sized lymphocytes. these cells appear shortly after immunisation but increase in number with time. immunocytochemical study has shown that many of these cells are reactive to antisera to thy-1.2 and lyt-2 (cd8) features of suppresser/cytotoxic t lymphocytes. later appearing cells demonstrate features of plasmacytoid lymphocytes and contain immunoglobulin of mainly igg class (hayashi et al., 1985) . these authors therefore suggested the sialoadenitis appeared as both a result of cytotoxic/suppresser t-cell activity and an antibody-dependent cell-mediated cytotoxicity. in the hamster salivary glands, interstitial infiltrates of lymphocytes and plasma cell are quite common and may become more marked with advancing age (mcmartin, 1979) . whereas necrosis of the parotid gland of uncertain aetiology sometimes occurs in the dog, mild focal chronic inflammation is quite a common incidental finding in canine salivary glands and has been reported in about 5% of normal beagle dogs (kelly et al., 1982) . although the inflammation in salivary tissue which results from ionising radiation is only indirectly relevant to drug safety evaluation, it is of interest in view of the notable species differences in sensitivity to this form of insult. serous acinar cells in man and rhesus monkey appear least resistant to the effects of ionising radiation, where damage is characterised by widespread degranulation and degeneration of acini, infiltration by polymorphonuclear cells followed by lymphocytes, plasma cells and subsequent atrophy and fibrosis . these changes contrast with the lesser effects of ionisation radiation on the rodent salivary glands in which there is little or no acute inflammatory response. lymphoid bodies are sharply circumscribed collections of lymphoid cells generally located between the parotid and sublingual glands close to a cervical lymph node in mice. they are apparently normal aggregates of lymphoid tissue. like many other glandular organs, the size of the secretory tissue of the salivary gland is responsive to functional demand and is subject to age-related changes. in man, the gland parenchyma frequently becomes atrophic and replaced by connective tissue or fat with advancing age, possibly partly related to vascular changes (waterhouse et al., 1973; scott, 1977) . in ageing rats, the extent and height of granular ducts and their content of mature secretory granules has also been shown to decrease with age (sashima, 1986) . dietary factors influence salivary gland size. decreased food consumption or protein starvation can reduce the weight of salivary glands in rats. there is shrinking of mucous and serous glands and loss of zymogen granules associated with decreased rna but unchanged dna content, attributable to the reduced requirements for protein synthesis , mcbride et al., 1987 . as salivary gland function is responsive to adrenergic stimulation, it is not surprising that atrophy occurs following adrenergic blockade. the weights of the submandibular gland in mice were shown to decrease following administration of the î²-adrenergic blocking agent, propranolol (smith and butler, 1977) . this was associated with a reduction in stainable neutral mucins and a decrease in the thickness of the acinar cells making the gland lumens appear larger than normal. the cytotoxic agent, alloxan, known primarily for its specific effect on pancreatic b cells, has also been shown to produce weight loss of the rat submandibular gland, associated with lipid inclusions in the acinar cells, capillary basement membrane thickening and reduced salivary flow (reuterving et al., 1987) . it is probable that alloxan exerts a cytotoxic effect on the acinar cells of the rat submandibular gland (sagstrã¶m et al., 1987) . methotrexate, a folic acid antagonist, has also been reported to cause vacuolization of acinar and ductular cells with reduction of secretory granules in rat salivary glands (mcbride et al., 1987) . ligation of the main excretory ducts has frequently been used as an ex-perimental model for study of salivary gland atrophy as well as the regeneration that follows removal of the ligature. there is marked atrophy of all cell types but most markedly the acinar cells through apoptosis. although overt necrosis has been reported following ligation of the excretory duct, it appears that this may have been the result of constriction of the vasculature for acinar cells are relatively intolerant to a decrease in oxygen and nutrient (denny et al., 1997) . a number of therapeutic agents increase salivary gland size in man, although the scarcity of biopsy data precludes a critical assessment of the precise mechanism in many cases. drugs reported to produce salivary gland enlargement in man include iodide-containing radiological contrast media, isoprenaline and anti-inflammatory agents phenylbutazone and oxyphenbutazone. enlargement may also occur after endotracheal anaesthesia and upper gastrointestinal tract endoscopy in man (riddell, 1982) . some of these agents and procedures may produce spasm of large salivary ducts and retention of secretions. several pharmacological agents, particularly sympathomimetic amines, have been shown to produce increases in salivary gland size in rodents following repeated dosing (brenner and stanton, 1970) . there is an intimate relationship of sympathomimetic amines with the control of the secretory process in salivary tissue. whereas a single injection of isoprenaline (isoprotorenol) in the range of 20-200 mg/kg induces discharge of preformed secretory granules followed by gradual re-synthesis and reconstitution, repeated injections produces an increase in the size of salivary glands (simson et al., 1974) . histologically, the enlarged glands are composed of secretory cells of increased size that contain increased amounts of secretory substances in the cytoplasm (simson et al., 1974) . although these histological features are principally those of diffuse cellular hypertrophy, the increase in dna content and radioactive thymidine uptake described in the salivary tissue following repeated administration of isoprenaline suggests that a degree of hyperplasia also occurs (barka et al., 1972) . these effects do not depend on the integrity of the autonomic nerves because they occur after ablation of the autonomic ganglia (barka et al., 1972) . they appear to be mediated by an effect on adrenergic î²-receptors. the effects can be blocked by propranalol, a î²-receptor antagonist but not by phenoxybenzamine, an î±-receptor antagonist (brenner and stanton, 1970) . as theophylline and caffeine also elicit salivary gland enlargement in rats, a role for cyclic 3',5-adenosine monophosphate (camp) in salivary gland enlargement has been postulated (brenner and stanton, 1970) . detailed study of hypertrophy, protein synthesis, and intracellular camp activity in the salivary glands of rats treated for 10 days with isoprenaline (isoproterenol), a series of î²-adrenergic receptor agonists and the phosphodiesterase inhibitors, theophylline and caffeine, showed that similar effects occurred with all agents although differences in the degree of hypertrophy, the nature of pro-tein and glycoprotein synthesis and golgi membrane enzyme activity were recorded (wells and humphreys-beher, 1985) . the parotoid gland showed the most pronounced hypertrophy followed by the submandibular gland but the sublingual gland appeared to be unaffected by treatment. the degree and nature of the changes induced by the various î²1/î²2 receptor agonists suggested that most of these effects were mediated through î²1 receptors which are present in greatest numbers on the parotid and salivary cells. it was suggested that the effects of î²-adrenergic agonists on salivary gland are produced by a receptor-mediated stimulation of adenylate cyclase activity causing an increase in levels of intracellular camp. however, other factors may be important for wells and humphreys-beher (1985) also showed that although isoproterenol and caffeine increased salivary cell camp to comparable levels, the hypertrophy was greater with isoproterenol. cardioactive phosphodiesterase inhibitors were shown to produce submaxillary hypertrophy in rat subacute toxicity studies (rogers et al., 1985; jayasekara et al., 1986; smith et al., 1988) . parotid and submaxillary glands were those most affected by the inotropic phosphodiesterase inhibitor ici 153,110 (westwood et al., 1991) . as the agents produced their positive inotropic action via selective inhibition of the cardiac phosphodiesterase subfraction iii specifically requiring camp as its substrate, it was suggested that the salivary gland hypertrophy was a result of phosphodiesterase inhibition (smith et al., 1988) . other classes of drugs can also produce salivary gland enlargement in rats in repeated dose studies. doxylamine, a representative of the widely used ethanolamine group of antihistamines, has been reported to produce marked cytomegaly in the fischer 344 rat parotid gland. enlarged cells were characterised by a basophilic and coarsely granular or vacuolated cytoplasm (jackson and blackwell, 1988) . the b6c3f1 mouse did not develop these changes after a similar treatment schedule. in view of the presence of considerable amount of epidermal growth factor in salivary glands, it is of interest to note the effects of its administration to laboratory animals. salivary gland weights were increased in rats and cynomolgus monkeys infused with high doses of recombinant human epidermal growth factor (breider et al., 1996; reindel et al., 1996) . however histological features seen are primarily those of ductular epithelial hyperplasia (see below under hyperplasia). epithelial cells characterised by abundant granular eosinophilic cytoplasm as a result of the accumulation of mitochondria are often referred to as oncocytes, a term used by hamperl (1950) to describe similar cells in hã¼rthle tumours of the thyroid gland. they may be found in various focal nodular and neoplastic states of the salivary glands in both man and laboratory animals. the precise significance of these cells is uncertain. the mitochondria usually appear unremarkable except for lack of dense granules and it has been suggested that the mito-chondrial changes represent an adaptive phenomenon or compensatory hyperplasia (ghadially, 1982) . in human salivary tissue their prevalence seems to increase with advancing age and they can be associated with hyperplastic lesions or neoplasms such as oxyphil adenomas and adenolymphomas. eosinophilic cells also occur in the salivary glands of certain strains of aged rats (bogart, 1970) and in mice with experimentally induced autoallergic sialoadenitis (takeda et al., 1985) . in the study of takeda et al. (1985) the eosinophilic cells appeared to arise predominantly in the secretory (glandular) ducts of the submandibular glands, although eosinophilic cells can apparently develop from either duct or acinar cells. well-defined, unencapsulated foci of enlarged acinar cells occur spontaneously in the salivary glands, particularly the parotid of rats, mice (chiu and chen, 1986) , and hamsters although their reported incidence varies between laboratory. the enlarged cells possess greatly expanded cytoplasmic volume that retains a vesicular, vacuolated or foamy appearance or possesses a pale eosinophilic granular texture. the basal parts of the cells usually stain intensely blue in haematoxylin and eosin stained sections and contain large, dense, irregular hyperchromatic or pyknotic nuclei showing little evidence of mitotic activity. although there has been little ultrastructural study of these foci, the cytoplasmic alterations appear to be distinct from those of so-called oncocytes that characterised by granular eosinophilic cytoplasm packed with mitochondria. the biological nature of these foci is uncertain. the lack of any prominent mitotic activity, cell proliferation or expansive growth suggests that they are most aptly regarded as hypertrophic lesions (chiu and chen, 1986 ). although they increase in prevalence with increasing age in certain strains of rat, there is no evidence to suggest that they represent pre-neoplastic lesions or possess any relationship with development of neoplasia in salivary tissue (dawe, 1979) . hyperplasia and squamous metaplasia of the salivary ducts are common features of many inflammatory and reactive conditions in the salivary glands of rodents, dogs, non-human primates and man and can be associated with the presence of stones and calculi with the duct system. squamous metaplasia and regenerative change in the ducts occurs in rats afflicted with sialodacryoadenitis (carthew and slinger, 1981) . it is also described specifically located in the ducts of the sublingual glands in the wistar rat in the absence of obvious sialodacryoadenitis or evidence of any specific disease. similar regenerative hyperplastic duct changes are also seen in necrotic and inflammatory conditions in the dog salivary gland . detailed morphological examination with immunocytochemical study of epi-dermal cytokeratins of the rat salivary gland after arterial ligation has shown that the acinar units can also undergo squamous metaplasia (dardick et al., 1985) . it appears that the acinar-intercalated duct complexes can rapidly reprogram to produce epidermal cytokeratin filaments in ischaemic or inflammatory states. hyperplasia of the ductular epithelium appears to be the principle result of the administration of epidermal growth factor to rats and cynomolgus monkeys. in rats histological features were primarily of ductular epithelial hyperplasia without evidence of significant acinar hyperplasia (breider et al., 1996; reindel et al., 1996) . in primates the changes were most striking in the interlobular and large intralobular ducts where the epithelium showed multilayered and papilliform projections. however, mitotic activity was evident throughout the duct epithelia and acinar cells showed hypertrophy with depletion secretory granules and the presence of large vesicular nuclei . focal duct and acinar hyperplasia, showing minimal compression of the surrounding parenchyma and distinct from focal hypertrophy is also described in the classification of rat salivary lesions (mohr, 1997) . primary neoplasms of salivary glands are uncommon in the usual strain of rats and mice employed in carcinogenicity bioassays (haseman et al., 1998) . acinar and tubular adenomas and adenocarcinomas as well as squamous carcinomas are reported in rats (mohr, 1997) , mice (frith and heath, 1994) and hamsters (takahashi and okamiya, 1996) . some carcinomas showing squamous or glandular differentiation may be observed infiltrating the salivary gland that originate in other local structures of the head and neck region. occasionally, salivary gland neoplasms show adenomyomatous differentiation. mixed glandular and lymphoid tissue patterns resembling wartin's tumour in man are also sometimes seen. neoplasms of soft tissues also develop in and around the major salivary gland in rodents (see integumentary system chapter i). in humans the oesophagus is not considered a common site for drug-induced injury although some studies have suggested that medication-induced changes are more prevalent than previously supposed (bonavina et al., 1987) . severe damage can occur following prolonged contact between mucosa and ingested tablets or capsules which results in local high concentrations of potentially irritant substances (bott and mccallum, 1986; brors, 1987; kikendall, 1999; levine, 1999) . damage as a result of local contact may be more common in elderly subjects as the amplitude of oesophageal contractions decrease with age and capsules more liable to lodge in the lumen of the oesophagus (bonavina et al., 1987) . however patients of all ages may be affected. women have been injured more frequently than men probably because of the greater likelihood of their being treated with potentially injurious drugs (kikendall, 1999) . the shape and surface coating of tablets may influence their tendency to adhere to the mucosa and lodge in the oesophagus (marvola et al., 1983) . a wide variety of drugs have been implicated. in the united states the majority of cases appear to be caused by ingestion of tetracycline or doxacycline (levine, 1999) . some of the causative agents such as potassium chloride, aspirin and other non-steroidal anti-inflammatory drugs are also implicated in ulceration lower in the gastrointestinal tract. over recent years the bisphosphonate, alendronate has been one of the most commonly reported causes of adverse effects in the oesophagus with severe injury being reported. although injury is linked to ingestion without water or failing to remain upright after swallowing the medication, alendronate is particularly caustic (kikendall, 1999) . oesophagitis due to candida albicans is a well-described complication of antibiotic therapy. administration of immunosuppressive drugs may predispose to viral infections in the oesophagus. a number of agents affecting neuromuscular co-ordination may also predispose to gastro-oesophageal regurgitation and reflux oesophagitis (bott and mccallum, 1986) . in laboratory rodents spontaneous lesions of the oesophagus are occasionally seen. oesophageal impaction has been described in untreated srl:bhe rats. this is characterised by massive dilatation of the oesophagus with food or bedding (ruben et al., 1983) . histologically, the muscle fibres in the wall of the oesophagus show varying degrees of degeneration including swelling or shrinking of fibres, myofibrillar fragmentation, cytoplasmic vacuolation and mineralisation. so-called megaoesophagus, characterised by enlargement of the oesophagus, degeneration of muscle fibres and ganglion cells in the myenteric plexus has also been described in certain strains of rats and mice (harkness and ferguson, 1979; randelia and lalitha, 1988) . its cause is unknown. a commonly occurring lesion reported in fischer 344 rats is oesophageal hyperkeratosis, which occurs at all ages . in the study by maeda et al. (1985) , it occurred more commonly in rats fed a protein-restricted, calorie unrestricted diet than in rats fed ad libitum with normal diet. it was suggested that the particular high prevalence of oesophageal hyperkeratosis observed in all groups in this particular study was related to acidification of drinking water . another pathological findings in rodents is perforation of the oesophagus as a result of a gavage accident. under these circumstances there is a variable inflammatory and purulent exudate localised around the perforation or spread within the pleural or occasionally the pericardial cavities. the oesophagus and surrounding tissues need careful examination by the pathologist for it is not always clear from clinical findings that oesophageal damage has occurred. spontaneous oesophageal lesions are uncommon in laboratory beagles, even though emesis and vomiting are frequent responses of this species following dosing in toxicity studies. local oesophageal irritancy potential of drugs has been assessed in a number of animal models, notably the cat and pig (carlborg and densert, 1980; olovson et al., 1983) . in these models, the test drugs are placed in the upper oesophagus using endoscopic techniques for periods of several hours to allow dissolution of the preparation. subsequently, the animals are followed for 3-6 days and histopathological assessment performed on the oesophagus. the degree of inflammation, erosion of mucosa or deep ulceration is recorded in a semiquantitative manner. the degree of ulcerogenic activity of drugs in these models seems to correlate with reported ulcerogenic activity in the human oesophagus (carlborg et al., 1983) . systemic administration of drugs with radiomimetic or antimitotic activity can cause hypoplastic changes in the oesophageal mucosa as well as the remaining gastrointestinal tract mucosa (tucker et al., 1983) . conversely, hyperplasia with increased keratinization has been reported in the oesophagus of the rat following chronic high dose administration of alcohol (mascrã¨s et al., 1984) . acanthosis with hyperkeratosis and parakeratosis has been reported in the oesophagus but not stomach of rats treated for up to 18 months with mesuprine hydrochloride, a î²-adrenergic receptor stimulator (nelson et al., 1972) . as part of its effects on the gastrointestinal tract, the oesophagus in rats and primates has been reported to develop uniform hyperplasia of the squamous epithelium following infusion of recombinant epidermal growth factor (breider et al., 1996; reindel et al., 1996; vinter-jensen, 1999) . in the rat, mouse and hamster the forestomach occupies about two-thirds of the proximal stomach area and is lined by cornified stratified squamous epithelium. the limiting ridge is a distinct elevated mucosal fold at the junction between the forestomach and the mucosa of the glandular part of the stomach. as humans lack a forestomach, the relevance of changes produced by drugs and chemicals in the rodent forestomach is disputed. studies in rats in which the forestomach has been removed have suggested that the forestomach acts as a storage organ releasing relatively undigested food into glandular stomach in response to energy demand (gã¤rtner and pfaff, 1979) . hence, the forestomach mucosa may be exposed to xenobiotics mixed in undigested food for far longer periods than elsewhere in the gastrointestinal tract. the interpretation of forestomach changes should take into account physiological factors, residence time and exposure differences to drugs between the rodent forestomach and human oesophagus. however the squamous mucosa lining the oesophagus in species without a forestomach may react to xenobiotics in a similar way to the forestomach mucosa of rodents if equivalent exposure levels are attained. inflammation and ulceration of the forestomach mucosa are some of the commonest spontaneous gastrointestinal lesions in laboratory rats, mice and hamsters. the prevalence of these gastric lesions varies between species, strains of laboratory rodents as well as between different laboratories. the precise causes of forestomach ulceration remain unclear although a variety of factors have been associated with its development including advanced age, infection, parasitism, diet, feeding regimens and stress. in rats, conflict-induced ulceration occurs in the forestomach and there is an age-related susceptibility, older rats developing more ulcers than younger rats (sawrey and sawrey, 1966) . in rats and mice dying of spontaneous disease, ulceration of the forestomach is also quite frequently observed. protein restriction or starvation has also been shown to produce forestomach ulceration in rats . ulceration of the forestomach in rodent toxicology studies may be incidental, particularly if the lesions are few and show no clear relationship to dose. if limited to high dose groups, ulceration may be a result of non-specific toxicity and stress-related. however, administered chemicals may have direct local effects of sufficient severity to cause focal damage to the forestomach mucosa. histological features of ulcers and inflammatory lesions of the forestomach are similar in rats, mice and hamsters. in mild cases, a scattering of acute inflammatory cells is seen in the intact squamous mucosa. ulcers can be single or multiple and are characterised by loss of squamous epithelium with a variable accumulation of neutrophils, mononuclear cells, cellular debris, fibrin and hair fragments in the ulcer crater. the inflammatory process may extend deeply into the stomach wall and be associated with intramural inflammation, oedema, endarteritis and fibrosis. haemosiderin pigment is also found in the ulcer margins. profuse haemorrhage may follow erosion of large blood vessels and complete perforation of the stomach wall with peritoneal involvement also occurs (greaves and faccini, 1992) . in long-standing cases of ulceration, hyperplasia of the adjoining squamous epithelium occurs, characterised by irregular acanthosis and down-growths of squamous epithelium into the submucosa . xenobiotics may produce inflammatory changes in the forestomach mucosa following initial dosing but subsequently, repair occurs even though treatment continues. an example of this phenomenon is illustrated by butylated hydroxyanisole. after 1 week of administration of this agent in a 2% mixture in diet to rats, a vesicular inflammatory reaction characterised histologically by the presence of subepithelial vesicles containing inflammatory cells and exudate was seen (altmann et al., 1985) . after further treatment, only hyperplasia of the squamous epithelium was evident, presumably as an adaptive response to the effects of the continued insult. hyperkeratosis associated with hyperplasia of the squamous epithelium is seen sporadically in untreated aged rodents. these changes may be localised to the margins of chronic forestomach ulcers or they can be associated with diffuse inflammation of the mucosa. occasionally, the forestomach mucosa of untreated, aged rodents exhibits hyperkeratosis with hyperplasia without inflammation (fig. 45 ). such changes may be diffuse or focal, but they are often localised to the zone adjoining the glandular stomach mucosa. there may be evidence of basal cell proliferation and downgrowth of the epithelium into the underlying stroma. dietary factors also influence the thickness of the forestomach mucosa. vitamin a deficiency, known to produce squamous metaplasia in glandular tissues may produce forestomach hyperplasia and hyperkeratosis in rats. when spf fischer 344 rats were maintained in a vitamin a deficient state for over 3 months, hyperplasia with hyperkeratosis, not unlike that produced by known carcinogens was reported (klein-szanto et al., 1982) . administration of a wide range of both industrial chemicals, therapeutic agents including both genotoxic and non-genotoxic carcinogens produces hyperkeratosis and hyperplasia of the forestomach epithelium which may be followed by preneoplastic lesions and squamous carcinoma (see below). histologically, the changes are characterised by hyperkeratosis, parakeratosis with varying degrees of acanthosis and papillomatosis (greaves and faccini, 1992) . the changes can be florid and it may be difficult to make a clear distinction between severe hyperplasia and neoplasia. nevertheless, it has been shown that the florid hyperplasia of the forestomach epithelium without evidence of cellular atypia can be completely reversible following the withdrawal of an inciting stimulus, ethyl acrylate (ghanayem et al., 1991) . hence, a critical feature may be the presence of cellular atypia in view of its association with agents with potent (genotoxic) carcinogenic activity. neoplasms arising in the forestomach of rodents are usually squamous carcinomas although basaloid features are also seen (fukushima and leiniger and jokinen, 1994; mohr, 1997; tatematsu, 1997) . squamous carcinomas, as at other sites, show variable differentiation being composed of proliferating squamous epithelium with moderate to marked cellular atypia, pleomorphism and mitotic activity with clear evidence of invasion into the muscularis. although they are relatively uncommon spontaneous lesions in aged rodents, they can be induced in rodents by administration of nitroso compounds (tatematsu 1997) as well as a range of non-genotoxic agents (see below). some authors report basal cell carcinoma when basaloid features are pronounced (tatematsu, 1997) . a wide range of agents is capable of producing squamous hyperplasia of the rodent forestomach and a number of these also induce squamous carcinomas. in 1986 kroes and wester reviewed over 60 genotoxic and non-genotoxic compounds that were reported to produce hyperplasia and carcinoma in the forestomach of rats, mice or hamsters. a well-studied example is butylated hydroxyanisole (bha) an important food antioxidant (reviewed by whysner and williams, 1996) . structurally related phenols and acids produce similar changes (rodrigues et al., 1986) . ethyl acrylate, used in the production of materials for dental and medical devices is also capable of inducing marked squamous hyperplasia, papillomas and carcinomas after long-term treatment of f344 rats and b6c3f1 mice (ntp, 1986; ghanayem et al., 1991) . sk&f 93479, an experimental histamine h 2 receptor antagonist produced atypical forestomach hyperplasia in rats following administration by gavage for 1 year by a mechanism which appeared unrelated to the inhibition of the h 2 receptor (betton and salmon, 1984) . other therapeutic agents associated with squamous hyperplasia and neoplasia include the 3-hydroxy-3-methylglutaryl coenzyme a (hmg-coa) reductase inhibitors (kloss et al., 1991; bueld et al., 1996; akiba et al., 1998; physicians' desk reference, 1999) and aristolochic acid (gã¶ggelmann et al., 1982; schmeiser et al., 1988) . some cytoprotective prostaglandins appear capable of inducing hyperkeratosis and hyperplasia without neoplasia presumably through a mechanism related to their pharmacological activity (levin, 1988) . this occurs in rats treated with misprostol, a synthetic prostaglandin e1 methyl ester analogue with gastric anti-secretory and anti-ulcer activity (kotsonis et al., 1985) , cl115,574, a synthetic analogue of prostaglandin e1 type (kramer et al., 1985) and 16,16-dimethyl prostaglandin e2 (reinhart et al., 1983) . even the extensively used antibiotic, ampicillin has been associated inflammation, ulceration with acanthosis and hyperkeratosis in mice but not rats treated for 2 years (national toxicology program technical report, 1987) . sodium saccharin is also reported to produce hyperplasia without neoplasia of the forestomach in f344 rats (hibino et al., 1985) . among its wide range of pharmacological effects on the gastrointestinal tract, the forestomach has also responds to recombinant epidermal growth factor when infused into rats (breider et al., 1996; vinter-jensen, 1999) . histological examination showed hyperkeratosis and hyperplasia of the squamous epithelium. the large body of studies performed with butylated hydroxyanisole illustrates the various factors that can influence the development of treatment-induced hyperplasia and neoplasia of the rodent forestomach and subsequent assessment of human risk. butylated hydroxyanisole possesses little or no mutagenic activity in vitro but when administered to rats for 2 years as a 2% mixture in the diet, it produced squamous hyperplasia, squamous papillomas and squamous carcinomas of the forestomach. at 0.5% in the diet butylated hydroxyanisole induced only hyperplasia . it also produces proliferative lesions in the forestomach of both mouse and hamster (ito et al., 1986) . studies in which butylated hydroxyanisole was fed in the diet to rats for shorter periods have shown that squamous epithelial hyperplasia occurs after only 1 week of treatment preferentially over the lesser curvature, the site at which carcinomas developed in the 2-year studies (altmann et al., 1985) . after 13 weeks' treatment, mucosal hyperplasia characterised by pronounced hyperkeratosis, parakeratosis and acanthosis most pronounced over the lesser curvature, was present in rats given 2% butylated hydroxyanisole in diet but not in rats given 0.5, 0.25 and 0.1% mixtures . abundant mitoses were found in the basal cells layers and tritiated-thymidine labelling confirmed that the changes were accompanied by a high rate of cell proliferation. following cessation of administration of butylated hydroxyanisole after 13 weeks, the tritiatedthymidine labelling index rapidly reverted to control levels within about 1 week although hyperplasia took longer to regress. nearly complete regression of the hyperplasia occurred after about 9 weeks of normal diet . the distribution of squamous hyperplasia induced in the rodent stomach by butylated hydroxyanisole is influenced by the mode of administration. whereas following feeding of rats with butylated hydroxyanisole mixed in the diet lesions tended to be located near the limiting ridge, altmann et al. (1985) showed that gavage of butylated hydroxyanisole in corn oil produced similar changes at the apex of the forestomach. it was suggested that this difference was due to incomplete mixing of butylated hydroxyanisole in the stomach lumen when given by gavage and prolonged contact of the gavage mixture with the upper segment of the forestomach (altmann et al., 1985) . more recently, it was shown that fischer 344, shr, lewis and sprague-dawley rats differ in their response to the hyperplastic and carcinogenic effects of 2% butylated hydroxyanisole in pelleted diet. the most sensitive was the shr strain followed by the f334 rats and the differences correlated with the cytotoxic effects of butylated hydroxyanisole in the different strains (tamano et al., 1998) . it was suggested that the presence of vascular damage in the stomachs of the shr rats might have contributed to the response to cytotoxicity and subsequent carcinogenicity. residence time of administered compounds in the forestomach may influence the development of lesions. although it has been demonstrated that butylated hydroxyanisole does not produce hyperplasia in the oesophagus of animals without a forestomach, have shown that high-doses given to primates are capable of producing an increase in mitotic activity in the lower end of the oesophagus similar to that occurring at equivalent exposure levels in the rat. the implication is that these interspecies differences may simply be a question of differences in exposure of the squamous mucosa to compound. this underlines the fact that mechanisms of action and exposure levels of xenobiotics attained in the gastrointestinal tract of rodent and non-rodent species as well as of man need to be carefully assessed when hyperplastic changes are induced in the forestomach mucosa of rodents. such information can be helpful in facilitating regulatory decisions in this area (moch, 1988) . on balance, the evidence suggests that the tumour development by butylated hydroxyanisole in rodents represents an epigenetic phenomenon related to largely reversible cytotoxicity and increased cell proliferation (whysner and williams, 1996) . in view of the low levels of exposure to butylated hydroxyanisole that occurs with the usual use of this agent, carcinogenic hazard for the human stomach is therefore probably very small. similar phenomena have also been reported in studies of phenols and acids that are structurally related to butylated hydroxyanisole (rodrigues et al., 1986) . these agents include n-butyl and n-propyl-4-hydroxybenzoic acid esters, propionic acid and 4-methoxyphenol. however, these studies suggested that certain areas of the forestomach epithelium react differently to structurally related chemicals, possibly due to the variable levels of activating enzymes within different zones of the forestomach epithelium. co-administration of acetylsalicylic acid was shown to abrogate some of these effects, suggesting that prostaglandin synthetase may be involved in the hyperplastic response (rodrigues et al., 1986) . a number of hmg-coa reductase inhibitors with different chemical structures including marketed products such as lovastatin, simvastatin and fluvastatin are also associated with the development of squamous hyperplasia of the rodent forestomach. the