untitled AIDS and Africa – when will the epidemic level off MARKKU LÖYTÖNEN Löytönen, Markku (2003). AIDS and Africa – when will the epidemic level off. Fennia 181: 1, pp. 1–11. Helsinki. ISSN 0015-0010. Presidential address at the annual general assembly of the Geographical Soci- ety of Finland, 2002. Markku Löytönen, Department of Geography, P.O. Box 64, FIN-00014 Uni- versity of Helsinki, Finland. E-mail: markku.loytonen@helsinki.fi. MS received 15 November 2002. Introduction The principle of population dynamics declares that it is environmental factors such as changes in physical conditions, food supplies and the oc- currence of natural enemies that determine the population size of any species in existence in a given area at a given point in time. These factors vary all the time, as the organic and inorganic environment is in a constant state of change. Eve- ry change that takes place in a species’ living con- ditions is reflected in its numbers. Examined over a period of time, this takes the form of a constant search for a state of equilibrium, which in an un- disturbed situation would result in a fluctuation around a certain average figure. In the course of its endeavours to create a fa- vourable environment for itself, human society has systematically attempted to prevent the spread of epidemics (McMichael 1993). Even so, the causes of such disasters were still unknown in the early nineteenth century, and it may be said that they were thus the last factors effectively limiting human population growth. Many of the advanced cultures of prehistoric times were obliged to car- ry on a bitter struggle for existence against the re- peated threats posed by epidemics. Later the se- rious outbreaks of the plague in Europe and else- where can be identified in the population curves in form of dips in total population figures (Cliff & Haggett 1988; Epstein 1995; Löytönen 1995). At the micro level, an imbalance comparable to the laws of population dynamics prevails be- tween humankind and the microorganisms that bring about diseases. Viruses, bacteria and other microbes are constantly altering their genetic make-up in search of new properties that will give them a hold over the human body, which has of- ten become able to resist their attacks by virtue of its immune defence mechanism. One should also bear in mind that the human body for healthy life needs some microbes. And in some cases, the infection by a potentially harmful microbe does not always lead to a clinical disease. Tuberculo- sis bacillus – estimated to be carried by one third of the global human population – develops into a clinical disease in only 10% of the infected pop- ulation during their lifetime provided no cofac- tors are present. The discovery in the second half of nineteenth century of the microorganisms responsible for the diseases was a scientific breakthrough that revo- lutionised our concepts of many human illness- es. It meant that the cause of infectious diseases could be pinpointed and progress could be made in their treatment, in the sense that for the first time in the history of humankind it was known what agent was responsible for the plague, for example. Since that time, improved hygiene, an- tibiotics, inoculations and many other basic tech- niques of modern medicine have proved effective in combating infections, perhaps the best exam- ple of which lies in the total elimination of small- pox as a result of the international vaccination campaign (Cliff et al. 1981; Cliff et al. 1986; Cliff & Haggett 1989). The advances in the medical sciences together with increasing economic resources and better 2 FENNIA 181: 1 (2003)Markku Löytönen education resulted in improved quality of life and increased life expectancy – but mostly for people living in the industrial world. Healthy life and long life expectancy are so dependent on economic and intellectual resources and the organising of the society – health care provision, access to ed- ucation, and other similar elements – that many countries in the developing world have little to offer to their citizens. In this respect, Africa is probably the clearest example of how complex health-related problems can threaten the exist- ence of entire nations (Barnett & Blaikie 1992; Bond et al. 1997). The diffusion of the infection with the Human Immunodeficiency Virus (HIV) throughout the world constitutes a formidable example of a proc- ess of what has been said above. In just over two decades, it has managed to spread everywhere reminding people in both rich and poor countries that the balance between harmful microbes and the humankind during the second half of the twentieth century was nothing but one transient moment in the continuous search for a balance. While the growth curves of AIDS cases in most industrial countries are currently beginning to lev- el off (Fig. 1), the curves in many developing countries such as India are growing rapidly with no signs of levelling off (Chin & Mann 1988; Smallman-Raynor et al. 1992; Pan American... 