hyperplasia is time and dose dependent and may be associated with oedema and some inflammation of the submucosa. some, but not all of these agents are also capable of producing squamous neoplasia of the forestomach mucosa of rats, or mice or both after long-term treatment (kloss et al., 1991; bueld et al., 1996; akiba et al., 1998; physicians' desk reference, 1999) . the mechanism of action remains unclear although the degree of hyperplasia seems related to pharmacological potency. their carcinogenic potential in rodent bioassays does not seem to relate to the degree of hyperplasia in shortterm studies. moreover, the development of hyperplasia depends on local high concentrations of drug because when administered by non-oral routes, hyperplasia does not occur (kloss et al., 1991) . as most of these drugs are non-mutagenic, these findings are presumably also epigenetic in origin and possess relatively little risk for humans when given in the usual therapeutic doses. a contrasting example is provided by aristolochic acid, a nitrophenanthrene derivative of the ancient medicinal plant aristolochia clematis which was used as an anti-inflammatory component in a number of medicinal preparations in germany until 1982 (gã¶ggelmann et al., 1982; schmeiser et al., 1988) . aristolochic acid is a direct acting mutagen in salmonella typhimurium. when fed to rats at doses of 1.0 and 10 mg/kg/day, aristolochic acid produced severe papillomatosis of the entire forestomach within a period of 3 months. this was characterised histologically by the presence of branched squamous papillomas up to 6 mm high with focal dysplastic features. invasive squamous carcinomas with metastases were found subsequently, 3 or 6 months later without further treatment (mengs et al., 1982) . even at a low dose of 0.1 mg/kg/day papillomas and squamous carcinomas developed 9 months after a 3-month period of treatment. quite clearly the complexity of the hyperplastic response of the rodent stomach to xenobiotics, the association of hyperplasia induced by non-mutagenic compounds with the development of forestomach carcinomas, and the similarity of response in the forestomach to that of the oesophagus, dictates the need for a careful analysis of hyperplasia induced by novel drugs in the forestomach. the prelude to this assessment is careful histopathological characterisation of the changes. unlike the mouth and oesophagus through which tablets, capsules, gavage fluids and drug/diet mixtures pass relatively rapidly, the human stomach mucosa remains in contact with high local concentrations of administered compounds for much longer periods of time. administration of compounds in liquid or solid form, particle size, fasting and feeding all affect the gastric motility pattern. in the fasted state there is a cyclical pattern of motility consisting of three main phases. the first is a quiescent phase, followed by a phase of irregular contractions that increase in amplitude and frequency to reach a maximum in a third phase. feeding results in the replacement of this cyclic pattern by regular tonic contractions that move food towards the antrum and mix it with gastric secretions. these patterns have been well studied in both dog and man and appear to be qualitatively similar in the two species (sarna, 1985) . these motility patterns may have an impact on the length of time drugs remain in contact with stomach mucosa. for instance, the residence time of large non-disintegrating capsules or tablets administered in the fasting state is more dependent on the frequency of powerful phase iii contraction than if drugs are given as fluids or mixed with diet. for dosage forms released in the stomach, gastric residence time will influence drug supply to the main absorptive surfaces in the small intestine, which in turn may affect drug absorption (dressman, 1986) . gastric acid is also important in making ingested salts soluble. although the presence of food in the stomach is a stimulus of acid production, the ph in the forestomach of rats is highest in full stomachs and lowest when empty, presumably as a consequence of the buffering action of food (ward and coates, 1987) . the glandular stomach is conveniently divided into the fundus characterised by mucosal folds or rugae and the smoother antrum, which opens into the pylorus and duodenum. in species devoid of a forestomach, the proximal stomach mucosa or cardia is also lined by glandular mucosa. the glandular mucosa is covered by surface epithelium of regular columnar cells that extends downwards to form small gastric pits or foveolae. the gastric glands are simple tubular structures usually considered to comprise three segments. the base is the deepest part, the neck the mid-region, and the most superficial is the isthmus, continuous with the gastric pit. the upper part of the gastric gland contains mucous neck cells. small cuboidal chief or zymogenic cells, which secrete pepsinogen and stain blue or purple in haematoxylin and eosin sections, are located in deeper parts of the gland. the eosinophilic-staining parietal (oxyntic) cells, which produce hydrochloric acid, are distributed more randomly throughout the gastric glands. parietal cells can also be visualised by immunocytochemical staining with antibodies directed at h + k + -atpase (canfield et al., 1996) . the gastric glands situated near the limiting ridge in rodents, show a modified structure. in species not endowed with a forestomach, the mucosa near the cardia is composed of simplified branched glands lined by columnar epithelium. the antral mucosa is covered by a surface epithelium with gastric pits similar to that of the fundus but mucous secreting columnar glands line the glands. the stomach mucosa is richly endowed with endocrine cells, not all of which have been well characterised. enterochromaffin cells are quite numerous in the basal parts of the gastric glands of the fundus, particularly in the rat . they are generally argyrophilic, staining with silver staining techniques such as that of grimelius (grimelius, 1968; grimelius and willander, 1980 ) that utilise exogenous reducing agents. these cells contain histamine and histamine-related enzymes such as histidine decarboxylase in the rat and other species . endocrine cells which are argentaffin in type stain with silver stains such as that of masson (1914) because of the presence of endogenous reducing substances including 5-hydroxytryptamine and catechol-amines are also reported in the mucosa of the fundus of some species including man but apparently not in the rat . enterochromaffin cells are characterised ultrastructurally by the presence of numerous rounded or oval, vesicular, electron-lucent granules frequently containing a small eccentric electron dense core. gastric enterochromaffin cells can also be stained by immunocytochemical techniques using antisera to histamine and histidine decarboxylase as well as to non-specific enolase and chromogranin a betton et al., 1988) . immunocytochemical study of the rat fundus using a battery of antisera to a variety of gastrointestinal peptides has shown somatostatin containing (d) cells and glucagon staining cells but no cells with gastrin (g cells) or serotonin reactivity (bishop et al., 1986) . gastrin or g cells possess apical processes reaching the stomach lumen believed to be important in stimulation of gastrin release as a result in increases in antral lumen ph or the presence of amino acids or peptides . glucagon and serotonin containing endocrine cells have also been located in the rat antral mucosa (bishop et al., 1986) . increased gastric acid secretion is initiated by activation of central vagal efferent pathways but acid secretion is maintained by both neural and endocrine reflexes activated by the presence of food in the stomach. gastrin secreted from the g cells of antrum is the main stimulant of acid secretion. somatostatin is secreted from antral d cells when the luminal ph falls to below 3.5 to act by a paracrine mechanism to suppress g cell function thus forming a negative feedback loop (dockray, 1999) . the two main endocrine cell types from the body mucosa integrate neurohumoral stimuli rather than respond to luminal chemicals. although gastrin is capable of stimulating parietal cells directly, it has an even greater effect through stimulation of enterochromaffin cells to release histamine, a potent paracrine stimulator of parietal cells (hinkle and samuelson, 1999) . gastrin stimulates release of histamine from enterochromaffin cells of the body mucosa, which increases acid secretion through activation of parietal cell histamine-h 2 receptors. both parietal and enterochromaffin cells are inhibited by somatostatin released from the d cells of the body mucosa in response to a variety of neurohumoral stimuli such as noradrenaline, vasoactive intestinal peptide, calcitonin gene-related peptide, and cholecystokinin. it should also be noted that gastrin has a role in kinetics and differentiation of both parietal and enterochromaffin cells (dockray, 1999) . gastrin acts at the gastrin/cholecystokin-b receptor that is expressed by gastric epithelial cells and by neurones in the central nervous system (kopin et al., 1992; wank, 1995) . the gastrin receptor is simply the cholecystokinin-b receptor located in the stomach. the other cholecystokinin receptor, cholecystokin-a has high affinity for cholecystokinin. stimulation of this receptor in the stomach mediates secretion of pepsin from gastric chief cells and release of somatostatin from d cells resulting in inhibition of acid secretion. in the central nervous system cholecystokinin and its receptors contribute to the regulation of satiety, anxiety, analgesia and dopamine-related behaviour (wank, 1995) . finally, it is worth recording that both progesterone and oestrogen receptors have been identified in both normal and pathological gastric tissues of humans (wu et al., 1992) . generative cells in the gastric mucosa as shown by uptake of tritiated thymidine for dna synthesis are distributed principally in the isthmus (inokuchi et al., 1983) . tracing of cells using thymidine labelling have shown that most of the cells in the generative zone migrate in a successive manner to the mucosal surface to form columnar epithelium. the life span of surface epithelium in the stomach of rats, mice and hamsters has been calculated to be about 3-4 days. studies of cell cycle and dna synthesis time in the proliferative zones in the stomach of rat, hamster and man have suggested that the generative cells in the isthmus undergo mitoses at about 30-hour intervals in rodents and 40-hour intervals in man (inokuchi et al., 1983) . although this process of migration from the proliferating cell zone of the isthmus renews surface epithelial cells rapidly, cell migration to the lower parts of the gastric glands is much slower and more complex. detailed studies have show that undifferentiated cells in the region of the isthmus represent a common source for surface mucous cells and mucous neck cells (karam and leblond, 1995) . electron microscopic and ultrastructural cytochemistry has suggested note: saccharide binding specifications are much more complex than the inhibition by simpler sugars outlined above suggests. see review by nicholson (1974) . *source: nicholson, 1974; goldstein and hayes, 1978; schulte and spicer, 1983; giannasca et al., 1994. that chief cells in the adult rat stomach develop from undifferentiated stem cells in the isthmus (suzuki et al., 1983) . graft experiments in mice have also suggested that immature cells of the isthmus differentiate into chief cells as well as parietal cells (matsuyama and suzuki, 1970) . studies in transgenic mice have shown that mature parietal cells influence the fate of other gastric epithelial cells because targeted degeneration of parietal cells is associated with loss of chief cells suggesting interactions between these cell populations in determining their differentiation (li et al., 1996; canfield et al., 1996) . gastrin is also an important regulator of parietal cell and enterochromaffin differentiation and number (dockray, 1999; montgomery et al., 1999) . labelling experiments in the hamster stomach have shown that both chief and parietal cells possess a similar but quite long life span of about 200 days (hattori, 1974; hattori and fujita, 1976) . it has been suggested that the relative distribution of chief and parietal cells in the gastric gland represents an expression of their different migration patterns downwards from the proliferative zones in the isthmus. this type of migration pattern in which cells are able to overtake each other has been termed a 'stochastic flow system' (inokuchi et al., 1983) . the origin and kinetics of endocrine cells of the stomach has also been the subject of debate but the available morphological, cytochemical and kinetic evidence suggests that the majority of these cells develop from the same stem cells as the other non-endocrine cells of the gastric mucosa, although self replication also occurs (matsuyama and suzuki, 1970; inokuchi et al., 1983; solcia et al., 1986) . much of our knowledge about mucins produced by the epithelial cells lining the gastrointestinal tract has been obtained using histochemical techniques and these approaches continue to be helpful in the understanding of spontaneous and drug-induced gastrointestinal disease (sheahan and jarvis, 1978; filipe, 1979; tsiftsis et al., 1980; jass and roberton, 1994) . for these reasons, mucin histochemical techniques represent useful tools for the characterisation and elucidation of experimentally or drug-induced changes in the glandular mucosa of gastrointestinal tract. techniques commonly employed are presented in table 2 . the physiochemical properties of gastrointestinal mucins are dependent on their glycoprotein constituents. these glycoproteins are high molecular weight compounds with large numbers of sugar chains attached to a polypeptide backbone by o-glycosidic linkages between n-acetylgalactosamine and serine or threonine (berger et al., 1982) . the principle monosaccharides present are fucose, galactose, n-acetylgalactosamine, n-acetylglucosamine and sialic acid. traces of mannose may be present and ester sulphate residues are common (filipe, 1979) . due to this extensive glycosylation, mucins have a filamentous conformation, which is often negatively charged. this is believed to be important in forming a protective barrier to the cell. however, this property is a two-edged sword because when opposing cells have specific receptors for mucins, adhesion may become the predominant factor (van klinken et al., 1995) . although mucins are important in the gastrointestinal tract, it should be remembered that other products secreted by goblet cells might be important in mucosal defence. it has been recently recognised that trefoil proteins, a family of small proteins secreted by goblet cells and present on the mucosal cell surface, can also protect against a variety of deleterious agents, including bacteria, toxins and drugs (podolsky, 1999) . there are considerable regional variations in glycoprotein constituents in the gastrointestinal tract and these differences are probably related to physiological and functional factors. furthermore, synthesis and secretion of glycoproteins alter with changes in cell differentiation. alterations also occur in mucins in various inflammatory and neoplastic disease states as well as following administration of certain drugs and chemicals (ishihara et al., 1984) . terminal sugars or sugar sequences can be demonstrated histochemically by the use of labelled lectins, mostly plant proteins which combine non-enzymatically with particular sugar molecules, see table 1 (goldstein and hayes, 1978; debray et al., 1981; rudiger, 1982) . magenta: all mucosubstances containing hexoses schiff d-pas (pearse, 1968) and deoxyhexoses with vicinal glycol groups. some non-sulphated acid mucosubstances. neutral mucosubstances. periodate-borohydride/ magenta: pas activity following periodate borosaponification/pas, hydride/potassium hydroxide indicates pb/koh/pas presence of o-acylated sialic acids. (reid et al., 1973; periodate borohydride reduces periodate culling et al., 1974; generated aldehydes. potassium hydroxide removes o-acylesters from potential vicinoldiols and sialic residues linked glycosidically to a potential vicinoldiol. alcian blue ph 2.5 basophilia: weakly sulphated mucins. carboxyl (pearse, 1968) groups of sialomucins alcian blue ph 1.0 basophilia: sulphated mucins (lev & spicer, 1964) alcian blue ph 2.5 -magenta: neutral mucins periodic acid schiff, ab/pas basophilia: acid mucins (mowry & morard, 1957) purple-blue: neutral and periodate reactive acid mucins high iron-diamine, brown-black: sulphated mucins hid (spicer, 1965) unstained: sialomucins high iron-diamine-alcian brown-black: sulphated mucins blue ph 2.5, hid/ab basophilia: non-sulphated acid mucins (sialomucins) (spicer, 1965) source: adapted from filipe, 1979. when gastrointestinal mucins were studied in several species using histochemical techniques under uniform conditions, species differences were most obvious in the stomach and duodenum (sheahan and jarvis, 1976) . neutral mucins generally predominate in the stomach, contrasting with acid mucins in the small intestine, and sulphated mucins in the colon. in the stomach neutral mucins staining purple with the pas/alcian blue stain, predominate in the surface and foveolar mucosa, whereas mucous neck cells and antral glands contain acidic mucins that stain blue with pas/alcian blue procedure. sulphated mucins, as shown by the high iron diamine technique (hid) are also found in the deep glandular mucosa of the antrum in rat, mouse and man (filipe, 1979; jass, 1980; greaves and boiziau, 1984) . extremely heterologous staining patterns are seen in the gastric mucosa with labelled lectins, each lectin staining quite different cell populations. there are considerable interspecies differences in staining patterns with the same lectins (kuhlmann et al., 1983; suganuma et al., 1984) . the so-called paradoxical concanavalin a stain, in which conjugated concanavalin a is used to label mucins before and after periodate oxidation, has also been used to classify the alterations in mucins in proliferative and neoplastic conditions of the rat stomach mucosa (kobayasi et al., 1991; tatematsu 1997 ). although gastric erosions and ulcers in the glandular mucosa occur quite commonly in laboratory animals in toxicity studies, it is often difficult to determine whether such lesions in treated animals indicate a real ulcerogenic risk for the test compound. there is little that is histologically specific to drug-induced ulceration of the gastric glandular mucosa. mucosal haemorrhage, depletion of mucin, erosions and ulcers with or without inflammation may all be found. erosions represent mucosal breaks superficial to the muscularis mucosa. ulcers are lesions that extend through the muscularis mucosa. whilst the histopathological features of gastric erosions and ulcers are themselves relatively non-specific, it is important to look for any associated pathology in the stomach such as mucus depletion, epithelial hyperplasia or dysplasia, intestinal metaplasia and vascular lesions (see below). in humans, biopsy data suggests that drug-induced ulceration is characteristically devoid of an inflammatory component, but the most usual histological appearances are those of underlying gastric pathology (riddell, 1982) . formation of gastric and duodenal ulcers is dependent on the presence of both acid and peptic activity in gastric juice because acid without pepsin appears to have little digestive power. important predisposing factors in human patients with peptic ulceration include helicobacter pylori (campylobacter pylori) infection of the antrum, cigarette smoking and ingestion of non-steroidal anti-inflammatory drugs (soll, 1990) . helicobacter pylori is believed to infect over half the human population and its presence in the gastric mucosa is associated with chronic atrophic gastritis and peptic ulceration (cover and blaser, 1996) . it is a microaerophilic, gram-negative organism that possesses potent urease activity crucial for its survival at acidic ph. genome sequence analysis has shown that helicobacter pylori has well developed sequences for motility, scavenging iron and dna restriction and modification systems used by bacteria to degrade foreign dna. the link between helicobacter pylori infection and peptic ulceration is related to increases in gastrin release, perhaps through bacterial products or cytokines released from activated lymphocytes (richter-dahlfors et al., 1998) . although helicobacter pylori can infect other species, apart from non-human primates, the usual laboratory animal models do not appear to develop the inflammatory disease seen in humans (nedrud, 1999) . erosions and ulcers also develop following stress, reflux of intestinal contents and bile, changes in acid secretion and hypoxia, all of which may develop under the conditions occurring in high-dose toxicity studies. the requirement to give the test compound in high doses may also dictate the need to administer exceedingly high concentrations of test agent. this may produce damaging high local concentrations on the mucosa not relevant to therapeutic doses used in clinical practice. it has been demonstrated that hyperosmolar solutions of quite innocuous substances such as glucose can cause haemorrhage, erosions and ulcers of the rat gastric mucosa (puurunen et al., 1980) . the well-known association of gastric erosions and haemorrhage with uraemia may also be manifest following administration of high doses of drugs such as diuretics which severely derange fluid and electrolyte balance (garthoff et al., 1982) . stress ulceration may be linked to temporary ischaemia of the mucosa (dubois, 1987) . synergism between the ulcerogenic action of drugs and stress is a well-described phenomenon (rainsford, 1975; beattie, 1977) . protein depletion and starvation is also capable of inducing gastric ulceration in rats . although gastric pathology represents the largest cause of morbidity and mortality in man following therapy with non-steroidal inflammatory agents (fowler, 1987) , the reasons for this are probably multifactorial. the acidic properties of some of these drugs may cause direct local damage of gastric epithelial cells, demonstrable by the fact that appropriate formulation can reduce gastric toxicity of these agents in man (brors, 1987) . anti-inflammatory agents are capable of decreasing synthesis of glycoproteins and this may adversely influence protective mucus production of gastric mucosa (azuumi et al., 1980; ishihara et al., 1984) . it has also been suggested that non-steroidal anti-inflammatory agents cause cellular damage to the gastric mucosa by back-diffusion of gastric acid into mucosal tissues (davenport, 1964) or by causing damage to the gastric capillary bed with subsequent mucosal infarction (robins, 1980) . the theory that has gained widespread acceptance is that the ulcerogenic potential of non-steroidal anti-inflammatory drugs is related to their pharma-cological activity. vane (1971) proposed that the ulcerogenic potential of these agents was largely a result of their ability to inhibit prostaglandin synthetase, thereby reducing the protective effects of prostaglandins. pharmacokinetic factors may also be important. lipid solubility in the low ph environment of the stomach may influence local penetration into the mucosa (mccormack and brune, 1987) . moreover, it has been proposed that certain anti-inflammatory agents may possess lesser ulcerogenic potential in man because their inhibition of prostaglandin production is more limited to sites of inflammation, sparing gastric mucosa (whittle et al., 1980; whittle and vane, 1984) . it has also been demonstrated that factors altering the enterohepatic circulation of drugs can influence the expression of gastric damage (overvold et al., 1987) . comparative studies of the ulcerogenic activity of indomethacin in beagle dogs and domestic pigs has suggested that the dog may be an excessively sensitive species as a result of extensive enterohepatic circulation of indomethacin in this species (hanhijarvi et al., 1986) . prediction of ulcerogenic potential for man based on data from animal models is clouded by the lack of good comparative data on the relative ulcerogenic potential of non-steroidal anti-inflammatory agents in man because of extensive differences in side effect reporting (fowler, 1987) . moreover proper comparisons in man require not only equivalent therapeutic doses but also comparable dosage forms (brors, 1987) . in laboratory animals, a variety of different patterns of drug-induced gastric damage have been described. the study by shriver et al. (1975) in which a wide variety of different anti-inflammatory drugs were administered to fasted sprague-dawley rats under identical conditions, suggested the drugs could be divided into three groups based on their profiles of gastrointestinal toxicity. immunological agents such as azathiaprine, cyclophosphamide, methotrexate and d-penicillamine produced gastric mucosal haemorrhage whereas aspirin and related agents produced gastric mucosal haemorrhage and ulcers. the powerful nonsteroidal anti-inflammatory drugs indomethacin and phenybutazone produced gastric mucosal erosion and ulcers as well as small intestinal damage. comparative single oral dose studies of several different non-steroidal antiinflammatory agents at three different dose levels by suwa et al. (1987) in the rat using histology and measurement of faecal blood loss with 51 cr-labelled blood cells have also shown that different patterns of ulceration can be produced by different agents when administered under identical conditions. single oral doses of some non-steroidal anti-inflammatory drugs including aspirin produced widespread superficial damage and desquamation of gastric epithelium with little or no inflammation at 6 hours following dosing which completely healed 2 weeks later. this damage was associated with transient faecal blood loss. by contrast, indomethacin and ibuprofen produced both gastric damage and circumscribed, penetrating ulcers along the mesenteric border of the jejunum and ileum. furthermore, ulcers were still present after 2 weeks and were associated with prolonged or biphasic blood loss (see small intestine). in addition feeding conditions can influence the distribution of erosions and ulcers in laboratory animals. in fasted rats, erosions due to indomethacin treatment are found in the body of the stomach whilst in conventionally fed rats they are most prominent in the small intestine. detailed studies by satoh et al. (1981) showed that rats fed for 1 hour after a 24 hour-fast and given a single dose of indomethacin within 2 hours of re-feeding developed erosions and ulcers in the antrum primarily along the lesser curvature. indomethacin given to fasted rats produced erosions in the body mucosa. a further factor that needs to be kept in mind is that chronic administration of ulcerogenic compounds may produce quite different pathological appearances to those found following single dose administration. administration of aspirin to rats for 4 weeks has been shown to stimulate epithelial proliferation of the gastric body but not antral mucosa, possibly by an effect on cyclic adenosine 3',5' monophosphate (cyclic amp) or though increasing the rate of epithelial exfoliation (eastwood and quimby, 1982) . such a response may be the basis for increased resistance of the gastric mucosa to the chronic affects of these agents. it also may explain the tendency for ulcers to occur in the antrum following chronic administration of aspirin-like drugs as the proliferative response and presumably the adaptive potential appears less in this part of the gastric mucosa. both interspecies variations and strain differences have been reported in the response to ulcerogenic compounds. rainsford et al. (1982) showed that extravasation of red blood cells and greater vascular damage was observed in rats treated with aspirin or benoxprofen than in pigs given similar doses. sprague-dawley rats appear less susceptible to the ulcerogenic effects of cold-restraint stress than wistar rats (goldenberg, 1973) . diuretics and some angiotensin converting enzyme (ace) inhibitors and angiotensin ii antagonists have been associated with the development of gastric erosions and ulceration when administered in high doses to laboratory animals ( fig. 46 ) (imai et al., 1981; garthoff et al., 1982) . however, these effects appear related to the severe electrolyte disturbances produced by excessive doses of these drugs. this is perhaps analogous to the well-known association of gastrointestinal tract erosion and haemorrhage with uraemia. dogs appear to have a particular predisposition to this effect where it may be associated with deposition of basophilic ground substance and mineral in connective tissues and blood vessels in the mucosa (barker and van dreumel, 1985) . although inflammatory conditions due to microorganisms are generally uncommon in the stomach, gastritis is reported in laboratory rhesus monkeys in association with the presence of helicobacter organisms (reed and berridge, 1988) . as in the analogous condition in man, the stomach of affected animals shows an infiltration of the central mucosa by small lymphocytes and plasma cells, associated with reactive or atrophic changes in the mucosa and the presence of small curved bacteria in glands, visualised best with the warthin-starry stain. infiltration of the stomach by lymphocytes in rats treated with human recombinant interleukin-2 without ulceration was reported as part of a multisystem involvement induced by this agent (anderson and hayes, 1989 ). decrease in gastric mucus secretion may accompany both spontaneous inflammatory conditions and drug-induced lesions in the stomach of man and experimental animals. mucus depletion is characterised histologically by the presence of an intact epithelial layer in which cells show loss of the normal clear cytoplasm replete with mucous substances by more basophilic cells that contain little or no mucin. qualitative changes in mucus composition can also accompany mucus depletion. gastric epithelium in man may show decreases in sulphated mucosubstances following stress, high alcohol consumption or after aspirin administration (filipe, 1979) . similar changes occur in laboratory animals subjected to ulcerogenic regimens. stress ulceration in the rat is accompanied by decreased sulphation of gastric glycoproteins, presumably an expression of the changes in gastric cellular activity accompanying stress (lambert et al., 1969) . administration of aspirin and other anti-inflammatory agents including adrenocortical steroids to laboratory animals also reduces the content of sulphomucins in the gastric mucosa, probably by reducing their synthesis (denko, 1958; gerard, 1965; ishihara et al., 1984) . rather surprisingly, administration of histamine h 2 -receptor antagonists and fig. 46 . section from the glandular stomach from a rat treated with ah high dose of an angiotensin ii antagonist that shows superficial degeneration and ulceration (erosion) of the mucosa. (he, ã�30.) proton pump inhibitors associated with reduction of gastric acid output and increases in gastrin secretion have also been associated with alterations in gastric mucus. administration of omeprazole or famotidine to rats for 4 weeks was shown to inhibit prostaglandins pge 2 as well as the synthesis of both total and sulphated glycoprotein synthesis along with histochemical evidence of reduction in pas staining of the surface mucus (yoshimura et al., 1996) . although the mechanism for this change is unclear, the reduction in mucus, particularly sulphated mucus that is believed to be particularly resistant to peptic digestion, may have implications for mucosal defence. intestinal metaplasia of the stomach is characterised by the presence of differentiated epithelium, which resembles small intestine on the basis of light microscopic and ultrastructural morphology, mucin patterns and enzyme histochemistry (morson, 1955; planteydt and willighagen, 1960; lev, 1966; goldman and ming, 1968; . it develops in man in gastric mucosa altered by chronic atrophic gastritis and its significance is due to the fact that a link exists between intestinal metaplasia and gastric cancer. although intestinal metaplasia is found much less commonly in laboratory animals, it has also been reported to occur in association with gastric cancer induced by polychlorinated biphenyls (ward, 1985) . in view of this association with gastric cancer, it has been suggested that intestinal metaplasia represents a pre-neoplastic lesion. however, over recent years prospective clinical studies and experimental data have suggested that it is an epiphenomenon, coexisting with, but unrelated to the development of cancer. in man, several forms of intestinal metaplasia have been described. these variants fall into two main groups, an incomplete type and a complete form (teglbjaerg and nielson, 1978; jass and filipe, 1979; jass, 1980) . complete intestinal metaplasia is characterised by the presence of goblet cells, paneth cells and absorptive cells with brush borders and variably developed intestinal villi. incomplete forms are more heterogeneous characterised by goblet and mucous columnar cells but no absorptive cells and variable patterns of mucin. the routine alcian blue: ph 2.5, periodic acid-schiff stain (ab/pas) ( table 2) distinguishes between the intestinal acid mucins (blue) from the neutral mucins of gastric type. however, variable sialomucin and sulphomucin staining patterns are seen in intestinal metaplasia in man with the high iron-diamine/alcian blue stain (hid/ab) (jass, 1980) . the incomplete form of intestinal metaplasia, showing marked sulphomucin secretion, has been found more commonly in association with gastric cancer in man (jass and filipe, 1979; jass, 1980; wells et al., 1982) . however, prospective studies have tended to indicate that intestinal metaplasia with sulphomucin secretion may be an age-related form of chronic atrophic gastritis and not a premalignant lesion (ectors and dixon, 1986) . it has been suggested that intestinal metaplasia represents an adaptive response to long-standing chronic inflammation and reduced acid secretion. it may also represent an adaptive defensive response to long-standing helicobacter pylori infection because intestinal mucosa is more resistant to these organisms (steer, 1984; ectors and dixon, 1986) . intestinal metaplasia has been found in association with gastric cancer in laboratory animals. fischer 344 rats treated with the polychlorinated biphenyl, aroclor 1254, mixed in the diet for 2 years developed foci of intestinal metaplasia in the stomach epithelium in association with gastric adenocarcinomas (ward, 1985) . these lesions were characterised by abundant mucin-containing cells and alkaline phosphatase activity typical of the small intestine (morgan et al., 1981; ward, 1985) . similar, but more diffuse intestinal metaplasia was reported in the stomach of primates treated with polychlorinated biphenyls, although unassociated with gastric neoplasia (allen, 1975; mcconnell et al., 1979) . intestinal metaplasia is also found in the stomach of laboratory animals treated with powerful genotoxic gastric carcinogens. although tsiftsis et al. (1980) showed hyperplasia and foci of atypical changes (dysplasia) but little or no intestinal metaplasia in rats following administration of n-methyl-nâ´-nitro-n-nitroguanidine, tatematsu et al. (1983) were able to show intestinal metaplasia in rats treated with the same agent. however, intestinal metaplasia can be induced in rodents by a variety of different procedures that are not associated with the development of gastric cancer. intestinal metaplasia can be induced in the glandular stomach of rodents by fractionated, localised, ionising radiation (watanabe, 1978; watanabe et al., 1980) , injection of xenogenic stomach antigens as well as propantheline bromide and the non-carcinogen, iodoacetamine shirai et al., 1985) . the characteristics of intestinal metaplasia in laboratory rodents are similar to those seen in man with early increases in intestinal enzyme activity (alkaline phosphatase, lactase, trehalase, sucrose and maltase), development of goblet cells containing neutral, sialo-, or sulphomucins, and intestinal crypts with or without paneth cells. both the fundus and antrum can show changes although as in man, males appear more prone to develop intestinal metaplasia than females . based on these experimental findings, have also proposed that intestinal metaplasia is not a precancerous condition but an adaptive response to a chronic elevation in ph in gastric secretion due to the early loss of parietal cell mass brought about by these various procedures. on balance therefore, the evidence to date suggests that although intestinal metaplasia is associated with cancer and may consequently be considered a helpful morphological feature in the evaluation of human gastric biopsies, the finding of isolated intestinal metaplasia in safety studies does not indicate a preneoplastic state. finally, a form of metaplasia in which hepatocytes have been found has been reported as rare incidental findings in the glandular stomach of mice sacrificed at the end of 2-year carcinogenicity bioassays (leiniger et al., 1990) . histologically, focal accumulation of well-differentiated hepatocytes were found in the submucosa and lamina propria adjacent to the limiting ridge with dilated adjacent gastric glands showing epithelial hyperplasia and mineralisation with herniation into the submusosa. whilst these foci were not believed to be related to treatment with xenobiotics, it is not clear whether represented metaplasia or congenital ectopia. the gastric glandular epithelium is predisposed to the deposition of calcium possibly as it is a site at which marked ion exchange normally takes place. focal aggregates or concretions of densely blue-staining mineral are fairly commonly observed in haematoxylin-stained sections from the stomachs of aged rats where they are associated with cystic dilatation of the gastric glands (greaves and faccini, 1992) . mice and hamsters occasionally show similar changes. small concretions are also observed in gastric glands in the beagle dog. these appear to represent aggregates of calcium around mucoid material. gastric mineralisation may become marked in rodents and dogs when there is disturbance of mineral metabolism, particularly in association with renal pathology. this has been well described in rats with severe renal disease (glomerulosclerosis) and parathyroid hyperplasia (snell, 1967) . a similar phenomenon has been described in the stomach of dogs in uraemic states (cheville, 1979) . identical changes result from the administration of drugs that induce prolonged azotemia or electrolyte disturbances. these changes are characterised by diffuse deposition of mineral in the intestinal tissue of the mucosa of the gastric body but not cardia, antrum or pylorus. mineral deposits develop around basement membranes surrounding epithelium and blood vessels. the lamina propria becomes expanded by oedema and fibroplasia of the interstitium also develops. the gastric glands themselves become distorted with swelling and degeneration of parietal cells and atrophy of chief cells. erosion of the glandular epithelium with haemorrhage occurs presumably as a result of the ischaemia caused by diffuse vascular injury and altered parietal cell function. focal atrophy of the gastric glandular mucosa is a sporadic occurrence in laboratory rodents, usually as a result of previous focal gastric inflammation, ulceration, mineralisation or vascular occlusion. these changes, characterised histologically by focal fibrosis of the mucosa, gastric glandular dilatation and atrophy variably accompanied by polymorphonuclear cells and mast cells are common in certain strains of rats when 2 years or more in age (anver et al., 1982) . whereas diffuse mucosal atrophy occurs following severe inflammatory insult, diffuse atrophy of the stomach glandular mucosa without inflammation can be a result of surgically or drug-induced reduction in trophic factors necessary for the maintenance of normal gastric morphology and function. this is observed in man and experimental animals following antrectomy because this removes the peptide-producing cells of the antrum (gjurldsen et al., 1968; neilsen et al., 1972) . in the rat, antrectomy is accompanied by hypogastrinaemia, reduced weight and height of the oxyntic mucosa and a reduced number of argyrophil cells (hã¥kanson et al., 1976 . this is in contrast to procedures such as antral exclusion that lead to hypergastrinemia and increased thickness of the oxyntic mucosa. mice with genetic deletion of the gastrin gene also show reduction in the thickness of the gastric mucosa. whilst all cell types are present, there is a most pronounced decrease in the numbers of parietal cells as well as enterochromaffin cells associated with an increase in surface mucous cells. these changes are linked to a profound decrease in acid secretion, which becomes unresponsive to histaminergic, cholinergic and gastrinergic stimulation (hinkle and samuelson, 1999) . analogous atrophic changes have been reported following pharmacological removal of trophic stimuli. for instance, administration of the cholecystokinin-b/gastrin receptor antagonist, ci-988 to cynomolgus monkeys for periods of up to 13 weeks was associated with an initial phase of multifocal degeneration of gastric glands primarily in the fundus followed by diffuse reduction in the thickness of the glandular mucosa with little or no qualitative changes to the cell populations (dethloff et al., 1997) . although bilateral vagotomy produces profound functional changes in the stomach, notably reduction of gastric acid secretion, morphological changes in the fundal mucosa are not marked either in experimental animals or in man (crean et al., 1969; aase and roland, 1977) . studies in the rat have shown that diffuse atrophy of the gastric glands characterised by a decrease in the number and size of parietal, chief and mucous cells occurs transiently following truncal vagotomy but histological features return to normal by about 1 month after surgery (nakamura, 1985) . by contrast, unilateral vagotomy in the rat leads to marked and persistent atrophy of the oxyntic zone on the denervated side. this is characterised histologically by reduced height of the mucosa and reduced numbers and staining intensity of argyrophil cells (hakanson et al., 1984) . hã¥kanson and his colleagues argued that this unilateral atrophy was due to the removal of the trophic action of the vagus. the lack of lasting atrophy after bilateral but not unilateral vagotomy was explained by the subsequent rise in gastrin that occurs after bilateral vagotomy as a result of lack of acid feedback inhibition of gastrin release (hã¥kanson et al., 1984) . removal or reduction in extra-gastric trophic factors or hormones may also reduce the thickness of the gastric mucosa. this is has been demonstrated in the rat by hypophysectomy which causes a reduction in thickness of oxyntic and antral mucosa, compared with pair-fed controls. although there was little or no change in peptic:parietal cell ratios, a significant decrease in cell volume and secretory activity of gastric glandular cells were demonstrated which suggested a widespread disturbance of synthesis and secretory mechanisms (bastie et al., 1985) . atrophic changes in the chief cells were observed in rats treated for 6 months with high doses of omeprazole, an inhibitor of acid secretion. the findings were considered to represent disuse atrophy secondary to the inhibition of acid secretion (hansson et al., 1986) . another inhibitor of gastric acid secretion, the tricyclic agent pirenzepin, also produced atrophy of the fundic mucosa of rats following 3 months but not 1 month of treatment (lehy et al., 1978) . the atrophy was characterised by reduction in parietal cell numbers associated with lower numbers of gastrin-containing cells in the antrum, features unlike those following prolonged treatment with histamine h 2 -receptor antagonists. an increase in the thickness of the gastric mucosa can be the result of hypertrophy or hyperplasia of the mucosal cells and this occurs both spontaneously or following administration of drugs and chemicals. in view of the different cell populations in the gastric mucosa and the variety of morphological alterations that occur, it is difficult to make a clear distinction between hypertrophy and hyperplasia without morphometric techniques. morphometric techniques have shown that hypertrophy of some mucosal cells can coexist with hyperplasia of other gastric cell populations. a distinction also needs to be made between diffuse or uniform hyperplasia involving one or more of the cell populations from the hyperplasia associated with proliferative or adenomatous overgrowth. adenomatous hyperplasia also needs to be evaluated for atypical cytological features (dysplasia), which are linked to development of gastric carcinoma (see below). cells of gastric glandular mucosa undergo increases in size or number in response to the effects of gastrointestinal trophic hormones or their synthetic analogues. similar changes also follow administration of compounds that inhibit gastric acid secretion or modify other trophic hormones or growth factors. when gastrin or its synthetic analogue, pentagastrin is administered subcutaneously to rats and mice for several weeks, there is both an increase in the number and size of parietal cells without concomitant increase in zymogenic chief cells (willems and lehy, 1975; crean et al., 1978; balas et al., 1985) . in addition, diffuse hyperplasia of enterochromaffin cells also occurs. by contrast, cholecystokinin, a trophic peptide found in the duodenum and sharing the same c-terminal tetrapeptide sequence as gastrin, increases in the number of chief cells but not parietal cells when administered to mice under similar conditions . drugs which inhibit or neutralise gastric acid secretion such as histamine h 2antagonists, proton pump inhibitors and antacids also induce hypertrophy or hyperplasia of the parietal cell population (witzel et al., 1977; crean et al., 1978; mazzacca et al., 1978; kaduk and hauser, 1980; betton et al., 1988; white et al., 1998) . these agents are associated with a rise in serum gastrin levels, probably as a result of loss of feedback inhibition of low antral ph on gastrin-producing g cells (witzel et al., 1977) . not all histamine h 2 antagonists produce identical effects. other cytological changes have been reported with famotidine, another h 2 -receptor antagonist. this agent produced a dose-related increase in the prevalence and degree of eosinophilic granularity in chief cells of the stomach in toxicity studies in rats but not dogs (burek et al., 1985) . electron microscopy showed an increase in electron density of zymogen granules and it was argued that these effects were the result of secondary inhibition of pepsin secretion or turnover due to inhibition of acid secretion. cytoprotective agents of prostaglandin type produce different forms of diffuse gastric hyperplasia. rats treated with 16,16-dimethyl prostaglandin e2 hourly for 3 weeks, not only developed forestomach alterations (see above) but also thickening of both the body and antral mucosa. in the body mucosa, these changes were the result of a proportional increase in the total mass of surface and foveolar mucous cells, mucous neck cells, chief cells, parietal and endocrine cells as well as connective tissue. this was largely as a result of increase in cell number, although parietal cells also increased in size (reinhart et al., 1983) . unlike treatment with gastrin and gastrin analogues there was an increase in number of surface and foveolar mucous cells associated with increase in mucus content. misprostal, a synthetic prostaglandin e1 methyl ester analogue also produced diffuse glandular hyperplasia, characterised by lengthening of gastric pits and increased mucous secretion in the preclinical safety studies in dogs and rats (kotsonis et al., 1985) . this glandular hyperplasia not only affected the body but also the antral mucosal. studies with tritiated thymidine showed that the labelling index was reduced in rats treated with misprostal, suggesting hyperplasia following administration of prostanoids is a result of an increase in cell survival and decrease in cell shedding rather than an increase in cell proliferation (fich et al., 1988) . levin (1988) has reviewed the effects of prostaglandins of the e series on the gastrointestinal tract of dogs and rodents. a dose-related diffuse hyperplasia of the gastric glandular mucosa has been reported in both rats and cynomolgus monkeys given human recombinant epidermal growth factor. the gastric mucosa was thickened and there was a increase in the number of undifferentiated cells particularly in the neck region and upper part of the gastric glands (breider et al., 1996; reindel et al., 1996) . mitotic figures were also numerous in the upper reaches of the mucosa. the lower parts of the gastric glands were generally less affected. the large increase in the number of undifferentiated cells may have a functional effect on gastric acidity and function (vinter-jensen, 1999) . administration of recombinant growth hormone has also been reported to induce thickening of the gastric glandular mucosa in dog toxicity studies along with typical growth hormone-induced changes in other organs, body weight increases and insulin-like growth factor (prahalada et al., 1998) . the pyloric and fundic mucosa showed histological evidence of hyperplasia of the mucous neck cells. thickening of the gastric glandular mucosa as a result of an irregular proliferation and cystic dilatation of gastric glands associated with inflammation characterises a number of non-neoplastic conditions in the stomach of man and laboratory animals. cystic change with chronic inflammation and foveolar hyperplasia is observed in biopsies taken from the edge of chronic gastric ulcers in man (franzin and novelli, 1981) . mã©nã©trier's disease ('polydadenomes en nappes'), a rare disease found primarily in middle-aged men is also characterised by enlarged gastric folds, foveolar hyperplasia and gastric glandular cystic dilation (berenson et al., 1976; wilkerson et al., 1998) . although its pathogenesis remains elusive, increased expression of transforming growth factor î± (tgfî±) and the epidermal growth factor receptor has been described (demsey et al., 1992) . as tgfî± is an epithelial cell mitogen that inhibits gastric acid secretion and increases gastric mucin, and transgenic mice that overexpress tgfî± in gastric mucosa develop a similar condition (see below), it was suggested by demsey et al. (1992) that tgfî± might have an important role in this condition. similar changes have been observed in animals in association with infestation of the gastrointestinal tract by parasites (jubb and kennedy, 1970; cook et al., 1981) . laboratory rodents may develop a similar pattern of changes spontaneously with advancing age, although the cause of this change remains uncertain. the distinction between adenomatous hyperplasia and adenoma is not clear cut. nevertheless, adenomas are usually defined as localised or focal proliferative lesions with well-ordered glandular patterns with a clear boundary with the surrounding normal mucosa. they are usually exophytic or polypoid in nature but adenomas with localised downward growth are described (mohr, 1997) . proliferation of the gastric glandular mucosa has been well characterised in the laboratory mouse because certain strains have a particular tendency to develop this condition spontaneously with advancing age (stewart and andervont, 1936; rowlatt et al., 1969) . hyperplasia also occurs spontaneously in conventional laboratory strains employed in carcinogenicity bioassays. its prevalence can be influenced by environmental factors such as housing (chvã©doff et al., 1980) , food restriction (rehm et al., 1987) and the administration of xenobiotics (poynter et al., 1985; betton et al., 1987) . similar gastric changes have also been reported to occur in mice thymectomised shortly after birth (suzuki et al., 1981) . histologically, these changes in mice are characterised by hyperplasia of the foveolar and neck regions of the body mucosa (fig. 47) . in advanced cases this is accompanied by elongated, tortuous, or dilated glands lined by simple columnar or cuboidal epithelium, devoid of parietal or chief cells. the abnormal cells show only mild cellular pleomorphism and mitotic activity. the abnormal glands dis-place normal glandular tissue and may penetrate through the muscularis mucosa to reach the muscularis externa and serosa. step sections demonstrate continuity between these glandular elements and a total absence of metastatic spread in the adjacent tissues and lymph nodes. the lamina propria also shows increased amounts of smooth muscle and collagen accompanied by variable numbers of lymphocytes and other chronic inflammatory cells. oedema may be observed and blood vessels are often dilated. the antral mucosa remains relatively unaffected. histochemistry has shown variable mucin secretion of the altered glands. some glands are devoid of mucin, others show an increase in sulphomucin as revealed by the high-iron diamine technique (greaves and boiziau, 1984) . it has been suggested that these features are similar to mã©nã©trier's disease in man and might have a similar pathogenesis (dempsey et al., 1992; takagi et al., 1992) . the aetiology of the spontaneous condition in the mouse is uncertain. the occurrence of similar lesions in thymectomized mice has given rise to the suggestion that autoimmune damage to the gastric mucosa may be responsible (kojima et al., 1980) . the presence of circulating anti-parietal antibodies and the decrease in the number of parietal cells in thymectomized mice suggested that autoimmune damage can occur to parietal cells with compensatory chronic stimulation and proliferation of the generative zones (suzuki et al., 1981) . however, based on findings in female han nmri mice, rehm et al. (1987) showed that this proliferative condition can develop in mice in the absence of antiparietal antibodies. they demonstrated that this change is associated with an increase in the number of antral gastrin cells, raising the possibility of a hormone or paracrine mechanism. a similar proliferative form of gastropathy has been reported in mice which overexpression transforming growth factor î± (tgfî±), a potent mitogen and member of the epidermal growth factor family of peptides. tgfî± acts by binding to and activating the tyrosine kinase of the epidermal growth factor receptor. transgenic mice overexpressing tgfî± develop severe adenomatous and cystic hyperplasia of the gastric glandular mucosa starting from about 2 months of age along with loss of mature parietal cell numbers and a diminution in gastric acid production (takagi et al., 1992) . the degree of change was to some extent dependent on the genetic background on which the transgene operated. an increased prevalence of similar changes have been reported in cd-1 mice treated with the novel histamine h 2 -receptor antagonist sk&f 93479 for 21 months (betton et al., 1987; . although treated mice developed hyperplasia of gastric neuroendocrine cells similar to that observed in rodents treated with other antisecretory agents, they also showed an increase in the severity of glandular hyperplasia. like the spontaneous condition, these changes were characterised by thickening of the mucosa by hyperplasia of the foveolar and neck regions, and downward proliferation of glandular elements into gastric glands (betton et al., 1988) . poynter et al. (1985) have also reported similar glandular hyperplasia in the mouse stomach associated with histamine h 2 -blockade with the agent ioxtidine. these findings were similarly associated with hyperplasia of neuroendocrine cells. adenomatous polyps have also been reported in the pyloric antrum of c57bl/ 10j mice treated for 52 weeks with the synthetic progestin, cyproterone acetate (tucker et al., 1996) . these were single, pedunculated and well-differentiated lesions showing little evidence of dysplasia. the mechanism for the induction of these polyps is unclear although they may have been hormonally mediated as progesterone receptors have been identified in gastric tissue (wu et al., 1992) . although usually less prevalent and less exuberant than in mice, the aged rat also develops proliferative gastric glandular changes spontaneously. these changes are characterised by hyperplasia of the foveolar and mucin-secreting cells of the body mucosa, development of cystic glands lined by simple mucous or flattened cells, accompanied by chronic inflammatory cells, prominent blood vessels and smooth muscle in the lamina propria (greaves and faccini, 1992) . the antrum remains relatively unaffected. proliferative alterations can be induced by administration variety of xenobiotics as well as following surgical procedures that induce chronic reflux of normal intestinal contents. for instance, hyperplasia of the gastric mucosa, notably over the lesser curvature has been described in rats following the so-called bilroth ii gastrectomy that allows reflux of intestinal and biliary secretions into the stomach (kobayashi et al., 1991) . a proliferative condition of the gastric mucosa has been shown to develop following long-term treatment of rats with an ulcerogenic regimen of aspirin. female sprague-dawley rats given 250 mg/kg of aspirin in 1% methylcellulose once daily orally by gavage for 6 months followed for periods of up to 18 months without treatment, developed focal proliferative changes at the sites of healed ulcers, mainly in the mucosa of antrum or antral-body junction (st john et al., 1977) . these lesions were characterised by the presence of proliferating gastric glands lined by columnar, cuboidal or flattened epithelial cells in the mucosa, which also extended through the muscularis mucosa. mucus content of these glands was variable but when present was principally acidic in type, as shown by staining with alcian blue at ph 2.5. the lesions were accompanied by increased collagen in the lamina propria, endarteritis and an infiltration of lymphocytes, plasma cells and mast cells. the lesions were not associated with the development of carcinoma following 18 months' observation and it is probable that they were the result of the chronic damage and repair induced by aspirin treatment. hyperplasia of the gastric glandular mucosa also occurs in rats following the administration of powerful genotoxic carcinogens, although characteristically in association with of atypical histological changes and ultimately carcinoma. these changes have been best characterised in sequential studies with the rat using the carcinogen n-methyl-nâ´-nitro-n-nitrosoguanidine at doses low enough to avoid overt gastric ulceration and regenerative hyperplasia. it was shown that hyperplasia developing under these conditions occurs diffusely both in the body and antral mucosa. furthermore, the changes occurred earlier in the antrum than in the body and were focal or polypoid in character (tsiftsis et al., 1980) . involvement of the antrum in this way is quite unlike the spontaneous hyperplasia of the rat gastric mucosa. histologically, this form of hyperplasia is characterised by lengthening of the foveolae and neck regions both in the antrum and body. hyperplastic pits or foveolae show increased secretion of sialomucins and sulphomucins with a concomitant loss of neutral mucins. polychlorinated biphenyls such as arochlor 1254, which produce intestinal metaplasia and adenocarcinoma in the stomach of rats, also induce proliferative alterations characterised by proliferative cystic lesions in the mucosa associated with inflammation and fibrosis (morgan et al., 1981) . in common with lesions induced by genotoxic agents, these changes are found primarily in the antrum and pyloric regions, zones of predilection for the development of gastric carcinoma in man and experimental animals. it is important to distinguish between the various hyperplastic and adenomatous conditions found in the gastric glandular mucosa in laboratory animals that are not associated with neoplasia from those which precede the development of carcinoma. this distinction is complicated by the fact that proliferative changes associated with the development of cancer both in man and laboratory animals possess features in common with lesions not associated with neoplasia. however, a key distinctive feature is the degree of epithelial dysplasia. dysplasia is considered to be the lesion common to gastric conditions in man such as atrophic gastritis and gastric polyps that have been linked with a significantly increased risk of gastric cancer. although the term dysplasia may be less widely employed in experimental pathology, similar dysplastic changes to those occurring in man have been characterised in laboratory animals in which precancerous gastric lesions have been studied (tsiftsis et al., 1980) . it therefore represents a unifying concept in the assessment of proliferative changes in the gastric glandular mucosa of laboratory animals. as defined by an international group concerned with the diagnosis of preneoplastic conditions in the stomach of man, the principle features of dysplasia are (i) cellular atypia; (ii) abnormal differentiation; and (iii) disorganised mucosal architecture (morson et al., 1980; nagayo, 1981) . cellular atypia is characterised by nuclear pleomorphism, hyperchromasia and stratification of nuclei, increased nuclear-cytoplasmic ratio and loss of cellular and nuclear polarity. abnormal differentiation is shown by reduction or alteration in the normal secretory products of the mucosa. disorganised mucosal architecture is shown by irregularity of crypt structure, back-to-back glands, budding and branching of crypts and intraluminal and surface papillary growths. it is important to assess gastric mucosa very carefully for the features of dysplasia when hyperplastic gastric changes are found in treated animals. in the rat gastric cancer model employing the carcinogen n-methyl-nâ´-nitro-n-nitrosoguanidine, dysplastic changes were shown to start in the proliferating neck region of hyperplastic zones (tsiftsis et al., 1980) . these changes were characterised histologically by irregular growth patterns of glandular cells showing reduced mucin secretion, numerous mitoses and enlarged pleomorphic nuclei. these atypical glands were observed to extend downwards, eventually replacing normal gastric glands and ultimately penetrating the muscularis mucosa forming infiltrating adenocarcinomas of variable differentiation. the antrum developed these changes earlier than the body mucosa (tsiftsis et al., 1980) . these considerations were important in the safety evaluation of the histamine h 2 receptor antagonist, tiotidine (ici 125,211) a guanidino-thiazole derivative that also produced proliferative gastric lesions in the stomach of rats in a 24month carcinogenicity study (streett et al., 1984; . these changes were found mainly in the pyloric region and were characterised histologically by superficial erosions and irregular pyloric glands lined by cells with basophilic cytoplasm and enlarged hyperchromatic nuclei. some atypical glands penetrated the muscularis mucosae. dysplastic lesions situated primarily in the pyloric region were also associated with the development of invasive carcinoma in some rats. extensive histological sectioning of the stomach in rats treated with tiotidine for only 6 months also revealed evidence of early proliferative changes (streett et al., 1988) . therefore, these lesions produced by tiotidine possessed more in common with those induced by powerful carcinogens such as n-methyl-nâ´-nitro-n-nitrosoguanidine than the benign, species-specific proliferative change of little or no relevance for human safety. interestingly, mice treated for 18 months with tiotidine were devoid of dysplastic changes in the gastric mucosa (streett et al., 1988) . one of the most remarkable examples of drug-induced gastric alterations in rodent bioassays is the hyperplasia of enterochromaffin cells and development of carcinoid-like neoplasms in the stomach of rats treated with omeprazole (havu, 1986) . omeprazole is a substituted benzimidazole which inhibits gastric acid secretion by blocking the enzyme h + , k + -atpase, the proton pump of the parietal cells, in a specific and dose-dependent manner (fellenius et al., 1981) . although in rats there is a increase in number of gastric argyophilic cells with increasing age, rats treated with omeprazole for 104 weeks showed a marked, dose-related and diffuse increase of argyophilic, non-argentaffin cells in the basal half of the oxyntic fundal mucosa (havu, 1986) . these changes were more marked in female than in male rats but were not observed in the bioassay in which cd-1 mice were treated with similar doses of omeprazole for 78 weeks. these diffuse changes in the rat stomach were associated with focal hyperplasia of argyrophilic cells. these focal lesions were also associated with a doserelated increase in larger focal nodular lesions of argyrophilic cells, some of which were undoubtedly locally infiltrating carcinoid tumours. these nodular argyrophil lesions posed the usual problems of differential diagnosis of endocrine hyperplasia and neoplasia (see endocrine system, chapter xiii) and distinction of hyperplasia from neoplasia was uncertain. histologically, nodular lesions were composed of multifocal anastomosing solid or pseudoacinar cords of proliferating, regular cells with uniform nuclei and moderately abundant fine granular pale cytoplasm. these nodules showed little or no cellular pleomorphism or mitotic activity but clear evidence of submucosal infiltration without involvement of the muscularis externa was observed in some cases. the overall light microscopic features were similar to those of gastrointestinal carcinoid tumours reported in man. the incidence of gastric carcinoids was reported to be as high as 40% in females in the high dose group but only a few cases observed in similarly treated males (ekman et al., 1985; havu, 1986) . electron microscopy of the altered argyophil cells confirmed the presence of electron-lucent, vesicular granules, frequently with small irregular dense cores characteristic of enterochromaffin cells of the stomach. immunocytochemical study showed that these cells contained histidine decarboxylase which is found normally in gastric enterochromaffin cells which produce and store histamine . other findings reported in rats treated with omeprazole have been a proportional increase in the number and size of non-endocrine cells of the fundus (blom, 1986) , an increase in the number and immunostaining properties of the antral gastrin-containing g cells and hypergastrinaemia (bishop et al., 1986; creutzfeldt et al., 1986) . all functional and morphological changes following treatment for 60 days were fully reversible after 42 days' drug withdrawal (creutzfeldt et al., 1986) . as a result of these treatment-related increases of these normally rare gastric carcinoids in the rat bioassay with omeprazole, clinical trials with this agent were suspended until it was agreed that the endocrine alterations were a result of prolonged drug-induced achlorhydria. it was postulated that omeprazole causes a prolonged inhibition of acid secretion in the rat, which causes activation, and subsequently hyperplasia of antral gastrin cells and marked hypergastrinaemia. hypergastrinaemia in turn stimulates enterochromaffin cells of the fundus, which in time results in enterochromaffin hyperplasia . this argument is supported by the fact that similar morphological findings are reported in chronic atrophic gastritis and other achlorhydric sites in man (solchia et al., 1986; mã¼ller et al., 1987) and that antrectomy in the rat prevents the appearance of enterochromaffin hyperplasia following treatment with omeprazole . although mild dose-related gastric argyrophil cell hyperplasia was noted in dogs treated with omeprazole for 1 year, neoplasms of the stomach were not observed during this time period. why mice neither developed neither argyrophil hyperplasia nor gastric carcinoids with a similar treatment regimen is not clear, as the mechanism of action of omeprazole is similar in rat, dog and mouse. however, as the duration of action of omeprazole is shorter in the mouse, it was postulated that sustained inhibition of gastric acid secretion over 24 hours is necessary to activate increased gastrin secretion from antral cells (havu, 1986) . it has also been suggested that the mouse possesses fewer gastric enterochromaffin cells than the rat and shows a much lower serum gastrin response to omeprazole treatment (ekman et al., 1985) . duration of action or potency may also be the explanation for the lack of reports of carcinoid neoplasms in rats following inhibition of gastric acid secretion by the histamine h 2 -receptor blockers cimetidine and ranitidine. neither of these drugs completely inhibits gastric acid secretion in the rat for 24 hours (leslie and walker, 1977; larsson et al., 1986) , although mild gastric neuroendocrine hyperplasia has been recently described in cimetidine-treated rats (hirth et al., 1988) . however, the long-acting h 2 -receptor antagonist sk&f 93479 produced gastric carcinoid neoplasms when administered at a high dose (1000 mg/kg) to rats for 2 years (figs. 48 and 49) (betton et al., 1987; . although this dose level of sk&f 93479 did not entirely suppress gastric acid secretion and control gastric ph over 24 hours, plasma gastrin levels remained elevated at 3-4 times control values over this period. in a 21-month oral carcinogenicity study in cd-1 mice at the same dose level (1000 mg/kg), a diffuse neuroendocrine cell hyperplasia and multifocal glandular hyperplasia or neoplasia was also observed (betton et al., 1987) . similarly, loxitidine, a potent, non-competitive, insurmountable histamine h 2 -antagonist produced hyperplasia of neuroendocrine cells and carcinoid tumours in the gastric fundus of both rats and mice after 2 years' treatment in diet and drinking water, respectively (poynter et al., 1985 (poynter et al., , 1986 . other histamine antagonists bl-6341 and ici 162846 have