1997; The status... 2000; European centre... 2002; AIDS epidemic… 2003). Recognising that predominant modes of trans- mission vary by world’s regions and that there is a relationship between health and development, an understanding of the diffusion process, of the future growth, and of the predominant variables involved in it at a particular place are essential for disease intervention and for planning of the provision of health care. Such demands are ac- complished by modelling and forecasting the spread of any epidemic. In this paper, I will first briefly discuss the HIV pandemic and different approaches to geographical modelling and fore- casting. At the end of my paper, I will focus on AIDS in Africa and what will be the final level of prevalence once the growth curves begin to level off. African origin – but human sponsored As far as is known at present, the HI virus origi- nates from East Africa where it, or some early pro- totype of it, first infected humans from three spe- cies of monkey found in the area. It was very probably endemic throughout the nineteenth cen- tury in the territory inhabited by these monkeys. No epidemic ensued, however, because African society was for a long time sufficiently static, with people’s movements restricted to their own village or tribal community (Shannon & Pyle 1989; Shan- non et al. 1991; Smallman-Raynor & Cliff 1990, 1991; Gao et al. 1999). The situation altered after the Second World War, as the structural changes in the society and the associated urbanisation, trade and industrial- isation led to greater mobility in Africa. It was at this stage that the HI virus began to spread to wid- er areas and that the disease assumed the propor- tions of an epidemic in Africa (Barnett & Blaikie 1992). It is possible that individual travellers vis- iting the continent in the 1950s may have become infected, but these cases will not have attracted any particular attention from the health authori- ties, and any deaths that occurred must have sim- ply remained aetiologically unresolved. Thus, the HIV pandemic started and developed unnoticed. Fig. 1. HIV carriers, AIDS cases and AIDS deaths in Finland since the beginning of the epidemic. After the initial phase, the number of new HIV carriers oscillated around a broad mean value until the late 1990s. Then, largely due to the rapid increase of iv drug users, the number of new HIV car- riers jumped to a new, much higher level. This is a typical example of the interaction between the population size and the factors controlling it in a given area. FENNIA 181: 1 (2003) 3AIDS and Africa – when will the epidemic level off The situation finally came to light in the early 1980s, when the number of cases escalated and the disease came to the attention of the health authorities in the USA (Gould 1993; Gould & Wallace 1994). From the modelling point of view, it is rather essential to ask how it is possible that the HI vi- rus – slow to migrate and easy to avoid with little knowledge of the few of modes transmission – managed to spread beyond Africa to cause a pan- demic extending over the entire inhabited world. The answer to this question is geographically a very interesting one – and challenging. The bio- logical regularities that maintain a state of equi- librium between a human being and the microbes that threaten him or her have developed to their present state over a period of 100–200 million years, and may be said to have safeguarded the existence and growth of the human populations fairly well under the conditions in which the hu- man species evolved. Populations were small, and the density of settlement was not very great, in addition to which the world was for a long time inhabited by isolated human groups with no con- tact with one another (e.g., Gould 1989). During the modern times, however, conditions for the occurrence and spread of epidemics have altered drastically (Flahault & Valleron 1992; Hag- gett 1994. There has been what can only be de- scribed as a population explosion; people have moved to live in agglomerations of tens of mil- lions residents, mobility has increased, and the speed of travel has become such that any point on the globe can be reached in a few hours. It is highly probable that the HIV pandemic would never have come to anything if human society it- self had not altered its environment and ways of living as we have seen over the past few centu- ries. One pandemic – many epidemics From the geographical point of view, the HIV ep- idemics in the various continents – and in vari- ous countries as well – appear to be developing in different ways (Arbona & Löytönen 1997; Lamptey 2002). During the first decade of the HIV epidemic, it was common to characterise the pan- demic in terms of four global patterns of HIV transmission. The rapid changes in the epidemio- logical patterns e.g., in India and Vietnam have shown that the earlier classification of four pat- terns no longer holds true. The more recent view of the pandemic emphasises the multiple patterns of HIV epidemic even in neighbouring countries (Stoneburner et al. 1994, 1996; Mertens & Low- Beer 1996).. Because HIV infection in many countries has occurred disproportionately in cer- tain high-risk groups (European centre... 