been reported to produce neuroendocrine neoplasms in the stomach of rats and rats and mice, respectively (hirth et al., 1988; streett et al., 1988) . wormsley (1984) reviewed the considerations in the risk-benefit analysis of agents intended for long-term administration for peptic disease. drugs of other classes also cause hyperplasia of gastrin-containing cells. immunocytochemical study using antigastrin antibody revealed increased gastrin cell numbers in the antral mucosa of dogs given high doses of adrenocorticosteroids for 4 weeks and these changes were accompanied by enhanced serum and tissue gastrin levels (delaney et al., 1979) . these results suggest that adrenocorticosteroids have a trophic effect on gastrin-containing cells. in human patients hypergastrinaemia is also produced by pharmacologically induced hypochlorhydria although this is usually only slight and hyperplasia of enterochromaffin cells has not been observed (soll, 1990) . a complicating factor in drug safety evaluation is the association of gastric cancer in both man and laboratory animals with n-nitroso compounds. some of the most effective stomach carcinogens in laboratory animals have proved to be nnitroso compounds particularly since sugimura and fujimura (1967) induced gastric adenocarcinomas in rats with n-methyl-nâ´-nitro-n-nitrosoguanidine dissolved in drinking water. furthermore, epidemiological evidence associating nnitroso compounds with human cancer is also fairly strong for the stomach (corea et al., 1975; pocock, 1985) . the formation of n-nitroso compounds is theoretically possible with a variety of compounds that contain amino groups. it has been suggested that the formation of nitrosamines occurs in vivo under the acidic conditions in the stomach following dietary ingestion of nitrite, nitrates and secondary amines (mirvish, 1975 (mirvish, , 1983 . low levels of preformed nitrosamines are also present in some commercial pelleted diets for laboratory animals, principally derived from fishmeal (edwards et al., 1979) . calculations based on dietary intake and nitrosatability of precursors and carcinogenicity of derivatives have suggested that the risk that arises from endogeneous nitrosation is highly variable but highest from ureas and aromatic amines (shephard et al., 1987) . a number of drugs in widespread clinical use have been shown to produce nnitroso products in acidic aqueous media, although the extent to which this occurs in actual therapeutic use is unclear (gillatt et al., 1985) . some clinical evidence suggests that nitrosation of therapeutic agents can occur in clinical practice. for instance piperazine, a cyclic secondary amine, widely used as an antihelmintic drug, has been shown to form small quantities of n-mononitrosopiperazine in the human stomach as measured by gas chromatography-thermal energy analysis (bellander et al., 1985) . however, n-mononitrosopiperazine has not been shown to be carcinogenic in rodents (love et al., 1977) . n,nâ´-dinitrosopiperazine, carcinogenic to the upper gastrointestinal tract in rodents (lijinsky and taylor, 1975) , was not detected in man after administration of piperazine under the same circumstances (bellander et al., 1985) . the possibility of nitrosation is not usually taken into account in the testing of carcinogenic potential of novel drugs as bioassays are usually only performed with parent compound. however, concerns about nitrosation have arisen in subsequent clinical practice. an example of this was the proposal that a few gastric cancers found in patients whilst being treated with the histamine h 2 receptor antagonist cimetidine, were the result of treatment (elder et al., 1979; reed et al., 1979; hawker et al., 1980) . it now seems likely that all those observed cancers associated with cimetidine were incidental (penston and wormsley, 1986) . however, at that time concerns were increased by the theoretical possibility that cimetidine has the potential be nitrosated in vivo (elder et al., 1982) . a further factor was the concept that the treatment-induced gastric secretory inhibition with subsequent bacterial colonisation of the stomach rendered the conditions conducive to the generation of n-nitroso compounds from normal dietary constituents (reed et al., 1981; penston and wormsley, 1986) . these concerns appear to be unfounded. long-term surveillance studies with cimetidine have shown no causal link between its clinical usage and gastric malignancy (colin-jones et al., 1965; langman, 1985) . in addition, carcinogenicity bioassays performed with cimetidine, cimetidine plus nitrite and nitrosocimetidine have not shown any tumorigenic effect in the gastric mucosa (anderson et al., 1985) . a 7-year study in dogs in which multiple gastric biopsies were taken at intervals of approximately 6 months have also shown no indication of gastric hyperplasia, dysplasia, intestinal metaplasia or neoplastic change . although complacency is certainly not warranted with respect to the nitrosation of therapeutic agents in vivo, the risks of development of gastric malignancy from such drugs when administered on a short-term basis are probably very small (world health organisation, 1978) . even for gastric antisecretory agents which may be administered for longer periods of time, the balanced view would also permit development of novel agents provided they are not obviously mutagenic or carcinogenic in the usual preclinical studies and are not particularly liable to undergo rapid nitrosation. most carcinomas of the glandular mucosa are adenocarcinomas, whether induced by the potent genotoxic carcinogens or therapeutic agents. these tend to develop in the antral region in rodents (streett et al., 1984; szentirmay and sugar, 1985) . they range from those with well differentiated tubular or papillary features to poorly differentiated forms with trabecular, mucoid or signet ring features (tatematsu, 1997) . squamous metaplasia within adenocarcinoma can also be observed. stroma may be abundant with pronounced chronic inflammatory infiltration and hyalinisation. metaplastic cartilage and bone has also been described (szentirmay and sugar, 1985; mohr, 1997) . gastric adenocarcinomas induced in dogs by n-methyl-nâ´-nitro-n-nitrosoguamide show similar histological features although their reported distribution in the stomach appears more variable (fujita et al., 1974) . histological criteria for the diagnosis of invasive adenocarcinoma in experimental animals may vary between individual pathologists. some retain the old criteria of stewart and co-workers (1961) who defined invasive cancer as a neoplastic growth reaching the serosa. it is now considered more appropriate to apply criteria of use in human diagnostic pathology (see mohr, 1997) . unequivocal invasion of the submucosa is sufficient evidence of an invasive and therefore malignant process (greaves and faccini, 1992) . the small intestine is of major importance in drug safety evaluation for it represents the primary site of drug absorption. in view of its length and the presence of villi, it possesses an enormous surface area of specialised absorptive epithelium. furthermore, ingested substances have an extended residence time in this part of the gastrointestinal tract. the canine model has been one of the most popular for the study of drug absorption because the dimensions of the canine gastrointestinal tract permit administration of dosage forms intended for clinical use in humans. for this reason, factors, which influence drug absorption, have been better studied in dog and man than many other species. however, data from dog and man not only suggest that diverse factors influence drug absorption from the small intestine but that there are considerable species differences. residence time is of particular importance for drugs, which are incompletely absorbed because differences in mucosal contact time can be expected to result in differences in the fraction absorbed. dressman (1986) has shown using the heidelberg capsule technique that small intestine transit time in dogs is varies from between 15 and over 200 minutes whereas in man equivalent times are between 180 and 300 minutes. these results suggest that absorption of poorly absorbable drugs is likely to be quantitatively less although more variable in dogs than in man. however, these differences do not explain why some poorly lipophilic drugs such as chlorothiazine, acyclovir and phosphalinic acid are more extensively absorbed in dogs than in man. intestinal ph is consistently higher in dogs than in man so that drugs with half maximal absorption ph in the range ph 5-7 may also be expected to be absorbed at different rates in man and dog (dressman, 1986) . physiological and anatomical differences in the small intestine of other test species particularly rodents and humans are also likely to have an impact on drug absorption although many of these factors are still poorly understood. in addition to the small intestine acting as an absorptive surface, it is becoming increasingly obvious that it plays an important part in the metabolism of drugs (breckenridge, 1987) . although monoxygenase activity is relatively low in the gut compared with the liver, conjugation mechanisms are efficient and activity of udp-glucuronosyltransferase and glutathione-s-transferase are as high or even higher than in the liver (hã¤nninen et al., 1987) . in addition, the gastrointestinal microflora not only possesses metabolic capacity itself but also can influence the turnover rate of mucosal cells and subsequent exfoliation and release of enzymes into the lumen (hã¤nninen et al., 1987) . gastrointestinal metabolising activity is important because that the mucosa is exposed to high concentrations of xenobiotics in toxicity studies, and this can influence their overall bioavailability (chhabra and eastin, 1984) . studies in untreated rats have shown that the concentration of total cytochrome p450 in small intestinal microsomes is only about 10% of that found in liver microsomes (bonkovsky et al., 1985) . however, it exists in at least two forms and as in the liver, its activity can be induced by xenobiotics. it has been shown that in the rat, the concentration of cytochrome p450 and drug metabolising enzyme activity increases in intestinal epithelial cells as they move from crypt to villous tips and they are found in greater concentration in the proximal two-thirds of the small intestine than in the distal third (hoensch et al., 1976; bonkovsky et al., 1985) . bonkovsky et al. (1985) also showed that the phenobarbital-inducible form of cytochrome p450 represents less than 5% of total p450 in the small bowel, but as in the liver, phenobarbital treatment can increase this form to about 50% of total cytochrome p450 in small intestine cells. furthermore, it has been shown that drug metabolising activity in the tips of the villi in the duodenum is greater in rats fed a conventional diet than a semisynthetic diet and that the activity depends critically on the absorption of iron from the intestine (hoensch et al., 1976) . glutathione is also present throughout the entire mucosa, although in rats, cells at the tips of the villi contain less than cells located more basally, whereas related enzymes î³-glutamyl transpeptidase and glutathione-s-transferase show highest activity in the villous tip region (ogasawara et al., 1985) . the fact that these enzyme activities are highest in the duodenum and lowest in the terminal ileum suggests that detoxification systems for exogenous compounds are greater in the proximal small intestine. the small intestinal mucosa is constructed not only to act as an absorptive surface but also as a barrier to potentially pathogenic substances and microorganisms. although the main cell population of the epithelium is composed of absorptive cells, other major epithelial cell types, the mucous (goblet) cells, paneth cells and endocrine cells have important protective functions. in addition, specialised epithelial cells, the microfold (membranous or m) cells are located in the epithelium over peyer's patches. these cells form part of the other important protective system of the intestine, the gut associated lymphoid tissue (galt) or mucosal associated lymphoid tissue (malt). the mucosal lining is in a constant state of renewal. enteric epithelium possesses the fastest rate of turnover of any tissue exceeded only by a few rapidly growing neoplasms . in normal circumstances, the constant turnover of small bowel mucosa is maintained by equilibrium between cell production in the crypts and cell loss at the tips of the villi. there are intrinsic controls within the mucosa itself. exogeneous substances, intraluminal secretions, mechanical and neural factors as well as alterations in blood flow all possess potential to influence mucosal cell kinetics . all main epithelial cell types are believed to arise from undifferentiation columnar cells at the crypt base (cheng and leblond, 1974) , although mucous cells may also arise by proliferation of partly differentiated mucous cells in the crypts. as cell division is limited to crypts, the cell population in the crypts have high activities of enzymes such as thymidine kinase that are involved in nucleic acid synthesis (imondi et al., 1969) . the complete cell cycle lasts about 10-17 hours in rodents and at least 24 hours in man. enteric epithelium is completely replaced within 2-3 days in mice and rats and within 3-6 days in man . after two or more divisions in the crypt cells migrate to the villus, lose ability to incorporate thymidine and differentiate into mature cells equipped with enzymes associated with nutrient absorption (imondi et al., 1969) . cell migration in the rat is completed more rapidly in the ileum than in the jejunum principally as a result of the lower villous height in the ileal mucosa (altman and enesco, 1967). migration terminates by loss of cells from the tip of the villi. surrounding the crypt is a sheath of fibroblastic cells. these cells also undergo synchronous division and migration with the epithelial cells, maintaining the intimate relationship between the epithelium and supporting tissues (parker et al., 1974) . mature absorptive cells are important in the active and passive transport of nutrients as well as in the endocytosis of macromolecules. they are characterised by the presence of a striated or brush border which is seen in haematoxylin and eosin-stained sections as a refractile bi-laminar band. the inner, wider lamina corresponds to the microvillous region that is associated with the presence of neutral mucosubstances in most species. the outer, thinner band corresponds to the glycocalyx, which is composed principally of acidic mucosubstances (sheahan and jarvis, 1976) . this outer band of the brush border shows histochemical staining predominantly for sulphomucins in most species including mouse, hamster, dog and rhesus monkey, although in the duodenum of the rats and in the entire human small bowel sialomucins predominate in this layer. electron microscopy of the absorptive cells shows that the surface of absorptive cells is covered by tightly packed and well developed microvilli approximately 1 âµm long and 0.1 âµm wide. these are considered the first site of entry of food substances into the cell. the plasma membrane of microvilli is associated with fine filamentous projections, which probably represent the polysaccharide chains of the glycocalyx (bennett, 1969) . as the glycocalyx is composed of a network of polysaccharides, it has been suggested that it may behave like an ionexchange resin, be able to bind certain lectin-like molecules or trap substances in its matrix so providing a site for efficient intraluminal digestion (bennett, 1969; goldberg et al., 1969; king et al., 1986 ). the plasma membrane shows a trilamina structure at ultrastructural level. freeze fracture replicas from the microvilli which cleave this membrane through the plane of apposed non-polar groups of the lipid bilayer, demonstrate smooth complementary surfaces studded with small particles. these particles represent integral globular proteins of the plasma membrane. some of these intramembranous particles, mostly those of 10 nm diameter show irregular outlines with a central pit and are believed to represent gap junctions or transport channels (yamamoto, 1982) . a particularly important aspect of the absorptive cell membrane is its high concentration of disaccharidases such as sucrase, maltase and lactase, related to the absorption of sugars. alkaline phosphatase activity is also abundant on the surface of absorptive cells, although its precise role here uncertain (owen and bhalla, 1983) . immunocytochemical demonstration of alkaline phosphatase provides a useful tool to examine the effects of xenobiotics on intrinsic membrane glycoproteins in the small intestine (hasegawa et al., 1987) . enterokinase, the glycoprotein enzyme, which initiates the activation of pancreatic zymogens by converting trypsinogen to trypsin, is also present in the brush border and glycocalyx of the small intestinal epithelium, both in man and animals. immunocytochemical studies have demonstrated that in man this enzyme is located in the duodenum and proximal jejunum but not ileum, colon and stomach (hermon-taylor et al., 1977) . these enzymes constitute integral structural proteins of the cell membrane with active sites protruding from the cell surface. they are synthesised by the absorptive cells (blok et al., 1984) . the lateral surfaces of absorptive cells are in direct contact with neighbouring cells and firmly attached to each other by terminal bars or junctional complexes. a terminal bar comprises an apically situated tight junction or zonula accludens, a central zone, the zonula adherens, below which is situated a desmosome or macula adherens. the junctional complexes are relatively impermeable to macromolecules. studies with labelled tracers in the rat jejunum have shown that horseradish peroxidase (molecular weight 40,000, diameter 5 nm) and ferri-tin (molecular weight 100,000, diameter 10 nm) do not penetrate junctional complexes (yamamoto, 1982) . below the junctional complexes, the cell membranes interdigitate and the intercellular space widens towards the cell base, a feature that may be important in the movement of electrolytes and water across the intestinal epithelium (rhodin, 1974) . the cytoplasm of absorptive cells contains smooth and granular endoplasmic reticulum, free ribosomes and mitochondria. the golgi apparatus is located above the nucleus. the apical part of the cytoplasm is devoid of organelles except for a tight meshwork of filaments called the terminal web. filaments of actin within the microvilli are linked to the terminal web and this is believed to be important in the movement of microvilli (moosker and tilney, 1975; moosker et al., 1978) . goblet cells are much fewer in number than absorptive cells in the small intestine but they increase in number from the duodenum to the lower ileum. they are important in the production of mucus, which remains on the surface of the mucosa as a viscous layer and acts as the first line of defence against intestinal pathogens. goblet cells are characterised by the presence of abundant mucous droplets formed by the golgi complex and which accumulate in the apical part of the cell cytoplasm. histochemical study shows that neutral mucosubstances are present in the goblet cells found in crypts and on the villi in the entire small bowel mucosa of most species including man but there is an interspecies variation in the population of sialo-and sulphomucins (sheahan and jarvis, 1976) . in the mouse, sulphomucins predominate but among rats considerable individual variation in the proportion of sialo-and sulphomucins is reported. in the hamster, sulphomucins are more prominent in the proximal and sialomucins in the distal small bowel. in the dog, both sulphomucins and sialomucins are found with predominance of one or other in individual animals. staining for acidic mucins is less intense in the goblet cells of the small intestine in man compared with non-human primates but sialomucins are predominant in both species. a few goblet cells in the distal ileum in man also contain sulphated mucins (sheahan and jarvis, 1976) . paneth cells remain located near the crypt base throughout the small intestine (sandow and whitehead, 1979) . they are found in rodents and man but typically not in carnivores such as dog and cat (rhodin, 1974; . they are characterised by the presence of numerous eosinophilic cytoplasmic secretory granules between about 1.0 and 2 âµm diameter that contain various enzymes and mucosubstances. particular care is needed in fixation and staining for optimal demonstration of paneth cells for they rapidly degranulate after death and granules are destroyed by acetic acid fixation. formalin and mercuric fixatives appear appropriate methods and they permit staining with methylene blue, lendrum's phloxine-tartrazine and masson's trichrome (lewin, 1969) . the apical parts of paneth cells show glucose-6-phosphatase, carbonic anhydrase and monoamine oxidase activity and they have been shown to contain lysozyme and immunoglobulins, particularly iga (speece, 1964; riecken and pearse, 1966; ghoos and vantrappen, 1971; . staining for lysozyme appears to be the most practical immunocytochemical stain for the detection of paneth cells in formalin fixed paraffin wax embedded tissue sections. however, other immunocytochemical reagents have been employed in human and rodent tissues. these include antibodies to the cd15 antigen (ariza et al., 1996) and to a rat paneth cell zinc binding protein (sawada et al., 1994) as well as the use of pokeweed lectin for murine paneth cells (evans et al., 1994) . the paneth cell granules not only contain lysozyme but also a range of antimicrobial peptides such as secretory phospholipid a 2 , î±-defensins, also called cryptins. these are believed not only to possess antimicrobial activity but also important in the regulation of cell volume, chemotaxis, mitogenesis and inhibition of natural killer cell activity (ouellette, 1997) . it has been shown that these substances are released from paneth cells when germ free rats are dosed with the intestinal flora from specific pathogen free rats (satoh and vollrath, 1986; . however, it has been shown that paneth cells develop under germ free conditions and do not require luminal bacteria or dietary material for their development (ouellette, 1999) . endocrine cells are also scattered throughout the small intestinal mucosa. they are of both argentaffin and argyrophil types and are situated predominately in crypts. immunocytochemical study shows that they contain a variety of different peptides although gastrin, secretion and serotonin-containing cells have been those most extensively studied (inokuchi et al., 1983) . in addition to the barrier formed by mucus and epithelial cells, lymphocytes, plasma cells, macrophages, dendritic cells and mast cells also form part of the protective function of the small intestine. some lymphocytes are located within the epithelium mostly, above the basal lamina but below epithelial nuclei (pabst, 1987) . these lymphocytes are termed intraepithelial lymphocytes and are predominantly of t-suppresser/cytotoxic type in man and laboratory animals (selby, 1981; martin et al., 1986; pabst, 1987) . most lymphocytes in the lamina propria are also t cells but t-helper (cd4 positive) cells outnumber the t-suppresser/cytotoxic or cd8 positive phenotype (hirata et al., 1986; pabst, 1987; bruder et al., 1999) . intraepithelial lymphocytes are also mainly t cells (bruder et al., 1999) . many plasma cells present in the lamina propria produce iga, the major immunoglobulin of mucosal secretions (michalek et al., 1975) . iga represents another important component of the mucosal barrier between the gastrointestinal mucosa and intraluminal antigens. the main function of iga is to effect immune exclusion by intimate co-operation with non-specific defence mechanisms (brandtzaeg et al., 1985) . plasma cells in the lamina propria produce dimeric iga with two dimeric molecules joined by a joint (j) piece. a secretory component (sc), a glycoprotein expressed on the basolateral surface of epithelial cells acts as a receptor for dimeric iga and as a transport system for iga to the gut lumen where monomeric iga is secreted (brandtzaeg et al., 1985) . morphometric analysis of iga-containing immunocytes in the rat ileal mucosa using immunocytochemical staining has shown that the number of these cells varies with alterations in the microbiological status of intestinal contents (rodning et al., 1983) . a significant reduction in iga-containing lymphocytes and plasma cells was observed following microbial reduction associated with gnotobiosis, probably reflecting decreased microbial antigenic stimulation. experimental studies using labelled mesenteric lymphocytes also suggests that local microenvironments are important in the distribution of there cells in the intestinal wall (mcdermott et al., 1985) . peyer's patches are the most prominent aggregates of lymphoid tissue in the gastrointestinal tract and constitute important sites at which antigens from the gut lumen encounter immune competent cells which are responsible for the initiation of immune responses. peyer's patches are located on the ante-mesenteric wall of the small bowel and consist principally of collections of lymphoid follicles. in man, peyer's patches are more common in the ileum (cornes, 1965) but in mice they are more uniformly distributed (owen and neumanic, 1978) . in rats they are also more numerous in the distal than in the proximal small intestine and the number of follicles in patches usually varies from 2 to 6 but sometimes many more may be seen in any particular section (martin et al., 1986) . peyer's patches vary between rat strains. a comparative study showed that in fischer 344 rats they are smaller than those in wistar rats (bruder et al., 1999) . particular care in selection and orientation of tissue blocks is therefore essential for any form of quantitative or semiquantitative assessment of peyer's patches. peyer's patches are more than simple aggregates of lymphoid follicles. they consist of lymphoid follicles surrounded by a corona of small lymphocytes principally of b-cell type. the interfollicular area contains post-capillary venules with specialised cobblestone type epithelium (yamaguchi and schoefl, 1983 ) and many t lymphocytes. beneath the epithelium, over the bulging follicles (dome area), mixtures of t and b lymphocytes, plasma cells and macrophages can be seen (pabst, 1987) . immunohistochemical study in rats demonstrates the presence of w3/13-positive t (cd43) cells in the interfollicular area. in the rat, many cells with macrophage morphology also stain with the w3/25 (cd4) monoclonal antibody in the interfollicular zone (bland and warren, 1985; martin et al., 1986) . immunocytochemical study of the peyer's patches in the mouse has also shown considerable heterogeneity of staining patterns, particularly under the dome epithelium (ermak and owen, 1986) . similar patterns have been observed following immunohistochemical study of peyer's patches in man (spencer et al., 1986) . the epithelium overlying the peyer's patch follicles (dome area) contains specialised epithelial cells, called microfold, membranous or simply m cells. these cells have been identified in many species including rats, mice, hamsters, dogs, monkeys and man (owen and bhalla, 1983; wolf and bye, 1984) . these cells differ functionally from other enterocytes by their ability to transport large molecules such as ferritin and horseradish peroxidase and particulate matter from the lumen to the underlying lymphoid tissue (owen, 1977; jeurissen et al., 1985) . they have also been shown to be the site of penetration of reoviruses into the epithelium and they can transport vibrio cholerae and other organisms (wolf and bye, 1974; smith et al., 1987) . m cells therefore form weak points in the intestinal wall which transport intact antigen and macromolecules to the follicles where they can be processed and be transported to lymph nodes with consequent iga immune responses. this contrasts with the uptake of soluble antigens, which can be taken up by ordinary epithelial cells and transported in the circulation of the villi to be ultimately trapped in the spleen possibly to evoke an igm/ igg response (jeurissen et al., 1985) . understanding of these cellular and molecular characteristics is critical for the design of mucosal vaccines for pathogens that exploit this pathway (neutra, 1998) . no simple, specific histological markers for m cells have been identified for use in routine histological sections although many specialist techniques can be applied (reviewed by gebert et al., 1996) . this has hampered their study because they are still best identified on the basis of their ultrastructural characteristics or ability to capture and transport macromolecules (hamzaoui and pringault, 1998) . the m cell shares tight junctions and desmosomes with adjacent epithelial cells but it has fewer and shorter microvilli than absorptive cells. there is lack of a well organised terminal web. vesicles are abundant in the apical cytoplasm but lysosomes are reduced in number. attenuated cytoplasmic processes may be seen embracing lymphocytes (owen and nemanic, 1978; wolf and bye, 1984) . a cardinal feature is the presence of a intraepithelial pocket on the basal membrane which acts as an internal docking site for lymphocytes, such that they are filled by b and cd4 positive lymphocytes, macrophages and dendritic cells that move between underlying follicles and the epithelium (ermak et al., 1990; farstad et al., 1994) . with care, these attenuated microvilli can be seen at light microscopy in semithin plastic sections (wolf and bye, 1984) . cytochemical analysis has also demonstrated that they can be distinguished at light microscopic level by markedly reduced alkaline phosphatase activity on their terminal surface, in contrast to the dense reaction product produced by other enterocytes (owen and bhalla, 1983) . as the glycoproteins on the surface of m cells are different to those on surrounding cells, they can be selectively stained by labelled lectins, although these patterns are different in different species. for instance, mouse m cells are labelled by the fucose-specific probe ulex europaeus and to some extent anguilla anguilla (giannasca et al., 1994) . mucosal mast cells also appear to be involved in the immunological defence of the gastrointestinal tract. they respond by proliferation, migration and discharge of granules during nematode infestations (miller, 1980) . it has been shown in the rat that mucosal mast cells of the gut differ in several ways from connective tissue mast cells. these differences result in poor preservation of mast cells of the gut if the usual metachromatic staining techniques employed for the demonstration of mast cells in tissue sections (wingren and enerbã¤ck, 1983) . histochemical study suggests that mucosal mast cells differ from connective tissue mast cells by a lower degree of sulphation of glycosaminoglycans and different spatial relationships of protein and glycosaminoglycans in their granules. these cross link following formalin fixation in a way, which is sufficient to prohibit cationic dye binding. wingren and enerbã¤ck (1983) showed that these staining difficulties can be surmounted in tissues fixed in formaldehyde by staining in toluidine blue for prolonged periods of time (5-7 days), a procedure which allows adequate penetration of the toluidine blue molecule. optimal histopathological study of the small intestine is complicated by its length and mucosal fragility. it is important to avoid vigorous washing procedures or any form of excessive manipulation of the unfixed bowel, as artefact caused by washing may confound interpretation of changes induced by xenobiotics (roe, 1984) . combination of artefact due to washing, autolysis and the presence of neutrophils can produce a histological appearance that mimics in vivo damage. although careful visual inspection of the intestine and sampling of appropriate segments for histological examination is usually sufficient for routine examination, various forms of 'swiss roll' techniques are helpful for more complete study. rolling the unfixed, opened rodent intestine around a wooden stick prior to freezing or fixation is one proposed method (moolenbeck and ruitenberg, 1981) , although this method risks undue manipulation of the unfixed tissue. another more versatile technique applicable to rodent, large animal and human intestine can be performed after fixation. the unfixed opened bowel is pinned flat on a cork or board and fixed in a bath of formal saline. after fixation, the full thickness of rodent intestine can be rolled, transfixed by a pin and embedded in paraffin wax. likewise the mucosa of the intestine of large animal species or humans can be rolled after fixation by separating it from the muscularis externa (filipe and branfoot, 1974) . inflammation and ulceration of the mucosa occurs as a result of stress, infection with bacteria, viruses, and infestation by parasites or as a direct result of the effects of xenobiotics or ionising radiation. antimitotic or radiomimetic agents as well as ionising radiation are liable to adversely affect the rapidly dividing cells of the small intestine with resulting breakdown of the mucosal barrier. the ulcerogenic activity of non-steroidal anti-inflammatory drugs may also be expressed in the small bowel mucosa. different agents also act in synergy to enhance damage to the small bowel mucosa. an important example is the effect of drugs that depress the immune system and permit the development of pathological infections by microorganisms of the opportunistic type in the small intestine. the histological features of the inflammatory process in the small intestine are not usually specific for a particular agent. it is important to search for evi-dence of microbiological organisms and viral inclusions, which can indicate the cause of intestinal inflammation and ulceration. associated features in non-ulcerated mucosa such as morphology of the villi, accumulation of abnormal cells or foreign substances and changes in lymphoid cells or blood vessels are also important in the assessment of these changes. a number of organisms including those, which are normal residents of the gastrointestinal tract, can cause inflammatory changes in the intestinal mucosa of laboratory animals. with the notable exception of non-human primates, inflammatory bowel disease caused by microbiological organisms is not usually evident or of concern in most toxicity or carcinogenicity studies. however, when animals are treated with antibiotics, immunosuppressive agents or other drugs, which alter the normal intestinal flora, conditions may favour the proliferation of potentially pathogenic organisms in sufficient quantities to cause overt damage to the mucosa. certain bacterial flora may also act synergistically with intestinal protozoans to produce pathological changes (boorman et al., 1973) . in non-human primate colonies, gastrointestinal disease remains one of the most important causes of death . in contrast to other laboratory species, histological evidence of intestinal infectious disease is relatively common and may confound the interpretation of gastrointestinal alterations occurring in toxicity studies. although the majority of potentially pathogenic organisms affect the primate colon, a number of bacteria, protozoa and metazoa occur in the small intestine. a detailed review of the protozoa and metazoa occurring in the primate gastrointestinal tract has been published (toft, 1982) . owen (1993) has reviewed the parasites of laboratory animals including those of the gastrointestinal tract and the principle reference for identification of metazoa in tissue sections remains that of chitwood and lichtenfels (1973) . organisms which can cause inflammatory disease in the small bowel but which are primarily agents that produce inflammatory disease in large bowel are reviewed under 'colon' (see below). bacillus piliformis is the agent responsible for tyzzer's disease produces intestinal inflammation and ulcers in rats, mice and hamsters. susceptibility of different species and strains to experimental infection with bacillus piliformis is variable. for instance, c57bl, balb mice and f344 rats appear more resistant to infection than outbred syrian hamsters (waggie et al., 1987) . lesions of variable severity usually occur in the ileum but may also extend into the caecum and colon. severe infections are characterised histologically by ulceration of the mucosa, oedema and acute inflammation of the submucosa and muscle coats. muscle may also show focal necrosis. non-ulcerated mucosa is typically infiltrated by polymorphonuclear cells and crypt abscesses form. there is blunting and fusion of villi and reactive hyperplasia and mucin depletion of the overlying epithelium . mucosal lymphoid tissue may also show reactive changes or hyperplasia. filamentous bundles of bacillus piliformis can usually be found in the cytoplasm of both epithelial cells and smooth muscle cells at the edges of necrotic zones. methylene blue, giemsa or silver impregnation techniques such as warthin-starry or levaditi stains are the best stains for the demonstration of these organisms although with care they can be visualised in haematoxylin and eosin stained sections (weisbroth, 1979) . they are gram negative and pas positive. intestinal infections due to salmonella species are relatively common in the mouse but also occur in the hamster and rat . salmonella typhimurium and salmonella enteritidis are regarded as the organisms typical of murine salmonellosis (weisbroth, 1979) . lesions occur in the ileum and may extend into the jejunum and caecum. they are characterised by the presence of ulcers covered by fibrinous exudate and associated with diffuse infiltration of the adjacent mucosa by macrophages, neutrophils and lymphocytes. intact crypt epithelium shows mucin loss and reactive proliferative changes. a characteristic feature is the presence of poorly defined granulomatous lesions composed of macrophages mainly in associated lymphoid tissue or peyer's patches. clostridia species, especially clostridia difficile which cause a pseudomembranous colitis in man and laboratory animals (especially hamsters) may also produce inflammation and ulceration in the terminal ileum with histological features similar to those found in the colon (see following). proliferative ileitis (transmissible ileal hyperplasia) is a striking lesion of hamsters affecting the distal segment of the ileum that is associated with intracellular invasion of the intestine mucosa epithelium by bacteria. the organism has not been cultivated, but has been suggested that it is a campylobacter (jacoby, 1985) although other organisms have been identified in this condition (fox et al., 1993; peace et al., 1994) . although it is characterised by hyperplasia of the ileal mucosa in its early stages, an inflammatory phase intervenes in which there is focal necrosis and haemorrhage of the mucosa, crypt abscesses and infiltration of the lamina propria by acute inflammatory cells and macrophages. the histological features of the associated hyperplasia are characteristic. the mucosa is covered by immature, mucin-depleted pseudostratified hyperchromatic epithelium with mitoses extending to the tips of villi and densely basophilic intracytoplasmic inclusions (jacoby, 1985) . helicobacter jejuni (campylobacter jejuni) is a common cause of diarrhoea in man and may be the causative agent in small intestinal inflammation in laboratory dogs and primates. campylobacter species may be more prevalent in beagle dogs and primates than commonly appreciated. it is important to recognise that dogs colonised with these agents may be susceptible to stress-induced, acute onset gastroenteritis (fox et al., 1988; reed and berridge, 1988) . in man this form of bacterial disease is characterised histologically by mucin-depletion, flat-tening and reactive changes in the small bowel epithelium, crypt abscesses, oedema and infiltration of the mucosa by neutrophils, lymphocytes and plasma cells. similar histological findings have been reported in dogs infected with this organism (prescott and monroe, 1982) . the organisms are gram-negative curved, slender rods, which can be visualised in tissue sections with the warthin-starry stain, a recognised technique for spiral bacteria. the carbol fuchsin technique of gimenez (1964) , first used for the identification of ricketsiae in yolk sac culture and a cresyl fast violet technique are also useful methods for the identification of campylobacter species in paraffin sections (burnett et al., 1987; mcmullen et al., 1987) . another campylobacter-like organism, helicobacter pylori (campylobacter pylori) has been identified in human patients with gastritis, gastric and duodenal ulcers. it is an aetiological or predisposing factor in these forms of gastrointestinal disease (marshall and warren, 1984; rouvroy et al., 1987; richter-dahlfors et al., 1998) . spironucleus muris (hexamitis muris) is also a cause of inflammation in the small bowel of rats, mice and hamsters. during overt infestation, organisms are seen extracellularly in crypts and intervillous spaces associated with blunting of intestinal villi, epithelial degeneration and mucin-depletion, reactive epithelial hyperplasia, oedema and leucocyte infiltration (boorman et al., 1973; wagner et al., 1974) . the morphological expression of damage is accompanied by decreased levels of disaccharidases such as maltase, sucrase and lactase, which may represent a direct effect of the trophozoites on the brush border enzymes (gillon et al., 1982) . trophozoites are characteristically elongated symmetrical flagellates approximately 2-5 âµm wide, 12-20 âµm long. giardia species represent marginally pathogenic flagellates, which are found in the upper gastrointestinal tract. they are opportunistic agents that can become important in both animals and man with depressed immune function. studies in mice infected experimentally with giardia muris have shown that an early response is an increased infiltration of the epithelium by lymphocytes, predominantly t cells (gillon et al., 1982) . this has led to the suggestion that the response to infection by these parasite is primarily a cell-mediated immune reaction similar to experimental graft versus host reaction in the small intestine of mice (mowat and furguson, 1981; gillon et al., 1982) . depression of the immune response by treatment with corticosteroids has been shown not only to increase parasite numbers in murine giardiasis but also cause recrudescence of occult infections (nair et al., 1981) . it has also been suggested that decreased gastric acidity can predispose to giardiasis in man (nalin et al., 1978) . giardia muris (lamblia muris) is sometimes found in the small intestine of rat and mouse but it is common in the hamster (fig. 50) . trophozoites appear in histological sections as crescent-shaped structures on the brush border of the intestinal mucosa or in the adjacent lumen. mucosal lesions may be totally ab-sent or there may be blunting of villi, reactive epithelial hyperplasia. a typical feature is increased infiltration of the epithelium and lamina propria by mononuclear cells (boorman et al., 1973) . another important finding is that lactase, sucrase and maltase levels have been shown to decrease in the small intestine in mice infested with giardia muris (gillon et al., 1982) . giardia lamblia may colonise the small intestine of non-human primates and of man and produce similar morphological appearances to those found with infestations in rodents by giardia muris. other flagellates such as tritrichomonas muris are also be found in the small intestine of mice, rats and hamsters. the coccidian protozoan parasite cryptosporidium represents a striking example of the close relationship between some human and animal diseases. this organism was first recognised in the gastric glands of mice by tyzzer in 1907 and has since been confirmed as a cause of diarrhoea in animals and as a human pathogen. it causes mild diarrhoea in normal subjects especially children and young adults but it can produce severe intestinal disease in immunocompromised individuals (casemore et al., 1985) . histological examination of the small intestine of laboratory animals infested by cryptosporidium reveals the presence of organisms attached to the mucosal surface, often associated, as in man, with other parasites or infections. they are rounded, weakly basophilic structures 1-4 âµm diameter in haematoxylin and eosin stained sections but are strongly basophilic following romanowsky staining. transmission electron microscopy reveals the detailed internal structure of cryptosporidium attached to the microvillous surface of the epithelial cells. the various stages in the life cycle have been visualised by light and electron microscopy. infection starts with ingestion of an oocyst containing four sporozoites, which are probably released by the action of digestive enzymes. these attach themselves to the intestinal mucosa and undertake their life cycle attached to the epithelial cells (casemore et al., 1985) . these organisms have been demonstrated in laboratory species including mice, hamsters, rabbits, dogs and nonhuman primates (cockrell et al., 1974; rehg et al., 1979; toft, 1982; fukushima and helman, 1984; davis and jenkins, 1986) . hymenolepis nana (dwarf tapeworm) and hymenolepis diminuta (rat tapeworn) are described in the intestine of rats, mice, hamsters, non-human primates and man (hsu, 1979) . a variety of other metazoan patients are found in the small intestine of non-human primates, see review by toft (1982) . certain viruses produce inflammatory small bowel changes in mice. mouse hepatitis virus (lethal intestinal virus of infant mice) can cause mucosal epithelial necrosis and inflammation with characteristic compensatory epithelial hyperplasia and the formation of epithelial syncytia . murine rotavirus (epidemic diarrhoea of infant mice) produces swollen enterocytes of small and large bowel with fine cytoplasmic vesiculation with little or no inflammation but dilated lymphatics and vascular congestion. cytoplasmic acidophilic inclusions, 1-4 âµm diameter, are characteristic findings . electron microscopic examination shows cytoplasmic vesicles arising from the rough endoplasmic reticulum, which contain virus particles and electron dense granular material . a mouse adenovirus may also produce large basophilic intranuclear inclusions in the epithelial cells of the small intestine and caecum. k virus infection produces lesions in the jejunal and ileal mucosa of mice. histological features are characterised by mild polymorphonuclear infiltration, with ballooning of occasional endothelial cells within intestinal villi. intranuclear inclusions can be demonstrated in these endothelial cells by light microscopy using appropriate fixation (greenlee, 1985) . a variety of viruses have been isolated from the gastrointestinal tract of nonhuman primates including viruses of man (kalter, 1982) . however, they appear to be relatively infrequent causes of gastrointestinal disease and when disease is caused by these viruses it usually affects other organs. not only can non-steroidal anti-inflammatory agents such as indomethacin and phenylbutazone produce gastric ulceration but also penetrating ulcers of the small bowel of laboratory animals (shriver et al., 1975) . imaging studies with 111 indium-labelled leukocytes in man have also suggested that subclinical intestinal inflammation is associated with long-term therapy with non-steroidal anti-inflammatory drugs (bjarnason et al., 1987) . it has been postulated that indomethacin-induced intestinal ulcers in rats and dogs are produced by a prostaglandin-independent mechanism, different from the manner in which gastric ulceration is induced (tabata and okabe, 1980; whittle, 1981) . conversely, satoh et al. (1981) have suggested that similar mechanisms are responsible for both gastric and intestinal ulceration. they showed that indomethacin-induced gastric ulceration developed in the body mucosa in fasted rats but in the antrum and small intestine in rats given indomethacin 30 minutes after a 1-hour period of refeeding following a 24-hour fast. there was good temporal correlation between the development of intestinal ulcers and inhibition of prostaglandin synthesis (satoh et al., 1981) . the ulcers in the small intestine were morphologically similar to those occurring in the stomach and they were distributed mainly in the mucosa on the mesenteric aspect of the bowel wall. single-dose studies with indomethacin and ibuprofen in rats conducted by suwa et al. (1987) have demonstrated differences between pathology of induced gastric and intestinal damage. gastric damage was superficial, occurred within 6 hours and was fully repaired 2 weeks after dosing. ulcers in the jejunum and ileum reached a maximum area at 48-72 hours after dosing, occurred on the mesenteric border, penetrated through the muscularis mucosa and were accompanied by inflammation and oedema. ulcers were still present 2 weeks later. rainsford (1978) has shown that potent intestinal ulcerogens such as indomethacin inhibit the incorporation of radioactive 35 s sulphate into glycoproteins of the upper intestinal mucosa as well as the stomach of rats. this may decrease the capacity of the mucus in the intestine to act as a buffer for hydrogen ions. indomethacin given orally to dogs in doses of 2.5 mg/kg/day for 1-23 days was also shown to produce intestinal ulceration. these ulcers were deep, punchedout lesions, many of which were lying over peyer's patches (stewart et al., 1980) . some ulcers involved the whole circumference of the small intestine wall. histologically, the ulcers were associated with an intense inflammatory response principally of mononuclear cells, which infiltrated the bowel wall to the serosa particularly adjacent to peyer's patches. it was suggested that this distribution of ulcers was a result of an exaggerated immune response to normal intestinal antigens. these antigens may have been produced by inhibition of suppresser cells in peyer's patches, following depression of prostaglandin synthetase by indomethacin (stewart et al., 1980) . special dye techniques, scanning and transmission electron microscopy have also shown that non-steroidal anti-inflammatory drugs also produce adverse effects on the small intestine mucosa without overt pathological changes being evident by light microscopy (brodie et al., 1970; djaldetti and fishman, 1981) . following administration of aspirin to mice for 5 weeks, shortening and erosion of microvilli and increased numbers of goblet cells were only demonstrated in the duodenum and jejunum by scanning and transmission electron microscopy (djaldetti and fishman, 1981) . morphometric studies of the intestinal mucosa of indomethacin-treated mice have also shown widespread alterations to columnar cells, goblet cells and paneth cells suggesting generalised effects on mitotic activity and crypt loss (ettarh and carr, 1996) . such submicroscopic findings support the idea that non-steroidal anti-inflammatory agents may induce damage to the small intestine more commonly than supposed. although non-steroidal, anti-inflammatory drugs are the best-known drugs with adverse effects on gastrointestinal mucosa, small intestinal inflammation and ulceration can also produced by other agents through different mechanisms. anticancer drugs and other therapeutic agents, which affect cell proliferation, depress the bone marrow or the immune system can also produce intestinal mucosal necrosis, haemorrhage, inflammation and opportunistic gastrointestinal overgrowth when administered to dogs or rodents in high doses (fig. 51 ) bregman et al., 1987) . lymphoid infiltrates without tissue fig. 51 . section of small intestine from a wistar rat treated with a 5-day course of an antiproliferative anticancer drug. the mucosa shows mucus-depletion, loss of villous architecture as well as a reparative and inflammatory response. (he, ã�25.) damage were also reported in the small intestine of rats treated with human recombinant interleukin-2 (anderson and hayes, 1989) . agents of particular toxicological interest are cysteamine, propionitrile and their structural analogues as well as 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (mptp) which are capable of producing ulcers of chronic type in the duodenum of rats and mice (szabo and cho, 1988) . these compounds vary in their ulcerogenic capacity but they are all able to produce ulcers of chronic type with crater formation, granulation tissue and reactive changes in adjacent mucosa in the anterior and posterior wall of the proximal segment of the duodenum of rodents. although these different agents influence gastric acid secretion in different ways, structure-activity relationships suggest that they produce duodenal dysmotility, decrease bicarbonate production and reduce its delivery from the distal to proximal duodenum. these factors decrease the neutralisation of gastric acid in the first part of the duodenum and this may contribute to the development of ulceration (szabo and cho, 1988) . furthermore, these effects can be attenuated or prevented by dopamine agonists or their precursors whereas dopamine antagonists can potentiate their effects. this suggests that the central or peripheral dopamine-mediated actions of these agents may be involved in the pathogenesis of duodenal ulceration (szabo, 1979; . using appropriate fixation and staining procedures, fine granular lipid droplets can be normally visualised in the apical parts of epithelial cells covering the upper third of small intestinal villi. administration of drugs and chemicals may produce an excessive accumulation of lipid through specific effects on lipid metabolism or as part of general cellular toxicity. in the preclinical toxicity studies with 2,6-di-tert-butylamino-3-acetyl-4-methylpyridine (sa h51-055), an inhibitor of glucose transport intended for use as an anti-obesity drug, lipid accumulation occurred in the lamina propria of the small intestinal villi of sprague-dawley rats and guinea pigs but not in dogs or primates (visscher et al., 1980) . after administration of sa h51-055 to rats, there was progressive accumulation of lipid droplets in the epithelial cells over the tips of the duodenal villi demonstrable by osmium tetroxide staining. ultrastructural examination revealed uniform electron-lucent droplets within profiles of the smooth endoplasmic reticulum and golgi apparatus. lipid droplets increased with time, accumulated and coalesced to form large droplets in the lamina propria. larger droplets were phagocytosed by macrophages in the lamina propria but there was no evidence of epithelial damage or necrosis. changes were most pronounced in the duodenum but were also noted to a lesser extent in jejunum and ileum but not in colon or stomach. sequential studies using electron microscopy showed that lipid rapidly accumulated within several hours in the profiles of smooth endoplasmic reticulum and golgi apparatus of the epithelial cells and formed droplets or chylomicra in the intercellular space. the absence of any other subcellular changes or evidence of derangement of protein synthesis suggested that sa h51-055 altered the pathways of lipid resynthesis or transport. this was consistent with the distribution of the lipid in the upper third of the jejunal villous epithelium, a zone most active in lipid absorption, resynthesis and transport (dobbins, 1969) . it was suggested fatty change might have taken place because of alterations in the sugar moiety of chylomicra brought about by interference with glucose transport (visscher et al., 1980) . lipid droplets which stained with oil-red-o in formalin-fixed frozen sections and showed uniform electron density characteristic of neutral lipid were also observed in the epithelial cells and macrophages in the lamina propria of jejunum and duodenum and mesenteric lymph nodes in rats given a synthetic 2'-dodecyl glutaramide ester of erythromycin (gray et al., 1974) . unlike the erythromycin base, rats poorly tolerated this ester. it appeared that the ester was absorbed unhydrolyzed and converted to chylomicron-like droplets, which then accumulated in the macrophages of the lamina propria and local mesenteric lymph nodes, without overt damage to epithelial cells. accumulation of lipid in epithelial cells of intestinal villi has been observed in rats following administration of puromycin (friedman and cardell, 1972) and ethionine (hyams et al., 1966) , agents which have inhibitory effects on protein synthesis. detailed morphological study of the intestinal epithelial cell in rats treated with puromycin have shown that there is concomitant accumulation of lipid with a decrease in the quantity of rough endoplasmic reticulum and golgi membranes (friedman and cardell, 1972) . these changes were in keeping with the concept that lipid accumulates as a result of inhibition of the synthesis of membrane components of the golgi by the rough endoplasmic reticulum which are important for the transport of lipid. in addition to lipid droplets forming as a result of altered lipid metabolism, they may form in the epithelial cells of the intestinal mucosal as a result of a direct toxic effect of the ingested drugs on the small intestinal mucosa. in such instances atrophy of villi and degenerative changes in the epithelial cells may also be observed (see following). the small intestinal mucosa is also one of the many sites at which drug-induced accumulation of polar lipids form laminated structures (myeloid bodies) or crystalloid structures within lysosomes. this form of lipid storage disorder is produced by diverse amphiphilic cationic drugs in both man and laboratory animals probably as a result of drug interaction with polar lipids rendering them difficult to digest (lã¼llmann-rauch, 1979) . species differences in susceptibility and tissue distribution of phospholipid are probably not only related to physiochemical characteristics of the inducing drugs which influences their ability to permeate selective biomembranes and react with different lipids, but also to tissue concentrations of drugs achieved and the ability of organs to metabolise parent drug to less amphiphilic products. in general terms, this form of disorder is characterised by membrane-bound, acid phosphatase-positive cytoplasmic inclusions, which on ultra-structural study are seen as lamellated or crystalloid structures in lysosomes. these appearances are characteristically reversible on cessation of treatment with the inciting agent. an example of this phenomenon occurring in the small intestine is provided by the iodinated amphiphilic drug, amiodarone, which has been used clinically in europe for the past 20 years in the treatment of angina and more recently in the control of supraventricular cardiac arrhythmias. its adverse effects in man are believed to be the result of accumulation of drug in lysosomes particularly in liver, skin and eye (d'amico et al., 1981; shepherd et al., 1987) . when high doses of amiodarone were administered orally to rats and beagle dogs, multilamellated lysosomal inclusion bodies accumulated first in the jejunal mucosa and mesenteric lymph nodes before becoming widely distributed in other organs particularly in the lungs (mazuã© et al., 1984) . in both rats and dogs the small intestinal lesions were characterised by the presence of foamy macrophages with pale finely vacuolated cytoplasm and condensed eccentric nuclei within the lamina propria of the jejunal villi (mazuã© et al., 1984) . mesenteric lymph nodes were also involved early after the onset of treatment. in the dog, jejunal villi were somewhat flattened and widened or showed a variable degree of villous atrophy, most marked in the proximal and middle jejunum (vic et al., 1985) . electron microscopy confirmed the presence of lamellated lysosomal bodies distending macrophages. the early accumulation of foam cells in the jejunal macrophages was probably a reflection of the disposition of drug following oral absorption. although similar lipidosis was seen in many organs following intravenous administration in dogs, more lipidosis was seen in the jejunum after oral dosing. moreover, there were species differences in sensitivity to these changes, baboons being relatively insensitive compared to dogs. fischer 344 rats were very sensitive to these changes compared to sprague-dawley rats and wistar rats were almost completely resistant to lipidosis induced by amiodarone under similar conditions (mazuã© et al., 1984) . similar cytological changes have been reported in cells of some organs in patients treated with amiodarone (d'amico et al., 1981; shepherd et al., 1987) . villous shortening or stunting results when the proliferative activity of the crypt epithelium is reduced or under circumstances in which crypt cell proliferation is insufficient to compensate for increased cell loss as a result of mucosal cell damage. decreased cell proliferation can be seen segments, which are surgically bypassed, or following decreased food intake, parenteral nutrition, hypophysectomy or thyroidectomy bastie et al., 1982) . as adrenergic factors are important in the control of small intestinal epithelial cell division, drugs that alter î± or î² adrenoreceptor activity may influence the proliferative capacity of the epithelium. in mice, increased î±1 or î² receptor stimulation by appropriate agonists (e.g. phenylephrine) diminishes proliferation of crypt cells but proliferation is increased by stimulation of î±2 receptor activity (kennedy et al., 1983) . yohimbine, a î±2-antagonist also reduces cell proliferation in the same animal model. some of the effects of these agents may be mediated by changes in splanchnic blood flow . the detailed morphological study of the small intestinal mucosa in the rat following hypophysectomy by bastie et al. (1982) has shown a reduction in the height of the small intestinal villi associated with reduction in mitoses in the crypt epithelium. the number of goblet cells was shown to fall particularly in the jejunum and the number of paneth cells increase in the ileum. ultrastructural examination showed decreased height of the microvilli of absorptive cells and a lower number of their intracytoplasmic organelles and ribosomes. there were also significant decreases in brush border enzyme activities of alkaline phosphatase, aminopeptidase, maltase and lactase reported about 1 week following hypophysectomy. substances which reduce mitotic activity and therefore lower regenerative capacity of the intestinal epithelium also produce shortening or stunting of small intestinal villi and eventually flattening of the mucosa. a wide variety of anticancer agents and antiviral drugs with radiomimetic properties interfere with cell division in the crypts thereby reducing the number of epithelial cells produced. histologically, the effects of such agents are characterised by blunting, shortening or complete atrophy of villi. mitotic activity is reduced in the crypts and the crypts become dilated and lined by flattened cells. the overlying epithelium loses its normal regular arrangement and cells show pleomorphic nuclei with irregular chromatin patterns. increased numbers of inflammatory cells may infiltrate the lamina propria and epithelium. ulceration, haemorrhage and secondary infection of the gut wall ensue if there is overwhelming cell damage. comparison of the gastrointestinal toxicity expressed by antimitotic anticancer drugs of different classes in rodents, dogs, non-human primates and man have suggested that there is a higher degree of correspondence between effects in man and dog than between man and other species (owens, 1962; schein et al., 1970) . in studies with the antiviral agent acyclovir, a radiomimetic effect was noted in the gastrointestinal tract of dogs at high doses but not rodents (tucker et al., 1983) . another example is the villous atrophy described in rats following treatment with an antibacterial agent ici 17,363. this was believed to arise as a result of both interference with cell division and a direct effect on the surface epithelial cells (murgatroyd, 1980) . the effects of ici 17,363 were characterised by atrophy of villi with dilatation of crypts and atypical features in the crypt epithelium suggestive of an effect on mitotic activity. in addition, vacuolated lipid-laden epithelial cells were observed over the tips of villi accompanied by reductions in the numbers of goblet cells and reduced activity of acid and alkaline phosphatase, esterase, adenosine triphosphatase, glucose-6-phosphatase and succinic dehydrogenase, compatible with a direct adverse effect on superficial mature epithelial cells. a variety of factors stimulate cell proliferation in the small intestinal epithelium. these include enterectomy, increased feeding and stimulation of autonomic nerves. administration of neurotransmitters, thyroxine, growth hormone, corticosteroids, testosterone, gastrin, glucagon and recombinant epidermal growth factor may also stimulate epithelial cell proliferation breider et al., 1996; reindel et al., 1996; vinter-jensen, 1999) . in rats hypothalamic damage, hyperthyroidism, tube feeding, diabetes mellitus and insulin injections have been shown to produce intestinal hyperplasia (mackay et al., 1940; levin and smyth, 1963; jarvis and levin, 1966; forrester, 1972) . most causes of greater cell production lead to increased villous height and mucosal hyperplasia, although intense crypt cell proliferation as a compensatory regenerative response can be associated with villous atrophy (see previous). the compensatory response to the surgically resected or bypassed intestine has been the focus of the most detailed studies of increased cell renewal in the small intestine. partial resection in both rats and man is accompanied by increased villous height and crypt length (hanson et al., 1977) . this is primarily the result of hyperplasia for it has been shown that the numbers of cells per unit length of villus remains unchanged (hanson et al., 1977) but there is an overall increase in the cell population of villus and crypt (hanson and osborne, 1971) . dna/rna ratios also remain largely unaltered . no gross changes in villous shape have been reported after resection and the total number of crypts remains constant. although increased intestinal uptake of substances from the bowel lumen occurs in hypertrophied segments per unit length of bowel, disaccharide and dipeptidase activities are normal or even decreased after resection suggesting a comparative immaturity of cells in the residual mucosa. functional adaptation therefore is achieved by a larger number of cells, the individual absorptive capacity of which is not increased . increased numbers of specific goblet cell populations are also seen in hyperfunctional states. following jejuno-ileal bypass operations in rats, increased numbers of pas-positive goblet cells develop in the villi and crypts of the hyperfunctional segments of the duodenum, jejunum and ileum (olubuyide et al., 1984) . mucin histochemistry using the high-iron diamine and alcian blue techniques have shown that the goblet cells in the hyperfunctional segments contain increased sialomucins in the villi and crypts of the jejunum and ileum but not in the duodenum and increased sulphomucins in the distal ileal segment. sialomucin production may reflect relative cellular immaturity of the more rapidly proliferating cells under these circumstances. however, as sialic acid conveys more viscoelastic properties to mucin, it has been suggested that the goblet cells change following intestinal bypass fulfils a protective function against the increased flow of gastrointestinal contents (olubuyide et al., 1984) . a number of nutritional factors, particularly dietary fibre, can influence the proliferative characteristics of the small bowel mucosa. carefully controlled studies in rats given different forms of dietary fibre have shown that the prolif-erative characteristics of the small intestine can be modified by both the quantity and the quality of the fibre. one study has shown a decrease in the length of villi, crypt cell hyperplasia and shorter transit times in rats fed pectin-supplemented diet but an increase in mucosal growth without alteration in relative differences in crypt and villous length with guar supplementation compared with rats fed fibre-free diet (jacobs, 1983) . another rat study has shown that pectin feeding leads to increased mucosal area and height associated with an increase muscle mass (stark et al., 1996) . these different effects may be the result of differences in solubility, gel formation, water holding capacity, effect on transit time and ion exchange activity or bile acid adsorption of the different fibres. interactions between dietary constituents are complex. for instance, in rats the effects of 2% dietary cholestyramine, a non-absorbable ion exchange resin, on small intestinal histomorphology have be shown to depend on interaction of dietary factors (burkhardt et al., 1998) . administration of an inhibitor of cholesterol biosynthesis, 5î±-cholest-8-(14)en-3î²-ol-15-one, to rats for up to 9 days was also shown to produce enlargement of the small intestine in a way which was morphologically similar to the changes found following intestinal bypass (smith et al., 1989) . the enlargement was most marked in the proximal segment of the small intestine and progressively diminished towards the ileocaecal junction, sparing the stomach, caecum and colon. histological examination and morphometric analysis revealed an increase in smooth muscle mass, lengthening of the villi as well as an increase in the depth and cellular proliferation in the crypts of lieberkuhn without evidence of cell damage or fatty change. like the changes following jejunal bypass procedures, there was also an increase in acid mucosubstances in the goblet cell population overlying the villous mucosa (smith et al., 1989) . the mechanism for this change in the rat was not clear, particularly as intestinal hyperplasia was not seen in baboons treated with this 15-ketosterol for long periods. however, it was suggested that it was an adaptive response, possibly related to inhibition of cholesterol metabolism and cholesterol absorption from the diet, particularly as the laboratory diet employed in the rat study was particularly low in cholesterol. local and systemic changes in hormones and various transmitter substances also influence the number of cells in the small intestinal epithelium. morphometric studies of the small intestinal mucosa in mice following gastrin administration have shown increases in villous area associated with decreases in microvillous area, increased number of goblet cells and paneth cells . studies in which rats were treated with the prolactin-inhibitor, ergocryptine, have shown that the total number of mucous cells and the number staining with alcian blue at ph 1.0 increase in the ileal crypts, possibly as a result of increased synthesis of sulphated mucosubstances (gona, 1981) . in this context, it is of interest that chronic treatment with the rauwolfia neuroleptic, reserpine, causes an increase in the sulphation of goblet cell mucin in the small intestine as demonstrated by alcian blue staining at ph 1.0 and the high iron diamine technique without changes in the goblet cell numbers (park et al., 1987) . agents which affect activity of the sympathetic nervous system can also alter epithelial cell proliferation in the small (and large) intestine. treatment of rats with adrenaline, isoprenaline, phenylephrine, phentolamine and yohimbine all result in decreased mitotic activity of jejunal and colonic crypt cells (tutton and helme, 1974; kennedy et al., 1983) . by contrast, administration of metaraminol, clonidine, propranolol, prazosin and labetolol as well as simultaneous injection of propranolol and adrenaline all resulted in an increased rate of crypt cell proliferation (kennedy et al., 1983) . these results suggest that agents that stimulate î±2 adrenergic receptor activity and those that are î±1-antagonists and î²-adrenergic receptor antagonists increase proliferative activity in the rodent intestinal mucosa. caffeine is also reported to produce an increase in thickness of the intestinal mucosa when administered in high doses to rats (lachance, 1982) . this raises the possibility that intestinal mucosal hyperplasia can be produced by phosphodiesterase inhibition and resultant increases in intracellular camp in a similar way to the hypertrophy induced in salivary tissue. this is supported by recent findings in rats treated for periods of up to 6 months with the inotropic vasodilator, ici 153,110, a phosphodiesterase inhibitor intended for treatment of congestive cardiac failure. administration of high doses not only produced salivary gland hypertrophy but also marked thickening of the small and large intestinal mucosa. this was characterised histologically by an increase in villous length and deepening of intestinal glands, with a relatively unchanged number of epithelial cells per unit length of gland or villus (westwood et al., 1990) . although prostaglandin e analogues produce most of their effects in the stomach, increased thickness of the small intestine characterised by longer villi, deeper crypts and increase in cell size have been reported in rats treated with these agents (levin, 1988) . focal hyperplasia, focal avillous hyperplasia, focal atypical hyperplasia, duodenal plaque, polypoid hyperplasia, polyp -mouse irregular, atypical single or multiple foci of glandular hyperplasia may be found in the small intestinal mucosa of several strains of aged, untreated mice. the lesions are usually located in the first part of the duodenum where they form discrete, raised plaques composed of elongated, irregular or branched glands which replace the normal villous structure of the mucosa (rowlatt and chesterman, 1979) . the glands are lined by hyperchromatic columnar cells, which show marked pseudostratification and proliferative activity. paneth cells and mucinsecreting goblet cells may also be prominent. some glands are cystic and the stroma is fibrous and infiltrated by chronic inflammatory cells. the lesions become pedunculated or polypoid in appearance and show a fibrovascular core that is infiltrated by inflammatory cells. they resemble adenomatous polyps described in man. the cause of these changes in the mouse small intestine is unknown but their prevalence can be altered by dietary fibre and panthothenic acid deficiency as well as by administration of drugs and chemicals (hare and stewart, 1956; seronde, 1965; ito et al., 1981) . in their study of dba mice, hare and stewart (1956) considered that the lesions were not genuine neoplasms since they were composed of a mixture of cell types, which normally populate the mucosa. furthermore, they suggested that the presence of an inflammatory component in the stroma and the fact that the prevalence of these lesions was increased in mice fed a high roughage diet were consistent with the concept that they represent an inflammatory adenomatoid hyperplasia. seronde (1965 seronde ( , 1970 reported these lesions in mice fed purified diets, particularly when deficient in panthothenic acid. panthothenic acid deficiency was also associated with inflammation and deep penetrating chronic ulcers of the duodenum in affected mice, compatible with an inflammatory aetiology of the lesions. an increase in the prevalence of these duodenal changes was described in cd-1 mice treated with the synthetic prostaglandin e1 analogue, misoprostol for 21 months . these authors suggested that the findings posed no real concerns for the safety of patients treated with misoprostol on the grounds that the mouse was unique in this aspect of the response to misoprostol because the mouse had a particular liability to develop such changes in the small intestine. the proliferative lesions were found in a few control cd-1 mice in the same study. in addition, it was also argued that the lesions were neither neoplastic nor preneoplastic in nature . they were not seen in rats treated with misoprostol for 2 years . lesions characterised by such intense proliferative activity may be difficult to distinguish from neoplastic lesion. indeed chronic administration of hydrogen peroxide to c57bl/6j mice in drinking water was not only shown to potentiate the development of a similar type of duodenal hyperplasia but also to produce frankly invasive adenocarcinomas (ito et al., 1981) . anatomically the large intestine is broadly similar in man and laboratory animals but there are significant functional differences. the rat colon is probably one of the best studied of the laboratory animal species because the rat is widely used for experimental work on colon carcinogenesis . as the canine model is popular for oral dosage-form testing, differences in colonic physiology between dog and man may be better understood than between man and many other species (dressman, 1986) . in man, as well as in the non-human primates, the large intestine can be divided anatomically into caecum, appendix, ascending colon, transverse colon, sigmoid colon rectum and anal canal. like the small bowel, the colon comprises mucosa, submucosa, muscularis mucosa and serosa. mucosal plicae are only found in the rectum although plicae semilumaris, formed by folds of the entire thickness of the bowel wall, are found in the colon. the large intestine of the dog resembles that of man more than that of most other domestic species. it is a simplified tubular structure only slightly larger in diameter than the small intestine. the colon of the dog is divided anatomically into ascending, transverse and descending parts, but there is no well-defined sigmoid segment. the caecum in dog is a small diverticulum, similar to that found in other carnivorous species and it communicates directly with the colon. the colon of the rat and mouse is shaped like an inverted v that can be divided into ascending and descending segments. there is no clearly defined transverse colon. a characteristic feature in both rat and mouse is the presence of a curved kidney-shaped caecum. its size is intermediate between the large and anatomically complex caecum of herbivores such as the rabbit and the small caecum of carnivorous species. this probably reflects the omnivorous nature and flexibility of the rat and mouse in their dietary habits, particularly their ability to breakdown cellulose (rerat, 1978) . the caecum of the rat and mouse is a blind pouch from which the colon and ileum exit in close proximity and in which antiperistaltic movements occur. this structure and the presence of bacteria undoubtedly contribute to its ability to function as a fermentation organ in which breakdown of substances can occur in a reasonably controlled milieu (snipes, 1981) . in rat and mouse, the caecum is the site of absorption of many substances including calcium, magnesium, water and electrolytes vitamin k and fatty acids (snipes, 1981) . caecectomy has been shown to decrease digestion of carbohydrates and protein and increase loss of faecal water in these species (ambuhl et al., 1979) . the activity of intestinal microflora in the metabolism of both endogenous and exogenous substances has been demonstrated in the rodent caecum rowland, 1988) . the usual stock diets for rodents contain abundant plant fibre which provides bulk and fermentable carbohydrate for the microbial population in the caecum. rats fed stock diets have been shown to possess high levels of reductive and hydrolytic enzyme activity (e.g. azoreductase, nitroreductase, nitrate reductase, î²-glucosidase and î²-glucuronidase) in their caecal contents compared with rats fed purified fibre-free diets . intestines of germ free animals have thinner lamina propria, lower cell turnover, enlarged caecum, altered metabolism of cholesterol, bilirubin and bile salts and larger amounts of mucin in faeces compared with animals possessing normal gastrointestinal microflora (midtvedt, 1987) . species differences in microflora are also reported. comparative studies with human and rat intestinal microflora have suggested that each population of organisms possesses a degree of autonomous self-regulation and capable of responding quite differently to dietary changes (mallett et al., 1987) . comparative studies have shown large differences in the numbers of facultative anaerobic gram-negative bacteria in the gastro-intestinal tract of mice from three, major specific pathogen free units in australian laboratories. it was shown that these differences could influence the immune system, susceptibility to infection and experimental results (o'rourke et al., 1988) . the colon and caecum in man and laboratory animals are lined by a fairly uniform mucosa devoid of villi. columnar cells of two main types cover the surface epithelium. these are absorptive and mucous cells similar to those found in the small intestine. intestinal glands or crypts of lieberkuhn extend downwards from the surface generally as simple, unbranched tubules lined principally by mucous cells with smaller populations of absorptive, endocrine and undifferentiated cells. the mucosa in man and laboratory animals is not entirely flat but shows a slightly corrugated or uneven pattern, which varies with the particular site within the colon. in histological sections of the colon in man, this corrugated pattern is seen as an anthemion-like structure of crypts reminiscent of a greek architectural feature (filipe and branfoot, 1974) . this is also seen in larger laboratory animal species. in rats and mice the crypts of the caecal mucosa are wider near the lumen than in the crypt base and crypts may be branched, features which may be related to the absorptive function of this zone (snipes, 1981) . the proliferative zone in the large bowel is found in the lower part of the gland and mitotic figures are normally limited to this zone. as in the small bowel, multipotent, undifferentiated stem cells situated in the gland base give rise to the principle cell types which migrate to the cell surface with subsequent differentiation and alteration of their enzyme activities and morphological features (chang and leblond, 1971 ). in studies with mouse aggregation chimaeras in which mosaic cell populations of the intestinal epithelium were localised immunocytochemically, it was demonstrated that the entire epithelium of each adult gland descended from a single progenitor cell (ponder et al., 1985) . the single progenitor may itself give rise to several stem cells responsible for the cell renewal in the complete crypt. absorptive cells are found most commonly in the surface epithelium but also to a lesser extent in the glandular epithelium. they are morphologically similar to those in the small intestine each possessing apical plasma membranes with uniform microvilli and a well-formed glycocalyx. microvilli of absorptive cells have been shown by electron microscopic study to become longer and denser with increasing distance distally in the gastro-intestinal tract. thus, they are longer and more dense in the ileum than in the caecum and least dense and shortest in the colon, perhaps reflecting their relative absorptive capacity (snipes, 1981) . this is reflected at light microscopic level by the less conspicuous brush border in the large intestine compared with that in the small intestine. there are species and regional differences in the glycoconjugates found in the brush border of the large intestine, although they generally contain predominantly acidic mucosubstances. in the mouse and rat, sialomucins with some neutral mucins are found. in hamsters, dogs, non-human primates and man both sulphomucins and sialomucins may be seen in the brush border (sheahan and jervis, 1976) . the glycocalyx is important in the protective function of the colonic mucosa for its disruption by agents such as salicylates has been shown to increase absorption of xenobiotics from the rat rectal mucosa (sithigorngul et al., 1983) . although there has been some debate about the precise nature of the mucous cell populations based on structural studies in mice and rats (chang, 1981; , for practical purposes two principle types of mucous cell can be defined. one of these is the typical goblet cell with cytoplasmic mucous droplets forming a goblet shape, which is found both in glandular and surface epithelium. the other type, the so-called vacuolated cell or mucous cell, is found only in the crypts (chang and leblond, 1971; thomopoulos et al., 1983) . these vacuolated or mucous cells show empty-appearing vacuoles in the cytoplasm which rather than being empty contain abundant sialomucins of a type different from those in goblet cells (spicer, 1965; wetzel et al., 1966) . detailed structural studies and cytochemistry using lectin probes have suggested that these vacuolated cells are able to differentiate into absorptive cells with the cellular apparatus to produce the cell surface glycoconjugates of the glycocalyx (thomopoulos et al., 1983) . sulphomucins, as demonstrated by the high-iron diamine technique (table 2) , generally predominate in the distal colonic segment. in both rat and mouse, goblet cells of the proximal colonic mucosa contain largely sialomucins in the lower parts of the crypts with sulphomucins predominating in the upper parts of the crypt. the distal colon contains largely sulphomucins. the only difference between rat and mouse appears to be the fact that the mouse caecum contains almost exclusively sulphomucins but sulphomucins and sialomucins are found in the rat caecal mucosa. in hamsters, the entire colon contains predominantly sulphomucins. neutral mucins and sulphomucins predominate throughout the dog colon with occasional goblet cells containing sialomucin. in non-human primates, neutral mucins, sialomucins and sulphomucins are seen throughout the colon with sialomucins generally more prominent in the proximal colon and sulphomucins in the distal segment. in man, neutral mucins are found mostly in the caecum. in the caecum and ascending colon, sulphomucins are found in the upper crypts and sialomucins in the crypt base. the converse occurs in the distal colon where sulphomucins predominate in the lower two-thirds of the glands and sialomucins in the upper third of the glands and in the surface epithelium. lectin-labelling (table 1) shows even greater heterogeneity of mucins in the colonic mucosal cell population, probably reflecting differentiation patterns and changes in glycosyltransferase activity as cells migrate upwards (freeman et al., 1980; thomopoulos et al., 1983) . it has been suggested by jass and roberton (1994) that the two principle changes in pathological processes in the human colonic mucins are loss of oacylation substituents at sialic acid c 4 and c 7,8,9 and increased sialylation, neither being specific for neoplasia. the colon, like many other tissues also possess drug metabolising activity, although less than in the liver. it has been shown that the activity of cytochromes p450 involved in hydroxylation of benzo[a]pyrene in microsomes prepared from the colons of sprague-dawley rats retain their activity and responsiveness to inducers better than those in the liver with advancing age (sun and strobel, 1986) . the lamina propria of the large bowel is arranged in a similar way to that of the small bowel. by virtue of the presence of lymphocytes, plasma cells, macrophages and dendritic cells as well as scattered small lymphoid aggregates or patches, it forms an integral and important part in the mucosal immune defence system. most of the lymphocytes in the lamina propria of the human colonic mucosa, like that of the ileum, have been shown to be t cells with helper t cells out numbering the t-suppresser phenotype (hirata et al., 1986; pabst, 1987) . this contrasts with intra-epithelial lymphocytes of the human colonic mucosa which are also t cells but more than 80% of which possess characteristics of the suppresser/cytotoxic phenotype and only 10-20% being helper t cells. this distribution of lymphocyte subsets is seen immunocytochemically in the rat colon using the monoclonal antibodies w3/13, w3/25 and mrc ox8 (see haemopoietic and lymphatic systems, chapter iii). the pan t-cell marker w3/ 13 shows the presence of t lymphocytes in the lamina propria and most of these are labelled by w3/25 demonstrating their cd4 helper phenotype (bland and warren, 1985) . mrc ox8 demonstrates that few lymphocytes in the lamina propria are of suppresser/cytotoxic (cd8) type, which contrasts, with the high proportion of mrc ox8 positive lymphocytes in the colonic epithelium (bland and warren, 1985) . the monoclonal antibodies mrc ox6 and mrc ox17, specific for the rat ia antigen, also label numerous cells with macrophage and dendritic cell morphology in the large bowel of the rat (bland and warren, 1985; martin et al., 1986) . mature, small b lymphocytes are relatively uncommon in the colonic lamina propria of man and laboratory animals. however, the lamina propria contains large numbers of plasma cells, which are mainly of iga type, followed by smaller numbers of igm an igg subtypes (pabst, 1987) . a feature of the colonic mucosa is the presence of lymphoid aggregates also called lymphoid nodules, patches, lymphoid-glandular complexes or microbursa. these are similar to peyer's patches of the small intestine as they are composed principally of lymphoid cells of the b-cell series arranged in follicles with germinal centres with interfollicular and perifollicular zones composed of t cells (pabst, 1987) . they are distributed along the entire length of the colonic mucosa although they are generally smaller than peyer's patches. in sprague-dawley rats, lymphoid aggregates are usually about 5 mm diameter except in the distal colon where they attain sizes of up to 10 mm in maximum diameter (martin et al., 1986) . unlike peyer's patches which are characteristically not associated with crypts or villi, the colonic lymphoid aggregates frequently contain irregular atypical mucosal glands which may enter deeply in the lymphoid tissue and penetrate below the muscularis mucosa both in man and laboratory animals scott, 1982; martin et al., 1986) . in some strains of rat, cells in these glands express the ia antigen, unlike the other parts of the colonic epithelium (martin et al., 1986) . these glandular structures, which are intimately associated with lymphoid tissue, may be important in the immune protection of the colonic mucosa, perhaps by acting as a special local receptor for antigens . it has been proposed that these glandular structures represent sites of predilection for the spread of inflammatory disease to the submucosa by allowing microorganisms to pass through the muscularis mucosa (scott, 1982) . it has also been suggested that they constitute physical weak points in the bowel wall and may play a part in the pathogenesis of diverticular disease of the colon in man . colonic carcinomas induced by dimethylhydrazine in the rat also appear to develop more commonly in the lymphoid aggregates than in other zones (martin et al., 1986) . m cells have been described over the lymphoid aggregates in the caecum of the mouse (owen and nemanic, 1978) , see small intestine. although microorganisms are important causes of inflammatory disease in the large intestine of man and animals, among laboratory animals they are usually only significant problems in non-human primates and hamsters. in the strains of rats and mice and in beagle dogs commonly employed in drug safety evaluation, spontaneous disease of the colon as a result of infectious agents is uncommon. nevertheless, treatment with some therapeutic agents may alter the normal bacterial flora to permit overgrowth of pathogenic organisms or disturb the normal balance between antigens in the lumen or control mechanism to evoke inflammation. inflammation induced by organisms may also confound the histological assessment of drug-induced changes in the colon. ulceration and inflammation of the colon as a direct result of administration of potential therapeutic agents is reported in humans although less commonly that in the small intestine. it has been suggested from studies of the effects of anticancer compounds on neoplastic colonic cells that intestinal cells may possess inherent protective properties in the form of an accelerated efflux pump which can serve to protect them from potentially damaging agents (klohs and steinkampf, 1986) . ulceration and inflammation can be induced by the local application of drugs and vehicles to the rectal mucosa. assessment of these effects in an appropriate animal model is important in the safety evaluation of preparations designed for use in man as rectal suppositories. although inflammatory conditions of the large bowel may possess morphological features typical for some inducing agents, inflammation of the large intestinal mucosa is usually characterised by non-specific histological features. in early or mild inflammation, the surface and glandular mucosa remains intact but shows mucin depletion. this is characterised by reduction in the mucus in goblet cells and increased cytoplasmic basophilia. scattered neutrophils may be seen in the epithelium and adjacent lamina propria. in more severe cases, crypts become filled or distended with acute inflammatory cells (crypt abscesses). the lamina propria is variably hyperaemic and congested and contains increased numbers of mononuclear cells. severe changes are characterised by attenuation or frank erosion of the epithelium and the formation of penetrating ulcers filled with fibrinous exudate and surrounded by intense inflammation, granulation tissue and eventually fibrosis. residual glands may be dilated and lined by flattened epithelium or show reactive changes and mitotic activity. regenerative hyperplasia, which can become florid in chronic ulcerative conditions, is characterised by lengthening, irregularity and cystic dilatation of glands which are often lined by hyperplastic epithelial cells and goblet cells distended with mucin. where ulcerative damage has destroyed glands and supporting stroma, regeneration of glands may not occur in the normal regular fashion and branching of crypts may be evident. clostridium difficile may cause inflammatory changes in the colon of laboratory animals, particularly hamsters, and this may extend into the distal ileum. as in man this form of colitis, often referred to as pseudomembranous colitis, is usually associated with antibiotic therapy. in man it was originally associated with lincomycin and clindamycin therapy but other antibiotics have been implicated. it has been shown that both in man and the hamster experimental model that the enteritis is the result of the toxin produced by clostridium difficile (bartlett et al., 1977; milligan and kelly, 1979) . in man this condition is histologically characterised by the presence of plaques or pseudomembranes on the colonic mucosal surface. the pseudomembrane is composed of mucus, fibrin, blood cells, inflammatory cells and cell debris, which has an appearance of streaming from the underlying mucosa. the mucosa may be partly necrotic or mucosal glands are dilated and lined by flattened or hyperplastic cells. the ileal mucosa may show similar changes (milligan and kelly, 1979) . similar features are observed in the antibiotic-treated hamster although the pseudomembrane is less prominent and it may be distributed more proximally with involvement of the terminal ileum (rehg, 1985) . in the hamster, the condition is characterised histologically by erosion of the colonic epithelium and the variable presence of a pseudomembranous plaque of mucin and cell debris. intact but affected mucosa is thickened with reactive changes accompanied by mucin loss in the epithelium and infiltration of a hyperaemic and oedematous lamina propria and submucosa by polymorphonuclear cells (rehg, 1985) . although most instances of this form of clostridia colitis in the hamster have been associated with antibacterial therapy, it has also been reported in untreated hamsters (rehg and lu, 1982) and those treated with antineoplastic drugs (cudmore et al., 1980) . similar changes have been reported in antibiotic-treated guinea pigs and rabbits (rehg and lu, 1981; rehg and pakes, 1981) . guinea pigs are particularly sensitive to antibiotics especially those active against gram-positive organisms (young et al., 1987) . as in man, these drugs are believed to alter the intestinal flora, permitting overgrowth of clostridium difficile as well as gram-negative organisms, resulting in a severe and frequently fatal enterocolitis. a study of the disposition of ampicillin administered parenterally to guinea pigs showed that this drug was rapidly eliminated from the systemic circulation and excreted in urine and bile, possibly favouring this effect on flora in the colon (young et al., 1987) . citrobacter freundii, a gram-negative, short, plump rod and member of the family of enterobacteriaceae, is the causative agent of naturally occurring transmissible colonic hyperplasia of mice. this agent usually produces a mild or even asymptomatic enteritis in susceptible mouse populations, although it is a cause of rectal prolapse in mice (ediger et al., 1974) . marked strain differences have been noted in mice infected with this organism. nih swiss mice show the most severe histological changes and c57bl/6j mice appear the least affected (barthold et al., 1977) . rats and hamsters seem to be unaffected by citrobacter freundii (barthold et al., 1977) . microscopic changes are found primarily in the descending colon, although proximal segments of the colon and the caecum may also be affected. an important morphological feature is epithelial hyperplasia, which occurs maximally 2-3 weeks after experimental inoculation with citrobacter freundii (barthold et al., 1977) . the colonic glands are elongated and lined by cells that show mucin depletion or loss of goblet cells, considerable immaturity and mitotic activity. the surface epithelium may be covered with numerous coccobacilli, which can be visualised in routine haematoxylin, and eosin stained sections. crypt abscesses, inflammatory cells in the lamina propria, mucosal erosions and ulceration are also features (barthold et al., 1976; . in regressing lesions there is a rebound increase in goblet cells, which are often distended with mucin. the colonic glands may be branched or irregular (barthold et al., 1978) . most laboratory animals are naturally resistant to shigella infections but this is not the case for most non-human primates (takeuchi, 1982) . in infections with shigella, the colon shows a superficial acute inflammatory reaction comprising oedema, congestion, haemorrhage and infiltration by acute inflammatory cells. the surface epithelium shows mucin loss and formation of microulcers where total destruction of the epithelium has occurred. ulcers can extend into the lamina propria but in general terms the inflammatory process remains relatively superficial (takeuchi, 1982) . organisms are also located predominantly in the superficial epithelium. another bacterial infection of the gastrointestinal tract, which affects the colon in primates, is that produced by non-tuberculous mycobacteria . large intestinal lesions are characterised by massive accumulation of epitheloid macrophages in the lamina propria, which may extend into the submucosa and muscular layers and along lymphatics to involve mesenteric lymph nodes. small intestinal lesions may also occur, characterised by the presence of similar large macrophages in the lamina propria of villus tips. superficial ulcers may occur in severely affected segments of intestine . acid-fast bacteria are typically found within macrophages. other organs, including spleen, liver, bone marrow and lungs, may also be involved by focal accumulations of bacteria-laden macrophages or occasionally discrete granulomas with multinucleated giant cells. numerous protozoa and metazoa have been described as inhabitants of the caecum and colon of the non-human primate (toft, 1982) . far fewer are observed in the usual laboratory rodents and beagle dogs. amoebiasis caused by entamoeba histolytica is a widespread disease among non-human primates. it is characterised histologically by the presence of necrotizing ulcers, which reach the muscularis mucosa to form typical flank-shaped ulcers containing or surrounded by trophozoites. extensive haemorrhage may be seen as well as an inflammatory infiltrate composed of neutrophils and mononuclear cells (toft, 1982) . the ciliate, balantidium coli, can also cause an ulcerative process in the colon of primates, characterised by ulcers which extend down to the muscularis mucosa accompanied by lymphocytic infiltrate and balantidium coli trophozoites of up to 150 âµm in greatest diameter (toft, 1982) . a variety of metazoan parasites can be observed in the primate colon and usually can be reasonably well identified in tissue sections (see review by chitwood and lichtenfels, 1973) . the nematode of species strongyloides is an important parasite, which may be observed in the intestinal mucosa of primates. oxyurids commonly known as pinworms are essentially innocuous parasites seen in man, non-human primates and rodents. enterobius vermicularis is found in the large intestine and appendix of man and non-human primates, syphacia muris and syphacia obvelata in rodents. oesophagostomum species (nodular worms) are especially common nematode parasites of non-human primates forming characteristic nodules up to 5 mm diameter most frequently on the serosal surface of the large intestine and caecum and adjoining mesentery as well as in other sites in the peritoneal cavity. histologically, the nodules are composed of parasite cell debris surrounded by fibrous tissue and a variable mantle of chronic inflammatory cells and occasional foreign-body giant cells. they are frequently found in close proximity to small arteries and arterioles in the submucosa and subserosa of the colon and may be associated with a local granulomatous arteritis (lumb et al., 1985) . the inflammatory process may spread to surrounding or draining tissues, particularly if nodules rupture. mild periportal hepatic chronic inflammation is sometimes associated with the presence of this parasite in the mesentery, which may confound interpretation of drug-induced hepatic changes in the non-human primate. although the stomach and to a certain extent the small intestine remain the primary sites of predilection for the ulcerogenic action of non-steroidal anti-inflammatory, the colonic mucosa may become involved under certain conditions. less common complications of non-steroidal anti-inflammatory drugs and potassium chloride therapy in humans are colonic strictures. it appears that nonsteroidal anti-inflammatory drugs produce local inflammation followed by focal scarring of the submucosa with constriction and formation of a mucosal diaphragm whereas potassium causes segmental full thickness scarring and constriction (fellows et al., 1992; haque et al., 1992; van velzen et al., 1996; wolfe et al., 1999) . another form of induced colon damage has been reported in children with cystic fibrosis, the majority of who take high strength pancreatic-enzyme supplements to control malabsorption (smyth et al., 1994; fitzsimmons et al., 1997) . this condition has distinctive pathological features. there is involvement a long segment of ascending colon by a fusiform stenosis primarily as a result of submucosal thickening by deposition of mature collagen. the mucosa appears relatively spared but shows some ulceration and reparative changes (van velzen et al., 1996) . although it has been suggested that the changes may have been linked to the methylacrylate copolymer used for enteric coating of the high-strength preparations, a case-control study showed a strong relation between high daily doses of the enzyme supplements, accentuated by more recent availability of high dose forms (fitzsimmons et al., 1997) . in view of their usage for over 50 years, preclinical data on this material is scarce. colonic damage can be induced experimentally by administration of therapeutic agents. dogs administered 2.5 mg/kg indomethecin orally each day for periods of up to 23 days developed not only gastric and small intestinal ulceration but also scattered haemorrhagic erosions in the colon and rectum. histologically, these lesions were characterised by loss of superficial epithelial cells, mucus-depletion of glandular epithelium, crypt abscesses, frequently with acute inflammation in adjacent lymphoid aggregates in the submucosa (stewart et al., 1980) . an example of chemically induced colitis of relevance to safety assessment of therapeutic agents is that induced by degraded carrageenans or synthetic sulphated dextrans. carrageenans are a heterogeneous group of sulphated polysaccharides composed mainly of long chains of d-galactose subunits (d-galactan) derived from red seaweed species which are widely used as food emulsifiers, stabilisers, thickeners and gelling agents (ishioka et al., 1987) . when carrageenans are degraded by acid hydrolysis into smaller molecular weight fragments of about 20,000-40,000 and administered orally in high doses (e.g. 10% of diet) to rats, mice, guinea pigs, rabbits and rhesus monkeys, colitis results (sharratt et al., 1970; marcus and watt, 1971; benitz et al., 1973; fath et al., 1984; kitano et al., 1986) . similarly, colitis has been induced in rats following administration of a 5% dietary admixture of dextran sulphate sodium, a sulphated polymer of glucose (a d-glucose) of molecular weight of 54,000 (hirono et al., 1981) and a very high molecular weight d-glucan, amylopectin sulphate (ishioka et al., 1987) . although histological features of this form of induced colitis vary between study, species and strain, the colitis is generally characterised mucosal ulceration mainly in the caecum but also in the distal ileum, distal colon and rectum. there is mucus-depletion with variable acute inflammatory infiltrate of the in-tact epithelium, crypt abscesses, inflammatory infiltrate of the lamina propria with oedema, hyperaemia and even vascular thrombosis in the submucosa (hirono et al., 1981; fath et al., 1984) . increased proliferative activity of the mucosa is confirmed by an increase in the tritiated thymidine index compared with controls (fath et al., 1984) . in the caecum of rats, ulcers are linear but often circulating the entire circumference of the intestinal wall with subsequent scarring and stricture formation (oohashi et al., 1981) . ulcerating lesions in the rectum and at the anal margin are associated with squamous metaplasia. both the squamous metaplasia and the regenerative hyperplasia of the columnar epithelium have been shown to progress even after cessation of treatment (oohashi et al., 1981) . foamy macrophages containing metachromatic material, presumably polysaccharide, are also seen in the lamina propria, submucosa, regional lymph nodes, liver and spleen (hirono et al., 1981; oohashi et al., 1981) . the cause of this colitis is unclear. low dose levels, which may be expected to mimic human exposure, do not produce colitis. dextrans, carrageenans and other polysaccharides of molecular weights outside the range 20,000-60,000 tend not to incite colitis. an exception to this is the agent amylopectin sulphate, which has a far higher molecular weight. however, amylopectin is composed of polysaccharide chains, which can be degraded by amylase, and therefore smaller molecular weight fragments may be formed in vivo (ishioka et al., 1987) . it has been suggested that colonic disease produced by these agents is in some way linked to induced changes in intestinal microflora (marcus and watt, 1971) although the evidence for this is conflicting (ishioka et al., 1987) . a recent study in guinea pigs and rats using permeability markers of different molecular weights has suggested that degraded carrageenans enhance intestinal permeability in the absence of overt ulceration (delahunty et al., 1987) . it was therefore proposed that carrageenan-induced colitis could be the result of increased intestinal permeability to antigenic or inflammatory substances normally resident in the large intestine. moreover, long-term administration of high doses of these agents to rats leads to the development of colorectal cancer despite their being devoid of any mutagenic potential (see below). the only obvious features, which are common to a number of these non-genotoxic agents, is chronic inflammation and increased proliferative activity. the rectal administration of therapeutic agents and surfactants may also induce similar ulcerative and inflammatory changes. chemical colitis resembling pseudomembranous colitis has been reported in man as a result of chemical cleaning agents accidentally induced by endoscopic examination (jonas et al., 1988) . cellular degeneration, with loss of mitotic activity and mucin depletion can also occur in the colon following treatment with antimitotic drugs. lymphoid infiltrates without tissue damage were reported in the large bowel of rats treated with human recombinant interleukin-2 (anderson & hayes 1989) . melanosis coli is a well-described phenomenon in man associated with chronic ingestion of anthraquinone purgatives. it is considered to be due to the excessive accumulation of lipofuscin-like pigment in the macrophages of the colonic lamina propria (schrodt, 1963; ghadially and parry, 1966; steer and colin-jones, 1975) . this pigment probably originates from organelles of epithelial cells or macrophages, which are damaged by treatment. similar morphological changes have been induced in laboratory animals (guinea pigs) by treatment with anthraquinones (walker et al., 1988) . as a result of these animal studies, walker et al. (1988) suggested that the primary process is a treatment-induced increase in apoptotic bodies in the surface colonic epithelium that are phagocytosed by intraepithelial macrophages and transported to the lamina propria. lipofuscin and iron pigment is occasionally observed in the lamina propria of untreated rodents, notably hamsters, presumably a result of ageing, previous inflammatory processes and haemorrhage. as in other glandular epithelial tissues, hyperplasia may be focal or diffuse with or without atypical cellular features. the term used for hyperplasia with atypical features is atypical hyperplasia in the iarc classification (mohr, 1997) although others use the term dysplasia. like small intestine, cell proliferation in the large intestinal mucosa can be stimulated by a variety of different factors although these functional adaptive responses have been less well studied. physical stimulation by distension or increased dietary bulk is sufficient to initiate hyperplasia including thickening of the muscle coats (dowling et al., 1967; stragand and hagemann, 1977) . one of the most clearly documented forms of compensatory hyperplasia is that which occurs as a response to surgical resection or bypass of a segment of the colon. following resection of a segment of colon in rats, barkla and tutton (1985) showed that the remaining proximal segment of the right side of the colon showed an increase in the thickness of the mucosa and the muscularis externa as well as enlargement of lymphoid aggregates. histologically, the mucosa of the right side of the colon was uniformly thickened showing accentuated folds, elongated mucosal glands with increased height of the surface columnar cells. the changes were most marked up to 30 days following surgery but were less pronounced after 72 days. there was also a significant increase in the mitotic index in the proximal segment at 7 days although at 14 days and later the mitotic index had returned to normal. the distal, down-stream segment showed little or no morphological change but rather a long-lived increase in mitotic activity. it was suggested that these differences were related to the different embryological origin of the segments (barkla and tutton, 1985) . a similar form of uniform colonic hyperplasia affecting principally the caecal and right-sided colonic mucosa also occurs in rats following oral administration of sulphated dextrans of molecular weight of approximately 400,000 (figs 52 and 53). oral administration of a wide range of compounds such as raw and chemically modified starches, various dietary fibres, caramels, sugar alcohols (lactitol, sorbitol, mannitol, xylilol), lactose, a synthetic polydextrose, polyethylene glycol and magnesium sulphate to rats or hamsters has also been linked to an increase caecal size and colonic mucosal hyperplasia (leegwater et al., 1974; roe and bã¤r, 1985; newberne et al., 1988; stark et al., 1996) . the characteristic histological appearance of the caecum following administration of these agents is lengthening of the mucosal glands which are lined by epithelium composed of increased numbers of enlarged epithelial cells (i.e. hypertrophy and hyperplasia) showing increased proliferative activity and more rapid incorporation of tritiated thymidine (newberne et al., 1988) . in addition, mucosal and submucosal oedema has been reported in association with the administration of lactose and increased mucosal lymphoid aggregates following lactose or xylitol feeding (newberne et al., 1988) . changes in the colon due to fibre are complex. morphometric analysis has shown that changes to the mucosa depend on the fibre type (stark et al., 1996) . there may also be an interaction between dietary fibre content and colonic microflora that influences mucosal growth, although the mechanism is unclear (whiteley et al., 1998) . hypertrophy of the muscularis external is also reported in rats fed high fibre diets (stark et al., 1996) . as the increase in caecal size and mucosal hypertrophy appears generally related to the osmotic activity of the caecal contents in rodents treated with these agents, it has been postulated that the changes represent a physiological adaptation to increased osmotic forces, irrespective of the contributing compounds (leegwater et al., 1974) . treatment of rodents with antibiotics also causes caecal enlargement or dilatation without significant histopathological changes, probably as a result of changes in caecal microflora. it has been suggested that the enlargement relates to accumulation of urea as a result of inhibition of bacterial ureases (juhr and ladeburg, 1988) . however, histochemical studies of the intestinal mucosa of rats treated with neomycin have also shown treatment-related reductions in activities of nad tetrazolium reductase, succinate dehydrogenase, esterase, alkaline and acid phosphatase in the distal ileum, suggesting that some antibiotics also posses the potential to directly influence absorption and metabolic functions of mucosal cells (van leeuwen et al., 1986) . long-term administration of 16,16-dimethyl prostaglandin e2 to rats also produced thickening of the proximal colonic mucosa, although this was less fig. 53 . similar area of colonic mucosa to that seen in fig. 52 at the same magnification but from a rat treated with 10% dextran (molecular weight 500,000) in the diet for 2 weeks. this shows diffuse hyperplasia of the mucosa with elongation of colonic glands that are lined by relatively normal epithelial cells with abundant mucin and prominent vesicular nuclei. (he, ã�40.) marked than in the stomach and duodenum (reinhart et al., 1983 ) and similar changes have been reported in rats treated with other prostaglandin e analogues (levin, 1988) . as in the small intestine, administration of epidermal growth factor to rats and cynomolgus monkeys induces hyperplasia of the colonic mucosa characterised histologically by hyperplasia and increased mitotic activity of crypt cells and reduction in goblet cell numbers with an increase in crypt depth and a slight increase in the numbers of infiltrating neutrophils (breider et al., 1996; reindel et al., 1996) . in common with other epithelial surfaces, atypical hyperplasia is associated with the development of colonic cancer in both man and laboratory animals. the early alterations observed in rats treated with colonic carcinogens are similar to those found in the immediate vicinity of human colorectal carcinomas. the changes are characterised by lengthening, dilatation and branching of glands. the epithelium lining these glands shows mucous cell hyperplasia (goblet cell hyperplasia, see fig. 54 ), goblet cells containing predominantly sialomucin instead of the normal sulphomucin (filipe and branfoot, 1974; filipe, 1975; olubuyide et al., 1985) . despite mucin alterations, activities of glucose-6-phosphatase, glucose-6-phosphate dehydrogenase and glyfig. 54 . section from the colon of an aged hamster from a colony that developed intestinal inflammation and neoplasia spontaneously. this shows focal mucous cell hyperplasia characterised by enlargement and lengthening of the colonic glands with lining cells replete with mucins. (he, ã�25.) ceraldehyde-3-phosphate dehydrogenase were shown to be normal in this epithelium in rats treated with 1,2-dimethylhydrazine (mayer et al., 1987) . in man, this form of hyperplastic mucosa associated with cancer, has been termed 'transitional mucosa' (filipe and branfoot, 1974) . as lesions become more atypical, these dilated, branched glands become more complex and lined by epithelium that shows increasing pseudostratification and vesicular cell nuclei. for example in rats treated with the carcinogens azoxymethane or 1,2-dimethylhydrazine, crypts show diminution of mucus secretion, increased cytoplasmic basophilia, prominent, rounded or enlarged nuclei which show variable degrees of pseudostratification and which eventually develop into frankly invasive glands . in contrast to goblet cell hyperplasia, these atypical zones show increased activity of glucose-6-phosphate, glucose-6-phosphate dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase (mayer et al., 1987) . described similar alterations in rats treated with azoxymethane and the non-genotoxic agent dextran sulphate. these authors also demonstrated that these atypical foci could be identified by low power microscopy as aberrant crypt foci by translumination of the whole mounts of the fixed mucosa stained with methylene blue. note: some compounds may induce qualitative and quantitative changes in mucin content in the colonic mucosa without marked morphological alterations. an example of this phenomenon was described in rats treated with reserpine for 7 days. the colonic mucosa showed an increase in sulphomucin (high-iron diamine positive) containing goblet cells in the surface epithelium (park et al., 1987) . adenomas and adenocarcinomas of the small and large intestine are infrequent spontaneous neoplasms in laboratory animals compared with man where colorectal carcinoma is one of the most prevalent neoplasms in the western world. adenomas and adenocarcinomas probably occur spontaneously in older dogs more than in any other animal species and as in man these are located most frequently in the distal colon and rectum (lingeman and garner, 1972) . in nonhuman primates glandular neoplasms of the intestine occur with increasing age in the ileum and in the colon with a predilection for the zones near the ileocaecal valve (depaoli and mcclure, 1982) . in rats and mice, spontaneous intestinal neoplasms are uncommon although adenocarcinomas are occasionally observed in the ileum or colon in mice and rats used in chronic toxicity studies and carcinogenicity bioassays (burn et al., 1966; wells, 1971; maeda et al., 1985; greaves and faccini, 1992; zwicker et al., 1992) . most of these arise in the small intestine and appear to originate in the distal part of the small intestine, caecum and right side of the colon. they may produce metastases, mostly to liver and lungs. in a review of spontaneous adeno-carcinomas developing over a 17-year period in wistar rats, vandenberghe et al. (1985) identified 17 adenocarcinomas, all in ascending colon. in 15 of these cases there appeared to be an intimate relationship with campylobacter-like organisms together with diverticulae and chronic inflammation. these authors suggested that the organisms and the associated inflammation were involved in the pathogenesis of these cancers. in view of the importance of colon cancer in humans, a number of new genetic mouse models predisposed to colon cancer have been developed over recent years (heyer et al., 1999) . some hamster colonies, liable to develop inflammatory bowel disease (see above), also have a high incidence of small and large intestinal polyps, adenomas and adenocarcinomas (fortner, 1957; van hoosier et al., 1971; personal observations) . poorly differentiated carcinomas may infiltrate local lymph nodes and it may be difficult to locate the primary neoplasm. polyps are predominantly adenomatous in nature although inflammatory or regenerative polyps are observed (van hoosier et al., 1971) . adenocarcinomas are induced experimentally in the rodent intestine by the carcinogens 1,2-dimethylhydrazine and azoxymethane. the histogenesis of these induced carcinomas has been extensively studied and it is generally accepted that they resemble human colorectal cancer (ward, 1974; freeman, 1983) . however, there are differences between reported studies. some have shown that these experimental carcinomas arise from pre-existing adenomas consistent with the 'adenoma-carcinoma sequence' theory (ward, 1974; ward et al., 1977) . others suggest that they arise 'de novo' from altered mucosa as microinvasive carcinomas (sunter et al., 1978; maskens and dujardin-loits, 1981; rubio et al., 1986) . these differences may be partly the result of different dosage schedules. rubio et al. (1988) have shown that a single dose of 1,2-dimethyhydrazine produces non-polypoid, micro-invasive carcinomas, particularly in the mucosa overlying lymphoid aggregates, whereas in their earlier studies using multiple doses in the same strain of rat, an adenoma-carcinoma sequence was more evident. in addition, there are undoubtedly species and strain differences in the response to these agents. teague et al. (1981) demonstrated clear differences in the distribution and both macroscopic and histological types of adenomas and adenocarcinomas between three different inbred strains of rat given a similar dosage regimen of 1,2-dimethylhydrazine. in general, many of these carcinomas develop in the distal colonic segments similar to the distribution of human colorectal cancer, although tumours also develop in the proximal colon and in the ileum in rats treated with this agent (ward, 1974; teague et al., 1981) . neoplasms occurring in the rat colon following administration of high doses of degraded carrageenans and sulphated dextran also commonly occur in the distal colon and rectum and are commonly polypoid adenomas and adenocarcinomas (hirono et al., 1981; oohashi et al., 1981; ishioka et al., 1987) . however, in these models adenomas and adenocarcinomas also occur in the caecum and proximal colon and squamous carcinomas are sometimes seen in association with squamous metaplasia at the colorectal junction (oohashi et al., 1981) . the pathogenesis of neoplasms induced by carrageenans and dextrans remains unexplained. although they are biologically active compounds, they are non-mutagenic in the usual short-term tests (ishioka et al., 1987) . it has been suggested that carrageenans act as tumour promoters as they potentiate the appearance of carcinomas in rats treated with standard intestinal carcinogens hirono et al., 1981) . conversely it has been proposed that these agents are tumour initiators based on the development of carcinomas in rats treated with degraded carrageenans for only 2 months (oohashi et al., 1981) . however, despite only a short period of treatment, inflammation, regenerative changes and squamous metaplasia persisted throughout a period of 18 months after treatment was withdrawn before development of cancer in these rats. the only consistent association of carrageenans with development of carcinoma in rats is that of chronic inflammation and increased cell proliferation. although dose levels needed to produce inflammation are far higher than any exposure likely to be achieved in man, interpretation of this inflammation-cancer sequence in rats remains a challenge in safety assessment for similar xenobiotics. this situation is interest in view of the unquestionable risk of carcinogenesis in ulcerative colitis in man (riddell et al., 1983) . a similar range of neoplasms can be defined histologically in both human and experimental pathology. it is appropriate, to use the same classification for all species including man. lingeman and garner (1972) who reviewed a range of tumours from domestic and laboratory animals were able to employ the classification for human gastrointestinal neoplasms. a similar approach has been used in the iarc classification of rat intestinal tumours (mohr, 1997) . this classification can be summarised as follows: these represent localised, sessile or polypoid neoplasms composed of proliferating tubular glands, which show varying degrees of nuclear hyperchromatism, pseudostratification and cellular pleomorphism. a useful scheme for grading the carcinogenic potential of hyperplastic mucosa and adenomatous polyps in man based on the degree of epithelial pseudostratification has been proposed by kozuka (1975) . although experimental neoplasms may not always show the full spectrum of these changes reported in man, this grading provides a useful baseline concept for the assessment of these non-invasive proliferative lesions. with increased nuclear pseudostratification and atypical branching of the glandular structures of these polyps becomes more prominent. if neoplastic cells or glands are seen in the stroma of the stalk or base the diagnosis of carcinoma is made. villous adenoma is a form of adenoma in which the epithelial proliferation takes the form of elongated villi with a sparse fibrovascular stroma. they can be graded in a similar way to other adenomas. these are glandular neoplasms of variable differentiation, sometimes originating in adenomatous polyps or villous adenomas but which show infiltration of the intestinal wall, i.e. beyond the boundary of the muscularis mucosa. squamous carcinomas also occur in the anorectal zone but are similar to those which occur in squamous epithelium elsewhere. similarly, mesenchymal neoplasms also are found in the small and large intestinal wall (see integumentary system, chapter i). light and electron microscopical studies of parietal cells before and one year after proximal vagotomy in duodenal ulcer patients six-month repeated oral toxicity study of nk-104 in rats response of the non-human primate to polychlorinated biphenyl exposure cell number as a measure of distribution and renewal of epithelial cells in the small intestine of growing and adult rats induction of early lesions in the forestomach of rats by 3-tert-butyl-4-hydroxyamisole (bha) effects of caecetomy in the young adult female rat on digestibility of food offered and libitum and in restricted amounts toxicity of human recombinant interleukin-2 in rats. pathologic changes are characterized by marked lymphocytic and eosinophilic proliferation and multisystem involvement effects of cimetidine, cimetidine plus nitrite, and nitrosocimetidine on tumors in mice following transplancental chronic lifetime exposure age-associated lesions in barrier-reared male sprague-dawley rats: a comparison between hap: (sd) and crl:cobs[r] cd[r] (sd) stocks expression of cd15 in normal and metaplastic paneth cells of the digestive tract an epizootic of klebsiella aerogenes infection in laboratory rats correlation of quantitative changes of gastric mucosal glycoproteins with aspirin-induced gastric damage in rats long-term comparative effect cholecystokinin and gastrin on mouse stomach, antrum, intestine, and exocrine pancreas the effect of 6-hydro-xydopamine on rat salivary glands and on their response to isoproterenol biologically active peptides in submandibular glands the alimentary system proliferative and morphologic changes in rat colon following bypass surgery cyclosporin and gingival overgrowth the etiology of transmissible murine colonic hyperplasia dietary, bacterial, and host genetic interactions in the pathogenesis of transmissible murine colonic hyperplasia transmissible murine colonic hyperplasia mouse hepatitis virus infection, intestine, mouse murine rotavirus infection, intestine, mouse adenovirus infection, intestine, mouse clindamycinassociated colitis due to a toxin-producing species of clostridium in hamsters antibiotic-associated pseudomembranous colitis due to toxin-producing clostridia histological variations jejunal and ileal mucosa on days 8 and 15 after hypophysectomy in rat: morphometric analysis in light and electron microscopy comparative study of histological and kinetic variations of the digestive mucosa and pancreatic parenchyma after hypophysectomy in the rat effect of drugs on rats exposed to cold-restraint stress adverse effects of anticonvulsant drugs: a critical review formation of n-mono-nitrosopiperazine in the stomach and its secretion in the urine after oral intake of piperazine intestinal effects of carrageenans in the rhesus monkey (macaca mulatta) the cell surface: components and configurations mã©nã©trier's disease. serial morphological, secretory, and serological observations structure, biosynthesis and functions of glycoprotein glycans. experientia pathology of the forestomach in rats treated for 1 year with a new histamine h2-receptor antagonist, sk&f 93479 trihydrochloride fundic mucosal ecl cell hyperplasia and carcinoids in rodents following chronic administration of the histamine h2-receptor antagonist sk&f 93479 and other antisecretory agents gastric ecl-cell hyperplasia and carinoids in rodents following chronic administration of the h2 antagonist sk&f 93479 and oxmetidine and omeprazole gastric regulatory peptides in rats with reduced acid secretion non-steroidal anti-inflammation in humans immunohistologic analysis of the t-cell and macrophage infiltrate in 1,2-dimethylhydrazine-induced colon tumors in the rat turnover of brush-border glycoproteins in human intestinal absorptive cells: do lysosomes have regulatory function? alterations in gastric mucosal morphology induced by long-term treatment with omeprazole in rats the effect of aging on the rat submandibular gland. an ultrastructural, cytochemical and biochemical study drug-induced esophageal strictures cytochrome p450 of small intestinal epithelial cells. immunocytochemical characterization of the increase in cytochrome p450 caused by phenobarbital synergistic role of intestinal flagellates and normal intestinal bacteria in a post-weaning mortality of mice medication-induced oesophageal injury. survey of the literature resistance to starvation in albino rats fed from weaning on diets containing from 0 to 81% of protein as casein diseases of the kidney the human gastrointestinal secretory immune system in health and disease clinical aspects: an overview single-dose and multiple-dose intravenous toxicity studies of bmy-25282 in rats cellular hyperplasia in rats following continuous intravenous infusion of recombinant human epidermal growth factor adrenergic mechanisms responsible for submandibular salivary glandular hypertrophy in the rat aspirin: intestinal damage in rats gastrointestinal mucosal lesions: a drug formulation problem intestinal t lymphocytes of different rats strains in immunotoxicity effects of propionic acid and pravastatin on hmg-coa reductase activity in relation to forestomach lesions in the rat famotidine: summary of preclinical safety assessment effects of cholestyramine and diet on small intestinal histomorphology in rats spontaneous carcinoma of the colon of the rat cresyl fast violet staining method for campylobacter-like organisms pigmentation of the jawbone and teeth secondary to minocycline hydrochloride therapy genetic ablation of parietal cells in transgenic mice: a new model for analyzing cell lineage relationships in the gastric mucosa esophageal lesions caused by orally administered drugs. an experimental study in the cat tetracycline induced esophageal ulcers. a clinical and experimental study diagnosis of silodacryoadenitis virus infection of rats in a virulent enzootic outbreak cryptosporidium species a 'new' human pathogen thyroxine accelerates the differentiation of granular convoluted tubule cells and the appearance of epidermal growth factor in the submandibular gland of the neonatal mouse. a fine structural immunocytochemical study renewal of the epithelium in the descending colon of the mouse. i. presence of three cell populations: vaculated-columnar, mucous and argentaffin two types of mucous cells in the colon crypt origin, differentiation and renewal of the four main epithelial cell types in the mouse small intestine. iii entero-endocrine cells cytology of the canine oral papilloma uremic gastropathy in the dog intestinal absorption and metabolism of xenobiotics in laboratory animals parasitological review. identification of parasitic metazoa in tissue section spontaneous basophilic hypertrophic foci of the parotid glands in rats and mice effects of housing conditions on food intake, body weight and spontaneous lesions in mice. a review of the literature and results of an 18-month study cryptosporidiosis in the intestines of rhesus monkeys (macaca mulatta) isolation of a mouse submaxillary gland protein accelerating incisor eruption and eyelid opening in the newborn animal post marketing surveillance of the safety of cimetidine: mortality during second, third, and fourth years of follow-up hyperplastic gastropathy in the rat due to taenia taeniaeformis infection: parabiotic transfer and hypergastrinaemia. gastroenterology number, size and distribution of peyer's patches in the human small intestine a model for gastric cancer epidemiology helicobacter pylori infection, a paradigm for chronic mucosal inflammation: pathogenesis and implications for eradication and prevention the effects of vagotomy on the gastric mucosa of the rat the effect of prolonged administration of large doses of cimetidine on the gastric mucosa of rats effect of short-and long-term feeding of omeprazole on rat gastric endocrine cells clostridial enterocolitis produced by antineoplastic agents in hamsters and humans odontogenic tumours in fischer rats the histo-chemical demonstration of o-acylated sialic acid in gastrointestinal mucins: their association with the potassium hydroxide-periodic acid-schiff effect a new histochemical method for the identification and visualization of both side chain acylated and non-acylated sialic acids amiodarone keratopathy, drug-induced lipid storage disease salivary gland components involved in the formation of squamous metaplasia gastric mucosal injury by fatty and acetylsalicylic acids cryptosporidosis and proliferative ileitis in a hamster specificity of twelve lectins towards oligosaccharides and glycopeptides related to n-glycosylproteins intestinal permeability changes in rodents. a possible mechanism for degraded carageenan-induced colitis adrenal corticosteroids cause gastrin cell hyperplasia possible role of transforming growth factor alpha in the pathogenesis of menetrier's disease: supporting evidence from humans and transgenic mice the effect of hydrocortisone and cortisone on fixation of 35s in the stomach the effect of phenylbutazone and its derivatives, oxyphenbutazone and sulfinpyrazole, on 35s sulfate incorporation in cartilage and stomach salivary glands: a paradigm for diversity of gland development gastrointestinal neoplasms in non-human primates: a review and report of new cases gastric gland degeneration induced in monkeys by the cck-b/gastrin receptor antagonist ci-988 the effect of aspirin on small intestinal mucosa morphologic aspects of lipid absorption gastric and gastric epithelial physiology two-year evaluation of misprostol for carcinogenicity in cd sprague-dawley rats distribution and incidences of calcified lesions in dba/2ncrj and balb/canncrj mice interaction of phenytoin and inflammation induces gingival overgrowth in rats the intestinal response to high bulk feeding in the rat comparison of canine and human gastrointestinal physiology stress ulceration-clinical relevance of animal and human studies pathology of laboratory rats and mice effect of chronic aspirin ingestion on epithelial proliferation in rat fundus, antrum and duodenum the prognostic value of sulphomucin positive intestinal metaplasia in the the development of gastric cancer colitis in mice with high incidence of rectal prolapse volatile nitrosamine contamination of laboratory animal diets toxicological studies on omeprazole possible role of cimetidine and its nitrostated products in human stomach cancer cimetidine and gastric cancer tumours of the oral cavity, check pouch, salivary glands, oesophagus, stomach and intestines differential distribution of lymphocytes and accessory cells in mouse peyer's patches phenotypically distinct subpopulations of t cells in domes and m-cell pockets of rabbit gut-associated lymphoid tissues morphometric analysis of the small intestinal epithelium in the indomethacin-treated mouse expression of pokeweed lectin binding in murine intestinal paneth cells heterogeneity of m-cell-associated b and t cells in human payer's patches degraded carrageenan-induced colitis in cf1 mice. a clinical, histo-pathological and kinetic analysis substituted benzimidazoles inhibit acid secretion by blocking (h ++ k + ) atpase nonsteroidal anti-inflammatory drug induced jejunal and colonic diaphragm disease: a report of two cases effect of chronic misoprostol ingestion on rat gastric morphology and turnover abnormal patterns of mucous secretion in apparently normal mucosa of large intestine with carcinoma mucous secretion in rat colonic mucosa during carcinogenesis induced by dimethylhydrazine. a morphological and histo-chemical study mucins in the human gastrointestinal epithelium: a review. invest high-dose pancreatic-enzyme supplements and fibrosing colonopathy in children with cystic fibrosis the number of villi in rat's jejunum and ileum: effect of normal growth, partial enterectomy and tube feeding spontaneous tumors including gastrointestinal neoplasms and malignant melanoma, in syrian hamster aspirin, paracetamol and non-steroidal anti-inflammatory drugs. a comparative review of side effects campylobacter jejuni/coli in commercially reared beagles. prevalance and serotypes antigen specificity and morphological characteristics of chlamydia trachomatis, strain sfpd, isolated from hamsters with proliferative ileitis acquired salivary dysfunction. drugs and radiation gastritis cystica profunda application of lectins for detection of goblet cell glycoconjugate differences in proximal and distal colon of the rat lectin histochemistry of 1,2-dimethylhydrazine-induced rat colon neoplasia effects of puromycin on the structure of rat intestinal epithelial cells during fat absorption morphological aspects on pancreatic islets of non-obese diabetic (nod) mice carcinoma and related lesions in dog stomach induced by oral administration of n-methyl-nâ´-nitro-n-nitrosoguanidine cryptosporidiosis in a pup with distemper. vet.pathol squamous cell carcinoma, forestomach, rat adverse effects of mouthwash use. a review. oral surg.oral med.oral pathol tyzzer's disease, intestine, mouse, rat, hamster salmonellosis, intestine, mouse, rat, hamster adequate substitution with electrolytes in toxicological testing of 'loop' diuretics in the dog the forestomach in rats and mice, a food store without bacterial protein digestion m cells in peyer's patches of the intestine histochemie de la muqueuse gastrique fundique du chien traitã© par des drogues ulcã©rigã¨ne an electron-microscope and histo-chemical study of melanoisis coli mitochondria. in: ultrastructural pathology of the cell and matrix sustainability of forestomach hyperplasia treated with ethyl acrylate for 13 weeks and regression after cessation of dosing the cytochemical localization of lysozyme in paneth cell granules regional differences in glycoconjugates of intestinal m cells in mice: potential targets for mucosa vaccines susceptibilities of drug to nitrosation under simulated gastric conditions features of small intestinal pathology (epithelial cell kinetics, intra-epithelial lymphocytes, disaccharidases) in a primary giardia muris infection staining rickettsiae in yolk sac cultures alterations of gastric mucosa following a graded partial gastrectomy tumours of the salivary glands the oral cavity aristolochic acid is a direct mutagen in s. typhimurim studies of the binding of trypsin and chymotrypsin by human intestinal mucosa study of cold plus restraint stress gastric lesions in spontaneously hypertensive, wistar and sprague-dawley rats mucins in normal and neoplastic gastrointestinal epithelium the lectins: carbohydrate-binding proteins of plants and animals prolactin and ergocryptine effects mucus glycoproteins of the rat ileum tumours of the jaws hypopigmentary changes with a platelet aggregation inhibitor (abstract no. 2526). fed.proc drug induced enteropathy characterized by lipid in macrophages altered patterns of mucin secretion in gastric hyperplasia in mice digestive system. in: rat histopathology. a glossary for use in toxicity and carcinogenicity studies k virus infection, intestinem mouse a silver nitrate stain for alpha-2 cells in human pancratic islets silver stains in the study of endocrine cells of the gut and pancreas effects of antrectomy or porta-aval shunting on the histamine-storing endocrine-like cells in oxyntic mucosa of rat stomach. a fluorescence histochemical, electron microscopic and chemical study the vagus exerts trophic control of the stomach in the rat activation and hyperplasia of gastrin and enterochromaffin-like cells in the stomach gastrin and the trophic control of gastric mucosa onkocytes and so-called hã¼rthle cell tumor interaction of microorganisms, epithelium, and lymphoid cells of the mucosa-associated lymphoid tissue dog and swine as models for testing indomethacin-induced gastrointestinal irritation sialodacryoadenitis in the rat: effects of immunosuppression on the course of the disease role of gut in zenobiotic metabolism epithelial cell kinetics in the small intestine of the rat 60 days after resection of 70 percent of the ileum and jejunum compensation by the residual intestine after intestinal resection in the rat proceedings of the lst international symposium on omeprazole a cecal diaphragm associated with the use of nonsteroidal anti-inflammatory drugs chronic gastritis of the glandular stomach, adenomatous polyps of the duodenum, and calcareous pericarditis in strain dba mice idiopathic megaoesophagus in rat immunocytochemical localization of alkaline phosphatase in absorptive cells of rat small intestine after colchicine treatment spontaneous neoplasm incidences in fischer 344 rats and b6c3f1 mice in two-year carcinogenicity studies: a national toxicology program update diphenyldydantoin (dilantin) gingival hyperplasia: drug-induced abnormality of connective tissue on cell proliferation and differentiation of the fundic mucosa of the golden hamster. fractographic study combines microscopy and 3 h-thymidine autoradiography tritiated thymidine autoradiographic study on cellular migration in the gastric gland of the golden hamster enterochromaffin-like cell carcinoids of gastric mucosa in rats after life long inhibition of gastric secretion gastric cancer after cimetidine in a patient with two negative pre-treatment biopsies induction of experimental allergi sialadenitis in mice spontaneous development of auto-immune sialodenitis in aging bdf1 mice i mmunofluorescent localization of enterokinase in human small intestine mouse models for colorectal cancer morphologic changes in the urinary bladder and stomach after long-term administration of sodium saccharin in f344 rats lessons from genetically engineered animal models iii. lessons learned from gastrin gene deletion in mice immunohistological characterization of intra-epithelial and lamina propria lymphocytes in control ileum and colon and inflammatory bowel disease induction of intestinal tumors in rats by dextran sulphate sodium gastric enterochromaffin-like hyperplasia and neoplasia in the rat: an indirect effect of the histamine h2-receptor antagonist bl-6341 oxidative metabolism of foreign compounds in rats small intestine: cellular localization and dependence on dietary iron clinicopathological studies of gastrointestinal disease in macaques non-tuberculous myobacterial disease in rhesus monkeys on human papillomaviruses the laboratory rat. biology, diseases inhibition of intestinal protein synthesis and lipid transport by ethionine experimental toxicity studies with captopril, an inhibitor of angiotesin 1-converting enzymes 2. one month studies of chronic toxicity of captopril in rats development of spontaneous tongue calcification and polypoid lesions in dba/2ncrj mice changes in enzyme levels accompanying differentiation of intestinal epithelial cells textbook of endocrinology acute disease of the submaxillary and harderian glands (sialodacryoadenitis) of rats with cytomegaly and no inclusion bodies cellular kinetics of gastrointestinal mucosa, with special reference of gut endocrinecells changes of gastric mucus glycoproteins with aspirin administration in rats induction of colorectal tumours in rats by sulpated polysaccharides induction of duodenal tumors in mice by oral administration of hydrogen peroxide carcinogencity of butylated hydroxyanisole in f344 rats modifying effects anti-oxidants on chemical carcinogenesis a 13 week feeding study of butylated hydroxyanisole: the subsequent regression of the induced lesions in male fischer 344 rat forestomach epithelium an 85-day study of butylated ydroxyanisole in the cynomolgus monkey subchronic studies of doxylamine in fischer 344 rats effects of dietary fiber on mucosal growth and cell proliferation in the small intestine of the rat: a comparison of oat, bran, pectin, and guar with total fiber deprevation transmissible ileal hyperplasia, hamster anatomic adaption of the alimentary tract of the rat to the hyperphagia of chronic alloxan-diabetes a variant of intestinal metaplasia associated with gastric carcinoma: a histochemical study role of intestinal metaplasia in the histogenesis of gastric carcinoma colorectal mucin histochemistry in health and disease: a critical review uptake of particulate and soluble antigens in the small intestines of the rat chemical colitis due to endoscopic cleaning solutions: a mimic of pseudomembranous colitis pathology of domestic animals intestinal accumulation of urea in germ-free animals: a factor in caecal enlargement digestive enzymes in the parotid and submandibular glands of mammals morphologishe verã¤nderungen der magenmukosa von ratten nach chronischer antazidagabe enteric viruses of non human primates increased accumulation of sulfated glycoaminoglycans in cultures of human fibroblasts from phenytoin-induced gingival overgrowth new insights into the stem cells and the precursors of gastric epithelium colonic lymphoid-glandular complex (microbursa): nature and morphology lymphoid tissue and lymphoid-glandular complexes of the colon: relation to diverticulosis histology of salivary gland infarction in the dog adrenergic factors involved in the control of crypt cell proliferation in jejunum and descending colon of mouse pill esophagitis immunogold localization of ingested kidney bean (phaseolus vugaris) lectins in epithelial cells of the rat small intestine epithelial dysplasia of the rabbit colon induced by degraded carrageenan hyperkeratinization and hyperplasia of the forestomach epithelium in vitamin a deficient rats intrinsic resistance of colon tumors to anthrapyrazoles and antracyclines may be linked with a detoxification mechanism of intestinal cells (abstract no. 1040) studies on the effects of 3-hydroxy-3-methylglutaryl coenzyme a reductase inhibitors on the rodent forestomach reversibility of adenomatous hyperplasia in the gastric stump after diversion of bile reflux in rats experimental production of possible autoimmune gastritis followed by macrocytic anemia in athymic mice expression cloning and characterization of the canine parietal cell gastrin receptor histomorphologic investigations on the effect of cyclophosphamide on dentinogenesis of the rat incisor. scand preclinical toxicology profile of misoprostol premalignancy of the mucosal polyp in the large intestine: i. histologic gradation of the polyp on the basis of epithelial pseudostratification and glandular branching morphologic changes in the gastric mucosa of rats and dogs treated with an analog of prostaglandin e1 forestomach carcinogens: possible mechanisms of action lectin-peroxidase conjugates in histopathology of gastrointestinal mucosa the rat incisor in toxicologic pathology the pharmacology and toxicology of caffeine incorporation of radiosulfate in the gastric mucosa of the rat subjected to restraint antisecretory drugs and gastric cancer plasma gastrin and gastric enterochromaffin-like cell activation and proliferation. studies with omeprazole and ranitidine in intact and antrectomized rats effects of cyclosporin a administration in cats gingival hyperplasia associated with nifedipine therapy the aetiology of caecal enlargement in the rat effect de l'administration prolongã©e d'un antisã©crã©toire gastrique, le pirenzepin, sur les populations cellulaires de l'estomac de rat hepatocytes in the mouse stomach tumours of the oral cavity, pharynx, oesophagus and stomach a toxicological profile of cimetidine specific staining of sulphate groups with alcian blue at low ph the mucin histochemistry of normal and neoplastic gastric mucosa the effect of the thyroid gland on intestinal absorption of hexoses structural changes of the gastrointestinal mucosa induced by prostaglandins structural aspects of salivary glycoproteins drug-induced disorders of the esophagus histochemical observations on paneth cells diphtheria toxin-mediated ablation of parietal cells in the stomach of transgenic mice carcinogencity of methylated dinitro-sopiperazines in rats comparative study of intestinal adenocarcinoma of animals and man dental dysplasia in rats and mice chronic oral administration of 1-nitrosopiperazine at high doses to mrc rats nifedipine-induced gingival hyperplasia: a histochemical and ultrastructural study drug-induced lysosomal storage disorders oesophagostomiasis in feral monkeys (macaca mulatta) hyperalimentation in normal animals produced by protamine insulin effects of cytotoxic chemotherapy on dental development nutritional influences on aging of fischer 344 rats: ii the use of rats associated with human faecal flora as a model for studying the effects of diet on the human gut microflora colonic ulceration in young rats fed degraded carrageenan unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration gut-associated lymphoid tissue and 1,2-demethylhydrazine intestinal tumors in the rat: a histological and immunoenzymatic study leucocyte and bone marrow effects of a thiomorpholine quninazosin antihypertensive agent localization of epidermal growth factor receptors in cells of the enamel organ of the rat incisor effect of dosage form and formulation factors on the adherence of drugs to the esophagus morphologic changes of esophageal mucosa in the rat after chronic alcohol ingestion experimental adenomas and carcinomas of the large intestine behave as distinct entities: most carcinomas arise de novo in flat mucosa la glande endocrine de l'intestin chez l'homme differentiation of immature mucous cells into parietal, argryophil, and chief cells in stomach grafs sequential histochemical and morphometric studies on preneoplastic and neoplastic lesions induced in rat colon by 1,2-dimethylhydrazine recovery from amiodarone-induced lipidosis in laboratory animals: a toxicological study parietal cell hyperplasia indued by long-term administration of antacids to rats methotrexate-induced changes in rat parotid and submandibular gland function a spontaneous outbreak of polychlorinated biphenyl (pcb) toxicity in rhesus monkeys (macaca mulatta): toxicopathology classical absorption theory and the development of gastric mucosal damage associated with non-steroidal anti-inflammatory drugs mesenteric lymphoblast localization throughout the murine small intestine: temporal analysis relating intestinal length and lymphoblast division morphologic lesions in ageing syrian hamsters histological identification of campylobacter using gimenez technique in gastric antral mucosal the carcinogenic action of aristolochic acid in rats rat immunoglobulins in serum and secretions: comparison of iga and igg in serum, colostrum, milk and saliva of protein malnourished and normal rats influence of ofloxacin on the faecal flora the structure, origin and function of mucosal mast cells. a brief review pseudomembranous colitis in a leukaemia unit: a report of five fatal cases formulation of n-nitroso compounds: chemistry kinetics, and in vivo occurrence the etiology of gastric cancer: intragastric nitro-samide formation and other theories forestomach lesions induced by butylated hydroxyanisole and ethylene dibromide: a scientific and regulatory perspective international classification of rodent tumours. part 1, the rat the 'swiss roll'. a simple technique for histological studies of the rodent intestine organization of an actin filament-membrane attachment in the microvilli of intestinal epithelial cells charaterization and localization of myosin in the brush border of intestinal epithelial cells development of the human gastrointestinal tract: twenty years of progress aroclor 1254-induced intestinal metaplasia and adenocarcinoma in the glandular stomach of f344 rats immuno-histochemical localization of epidermal growth factor in rodent submandibular glands intestinal metaplasia of the gastric mucosa precancerous conditions and epithelial dysplasia in the stomach hypersensitivity reactions in the small intestine. 6. pathogenesis of the graft-versus-host reaction in the small intestinal mucosa of the mouse the distribution of acid mucopoly-saccharides in normal kidneys as shown by the alcian blue feulgan (ab-f) and alcian blue-periodic acid-schiff (ab-pas) stains gastric endocrine cell hyperplasia and carcinoid tumors in atrophic gastritis type a histochemical classification of acini and ducts of parotid glands from artiodactyles, carnivores and rodents a morphological and histrochemical study of a drug-induced enteropathy in the alderley park rat the mouse submandibular gland: an exocrine organ for growth factors dysplasia of the gastric mucosa and its relation to the precancerous state ultrastructure of dog parotid gland corticosteroid treatment increases parasite numbers in murine giardiasis duct-associated lymphoid tissue (dalt) of minor salivary glands and mucosal immunity minor salivary gland duct-associated lymphoid tissues (dalt) in monkeys, changes with age quantitative light and electron microscopic studies on the effect of vagotomy on parietal cells in rats toxicology and carcinogenesis studies of ampicillin trihydrate in f344/n rats and b6c3f1 mice animal models for gastric helicobacter immunology and vaccine studies fundal gastritis after billroth-ii type resection in patients with duodenal ulcer mesovarial leiomyomas in rats in a chronic toxicity study of musuprine hydrochloride current concepts in mucosal immunity v. role of m cells in transepithelial transport of antigens and pathogens to the mucosal immune system the influence of food additives and related materials on lower bowel structure and function the interactions of lectins with animal cell surfaces immunohistochemical localization of carbonic anhydrase in submandibular salivary glands of mice and hamsters treated with phenylephrine, testosterone or duct ligation tumours of the oral cavity, pharynx, oesophagus and stomach distribution of glutathione and its related enzymes in small intestinal mucosa of rats the ulcerogenic effect on the oesophagus of three î²-adrenoceptor antagonists, investigated in a new porcine oesophagus test model goblet cell hyperplasia is a feature of the adaptive response to jejunoileal bypass in rats a study on carcinogenesis induced by degraded carrageenan arising from squamous metaplasia of the rat coloretum differences in thegastro-intestinal mirobiota of specific pathogen free mice: an often unknown variable in biomedical research paneth cells and innate immunity in the crypt microenvironment. gastroenterology mucosal immunity and inflammation iv. paneth cell antimicrobial peptides and the biology of the mucosal barrier entero-hepatic cycling in rats plays a major role in fatal nsaid intestinal ulcerogenicity sequential uptake of horseradish peroxidase by lymphoid follicle epithelium of peyer's patches in the normal unobstructed mouse intestine: an ultrastructural study antigen processing structures of the mammalian intestinal tract: an sem study of lymphoepithelial organs cytochemical analysis of alkaline phosphatase and esterase activities and of lectin-binding and anionic sites in rat and mouse peyer's patch m cells predicting anticancer drug effects in man from laboratory animal studies the anatomical basis for the immune function of the gut morphological and histochemical changes in intestinal mucosa in the reserpine-treated rat model of cystic fibrosis the pericryptal fibroblast sheath. iv. replication, migration and differentiation of the subepithelial fibroblasts of the crypts and villus of the rabbit jejunum comparative analysis of the 16s rna gene sequence of the putative agent of proliferative ileitis of hamsters histochemistry: theoretical and applied h2-receptor antagonists and gastric cancer diagnostic exercise: inter-mandibular swelling in rats depletion of salivary gland epidermal growth factor by silodacryoadenitis virus infection in the wistar rat biology and biochemistry of papillomaviruses lescol (fluvastatin sodium) the cytology of salivary glands salivary gland sexual dimorphism: a brief review enzyme histochemistry of the human stomach with special reference to intestinal metaplasia nitrates and gastric cancer mucosal immunity and inflammation v. innate mechanisms of mucosal defense and repair: the best offense is the best defense derivation of mouse intestinal rypts from single progenitor cells twenty-one month evaluation of misoprostol for carcinogenicity in cd-1 mice association of long lasting unsurmountable histamine h2 blockade and gastric carcinoid tumours in the rat changes in the gastric mucosa of the mouse associated with long lasting unsurmountable histamine h2 blockade pharmacological and toxicological effects of chronic porcine growth hormone administration in dogs campylobacter jejuni enteritis in man and domestic animals is ethanol-induced damage of the gastric muosa a hyperosmotic effect? comparative studies on the effects of ethanol, some other hypperosmotic solutions and acetyl-salicylic acid on rat gastric mucosa synergistic interaction between aspirin, or other non-steroidal antiinflammatory drugs, and stress which produces severe gastric mucosal damage in rats and pigs the effects of aspirin and other non-steroid anti-inflammatory/analgesic drugs on gastrointestinal mucus glycoprotein biosynthesis in vivo: relationship to ulcerogenic actions electron microscopic observations comparing the gastric mucosal damage induced in rats and pigs by benoxaprofen and aspirin, reflecting their differing actions as prostaglandin-synthesis-inhibitors megaoesophagus in icrc mice compylobacter-like organisms in the gastric mucosa of rhesus monkeys gastric cancer in patients who have taken cimetidine effect of cimetidine on gastric juice n-nitrosamine concentration cryptosporidium cuniculus in the rabbit clostridium difficile colitis in a rabbit following antibiotic therapy for pasteurellosis clostridium difficile antitoxin neutralization of cecal toxin(s) from guinea pigs with penicillin-associated colitis clostridium difficile typhlitis in hamsters not associated with antibiotic therapy clostridial enteropathies, hamster spontaneous non-neoplastic gastric lesions in female han: nmri mice, and influence of food restriction throughout life saponification-induced increase in the periodic-acid-schiff reaction in the gastrointestinal tract. mechanism and distribution of reactive substance structure and carbohydrate histo-chemistry of postnatal canine salivary glands recombinant human epidermal growth factor 1-48 -induced structural changes in the digestive tract of cynomolgus monkeys (macaca fascicularis) influence of long-term 16,16-dimethyl prostaglandin e2 treatment on the rat gastrointestinal mucosa digestion and absorption of carbohydrate and nitrogeneous matter in hindgut of the omnivorous non-reminant animal salivary glands in longterm alloxan-diabetic rats. a quantitative light and electron-microscopic study digestive system: intestines. in: histology. a text and atlas helicobacter pylori-infected human antral primary cell cultures: effect on gastrin cell function the gastrointestinal tract dysplasia in inflammatory bowel disease: standardized classification with provisional clinical applications histochemical study on the paneth cell in the rat ultrastructural observations on the pathogenesis of aspirin-induced gastric erosions dietary related periodontitis and oro-nasal fistulation in rats abnormal skeletal and dental growth in epileptic children tooth root resorption induced in rats by diphenylhydantoin and parathyroidecotomy light microscopic morphometric analysis of rat ileal mucosa. i. component quantitation if iga-containing immunocytes short-term effects of various phenols and acids on the fischer 344 male forestomach epithelium enzootic and epizootic adrenal medullary proliferative diseases of rats: influence of dietary factors which affect calcium absorption campylobacter pylori, gastritis, and peptic ulcer disease in central africa the effect of diet on the mammalian gut flora and its metabolic activities interactions of the gut microflora and the host in toxicology naturally occurring tumors and other lesions of the digestive tract in untreated c57bl mice pathology of tumours in laboratory animals. tumours of the mouse esophageal impaction in the bhe rats minimal invasive carcinoma of the colon in rats early effects of alloxan on rat submandibular gland cyclic motor activity; migrating motor complex age-related changes of rat submandibular gland: a morphometric and ultrastructural study indomethacin produces gastric antral ulcers in the refed rat quantitative electron microscopic observations on paneth cells of germ-free and ex-germ-free wistar rats immunohistochemical observations of immunoglobin a in the paneth cells of germ-free and formerly-germ-free rats pigmentary changes in the rat oral mucosa following antimalaria therapy monoclonal antibodies to a zinc-binding protein of rat paneth cells age, weight and social effects on ulceration in rats the evaluation of anticancer drugs in dogs and monkeys for the prediction of qualitative toxicities in man gingival fibromatosis, macaca mulatta identification and mutagenicity of metabolites of aristolochic acid formed by rat liver melanosis coli: a study with the electron microscope light microscopic detection of sugar residues in glycoconjugates of salivary glands and the pancreas with lectin-horseradish peroxidase conjugates. i. mouse light microscopic detection of sugar residues in glycoconjugates of salivary glands and the pancreas with lectin-horseradish peroxidase conjugates genetic and sex-related differences in the structure of subnmandibular glycoconjugates mucosal microhernias in the nonhuman primate colon: their role in the pathogenesis of colonic diseases quantitative age changes in the histological structure of human submandibular salivary glands immunohistological characterization of intra-epithelial lymphocytes of the human gastrointestinal tract chronic duodenal ulcers in pantothenate deficient mice focal avillous hyperplasia of the mouse duodenum colon epithelium. i. light microscopic, histochemical, and ultrastructural features of normal colon epithelium of male fischer 344 rats colon epithelium. ii. in vivo studies of colon carcinogenesis. light microscopic, histrochemical, and ultrastructural studies of histogenesis of azoxymethane-induced colon carcinomas in fischer 344 rats two types of mucous cells in the colon crypt carrageenan ulceration as a model for human ulcerative colitis. lancet, ii comparative histochemistry of gastro-intestinal mucosubstances assessment of risk of formulation of carcinogenic n-nitroso compounds from dietary precursors in the stomach granular cells as a marker of early amiodarone hepatotoxicity: a pathological and analytical study marked epithelial hyperplasia of the rat glandular stomach induced by long-term administration of iodoacetamide a profile of the gastrointestinal toxicity of drugs used to treat inflammatory diseases morphology and cyto-chemistry of rat salivary gland acinar secretory granules and their alteration by isoproterenol. i. parotid gland effects of sodium salicylate on epithelial cells of the rectal mucosa of the rat: a light and electron microscopic study the effects of long-term propranolol on the salivary glands and intestinal mucosa of the mouse effects of cardiotonic phosphodiesterase inhibitors on rat salivary gland. presented at the 39th annual acvp meeting inhibitors of sterol synthesis. morphological studies in rats after dietary administration, administration of 5 î±-cholest-8(14)-en-3î²-ol-15-one, a potent hypocholesterolemic compound m cell numbers increase after transfer of spf mice to a normal animal house environment strictures of ascending colon in cystic fibrosis and high-strength pancreatic enzymes pathology of laboratory rats and mice anatomy of the cecum of the laboratory mouse and rat gastric carcinoids and related endocrine growths pathogenesis of peptic ulcer and implications for therapy histochemical distribution of lysozyme activity in organs of normal mice and radiation chimeras human peyer's patches: an immunohistochemical study diamine methods for differentiating mucosubstances histochemically metabolic and morphometric changes in small and large intestine in rats fed high-fiber diets melanosis coli: studies of the toxic effects of irritant purgatives surface morphology of the gastroduodenal mucosa in duodenal ulceration the effect of vincristine on dentino-genesis in the rat incisor. scand autoradiographic investigation of dentine production in rats incisosrs after vincristine administration. scand acute and late radiation injury in rhesus monkey parotid glands. evidence of interphase death unique radiosensitivity of serous cells in rhesus monkey submandibular glands pathologic observations on the adenomatous lesions of the stomach in mice of strain i carcinoma of the glandular stomach of rats ingesting n,nâ´2 ulcerative enterocolitis in dogs induced by drugs aspirininduced glandular dysplasia of the stomach. histologic and histochemical studies in rats effect of lumenal contents on colonic cell replacement pathologic findings in the stomach of rats treated with the h2-receptor antagonist tiotidine morphologic stomach findings in rats and mice treated with the h2-receptor antagonists, ici 125211 and ici 162846 lectin-peroxidase reactivity in rat gastric mucosa tumour production in glandular stomach of rat by nmethyl-nâ´nitro-n-nitrosoguanidine ageing affects the drug metabolism systems of rat liver, kidney, colon and lung in a differential fashion oral papillomatosis in new zealand white rabbits diagnostic exercise: lingual growths in rabbits immunoperoxidase localization of papillomaviruses in hyperplastic and neoplastic epithelial lesions of animals hypergastrinemia after blockade of acid secretion in the rat. trophic effects pathological features of the colonic tumors induced in rats by the administration of 1,2-dimethylhydrazine comparative studies on the gastrointestinal lesions caused by serveral non-steroidal anti-inflammatory agents in the rats cells intermediate between mucous neck cells and chief cells in rat stomach fine structure of giant hypertrophic gastritis developed in thymectomized mice hyperplastic gingivitis in a child receiving sodium valproate treatment dopamine disorderin dodenal ulceration biology of disease. pathogenesis of duodenal ulcer disease from cysteamine to mptp: structure-activity studies with duodenal ulcerogens digestive system. monographs on pathology of laboratory animals role of salivary mucins in the protection of the oral cavity effects of 16,16-dimethyl prostaglandin e2-methyl ester on aspirin-and indomethacin-induced gastrointestinal lesions in dogs hypertrophic gastropathy resembling menetrier's disease in transgenic mice overexpressing transforming growth factor î± in the stomach tumours of the oral cavity, buccal pouch, oesophagus, forestomach and salivary glands nodular hyperplasia of oncocytes in mouse submandibular glands early colonic lesions in experimental shigella infections in rhesus monkeys: revisited variation in susceptibility to the induction of forestomach tumours by butylated hydroxyanisole among rats of different strains pathology of neoplasia and preneoplasia in rodents independent induction of intestinal metaplasia and gastric cancer in rats treated with n-methyl-nâ´-nitro-n-nitrosoguanidine the response of three inbred strains of rat to the carcinogen 1,2-dimethylhydrazine small intestinal type' and 'colonic type' intestinal metaplasia of the human stomach and their relationship to the histogenetic types of gastric adenocarcinoma light and electron microscopic cytochemistry of glycoconjugates in the recto-sigmoid colonic epithelim of the mouse and rat preclinical toxicologic evaluation of bleomycin (nsc 125 006), a new anti-tumor antibiotic oral squamous cell carcinoma in ad libitum-fed and food restricted brown-norway rats the pathoparasitology of the alimentary tract and pancreas of non-human primates: a review t cell attractant chemokine expression initiates lacrimal gland destruction in nonobese diabetic mice altered patterns of mucin secretion in the precancerous lesions induced in the glandular part of the rat stomach by the carcinogen n-methyl-nâ´nitro-n-nitrosogaunidine immunohistochemistry and radioimmunoassay of egf in submanidbular glands of mice treated with secretogogues carcinogenicity of cyproterone acetate preclinical toxicology studies with acylovir: acute and sub-chronic tests the influence of adrenoreceptor activity on crypt cell proliferation in rat jejunum a sporozoan found in the peptic glands of the common mouse cyclophosphamide-induced abnormalities in the incisors of the rat spontaneous adencarcinoma of the ascending colon in wistar rats: the intracytoplasmic presence of a campylobacter-like bacterium inhibition of prostaglandin synthesis as a mechanism of action of aspirin-like drugs spontaneous hyperplastic and metaplastic duct epithelium in the sublingual salivary glands of wistar rats spontaneous tuors of the syrian hamster: observations in a closed breeding colony and a review of the literature mucin gene structure and expression: protection vs. adhesion morphological effects of high dose neomycin sulphate on the small and large intestine comparative and experimental pathology of fibrosing colonopathy intestinal pathology in the dog induced by sublethal doses of amiodarone pharmacological effects of epidermal growth factor (egf) with focus on the urinary and gastrointestinal tracts chemically induced lipidosis of the small intestinal villi in the rat lesions of experimentally induced tyzzer's disease in syrian hamsters, guinea pigs, mice and rats hexamitis in laboratory mice, hamsters, and rats melanosis coli: a consequence of anthraquinone-induced apoptosis of colonic epithelial cells safety evaluation of cimetidine: report at the termination of a seven-year study in dogs hypopigmentation in dogs treated with an inhibitor of platelet aggregation gingival hyperplasia in dogs induced by oxodipine, a calcium channel blocking agent cholecystokinin receptors gastrointestinal ph measurement in rats: influence of microbial flora, diet and fasting morphogenesis of chemically induced neoplasms of the colon and small intestine in rats natural history of intestinal neoplasms induced in rats by a single injection of methyl (acetoxymethyl) nitrosamine proliferative lesions of the glandular stomach and liver in f344 rats fed diets containing aroclor 1254 experimentally induced intestinal metaplasia in wistar rats by x-ray irradition. gastroenterolgy effect of dietary undegraded carrageenan on colon carcinogenesis in f344 treated with azoxymethane or methyl-nitrosourea induction of intestinal metaplasia in the rat gastric mucosa by local x-irradiation the effect of sex difference on induction of intestinal metaplasia in rats distribution of the histaminergic neuron system in the central nervous system of rats; a fluorescent immunohistochemical analysis with histidine decarboylase as a marker replacement of functional parenchymal cells by fat and connective tissue in human submandibular salivary glands. an age related change canine papilloma: progression of oral papilloma to carcinoma bacterial and mycotic disease dental health of survivors of malignant disease analysis of protein synthesis in rat salivary glands after chronic treatment with î²-receptor agonists and phosphodiesterase inhibitors mucinous carcinoma of the ileum in the rat mucin histochemistry of gastric intestinal metaplasia long-term effects of an inotropic phosphodiesterase inhibitor (ici 153,110) on rat salivary gland, hardarian gland and intestinal mucosa the synovial membrane, liver, and tongue: target organs for a ricin a-chain immunotoxicin (zd0490) ultrastructural localization of acid mucosubstances in the mouse colon with iron-containing stains quantitation of glandular gastric changes in rats given a proton pump inhibitor for 3 months with emphasis on sampling scheme selection aberrant crypt foci in the colonic mucosa of rats treated with genotoxic and nongenotoxic colon carcinogen the interactions of diet and colonic microflora regulating colonic mucosal growth selective inhibition of prostaglandin production in inflammatory exudates and gastric mucosa temporal relationship between cyclooxygenase inhibition, as measured by prostacyclin biosynthesis, and the gastro-intestinal damage iinduced by indomethacin in the rat a biochemical basis for the gastro-intestinal toxicity of non-steroid antirheumatoid drugs butylated hydroxianisole mechanistic data and risk assessment: conditional species-specific cytotoxicity, enhanced cell proliferation, and tumor promotion menetrier's disease presenting as iron deficiency anaemia radioautographic and quantitative studies on parietal and peptic cell kinetics in the mouse: a selective effect of gastrin on parietal cell proliferation intestinal adaptation. structural, functionaland cytokinetic changes instestinal adaptation. mechanisms of control mucosal mast cells of the rat intestine: a re-evaluation of fixation and staining properties with special reference to protein blocking and solubility of the granular glycosaminoglycan effect of mixtures of dietary fibres on the enzyme activity of the rat caecal microflora effect of prolonged metiamide medication on the fundic mucosa the membraneous epithelial (m) cell and the mucosal immune system medical progress: gastrointestinal toxicity of nonsteroidal antiinflammatory drugs nitrosatable drugs: an assessment of the risks assessing the safety of drugs for the long-term treatment of peptic ulcers the mechanism of mitrazepam-induced drooling and aspiration blood vessels of the peyer's patch in the mouse. iii high endothelial venules ultrastructural basis of intestinal absorption the effects of omeprazole and famotidine on mucin and pge 2 release in the rat stomach forestomach ulcers in crj:b6c3 (c57bl/6ncrj ã� c3h/hencrj) f1 mice forestomach ulcers in crj:b6c3 (c57bl/6ncrj ã� c3h/hencrj) f1 mice an evaluation of ampicillin pharmacokinetics and toxicity in guinea pigs drug-induced gastro-intestinal disease naturally occurring intestinal neoplasms in aged crl:cd â® br rats