2002; The status... 2000) dramatic differences can be found even within a country or region. Significant differences in prevalence can also be observed between rural and urban areas (Lam & Liu 1994). In addition to these factors, social strata within a population have associated epidemiological char- acteristics that may facilitate or hinder transmis- sion of HIV. Perhaps the most striking difference in the number of AIDS cases is between the industrial countries and the developing world (Fig. 2). At the beginning of the new millennium, the UNAIDS and WHO estimated that 42 million adults and children are living either with HIV or AIDS. More than 20 million have already died of the disease. The vast majority – ca. 95% – of the people liv- ing with HIV or AIDS live in the developing coun- tries. This proportion is expected to continue to increase as poverty, poor health systems, and lim- ited resources for prevention and care promote the spread of the virus in these countries (The sta- tus... 2000). In North America, Western Europe, Australia, and New Zealand the principal modes of trans- mission are through homosexual and bisexual in- tercourse, intercourse with a drug user, and intra- venous (iv) drug use. Heterosexual transmission makes up a small yet growing percentage of the cases. Adult males constitute most of the cases while the proportion of women with HIV infec- tion is growing all the time. Consequently there is no marked paediatric epidemic. The countries within this pattern have advanced medical facili- ties to diagnose HIV infection and adequate eco- nomic resources to provide palliative care for those infected. Control of infection through blood transfusion has been effective since the mid 1980s. In Asia, rates of infection are generally at a clearly lower level. They reach two percent or more of the adult population in only three coun- tries: Thailand, Cambodia and Myanmar (The sta- tus... 2000). In many of the area’s densely popu- lated countries the current prevalence is less than 1%. There are, however, evidence that our under- standing of the HIV epidemic based on currently 4 FENNIA 181: 1 (2003)Markku Löytönen available prevalence statistics e.g., in India have little to do with the true situation, and that both incidence and prevalence are much higher than anticipated at least in India’s biggest cities. Two of the region’s countries – India and China – to- gether have a population that is over one third of the world’s population and thus the potential for an accelerating epidemic is almost immeasurable (Ramasundaran 2002). In Eastern Europe, Russia, and other parts of the former Soviet Union, the HIV epidemics contin- ue to be mainly concentrated in iv drug users. Most HIV infected people are found in the larg- est cities although cases are increasingly found in all administrative regions in Ukraine, Russia, Be- larus, and Moldova (Löytönen 1995, 1998; The status... 2000). The overall incidence figures have been smaller than in Western Europe. However, quite recently there have been marked outbreaks in Kaliningrad and Moscow among several thou- sand iv drug users suggesting that there exists quite a significant potential for the worsening of the epidemic. The relatively scarce financial re- sources and the poorly organised health care sys- tems in these countries cannot provide for the majority of the population the quality health care available in the industrial countries. In the Middle East and North Africa, HIV in- fection was relatively uncommon in the 1980’s. Currently, the number of people infected is in- creasing but due to lack of reliable data on prev- alence very little in known of the epidemics. It can be anticipated that transmission has taken place through sexual relations (homosexual and heterosexual) or intravenous drug use, but the true transmission patterns cannot be defined with con- fidence. Diagnostic capabilities are often but not always poor in this area (The status... 2000). In Latin America and the Caribbean, the epi- demic is currently evolving from predominantly homosexual and bisexual to heterosexual trans- mission with a rather diverse picture of the epi- demic. Rates are generally highest in Central America and the Caribbean where heterosexual spread of HIV is the predominant mode of trans- mission. In Haiti the current rate exceeds 5% – the only country with such a high rate outside Af- rica – and in the Bahamas the rate exceeds 4%. In South American countries, the epidemics are generally concentrated in sub-populations at highest risk, such as men having sex with men and iv drug users. The overall trends in the region seem to be similar to patterns of the epidemic in North America with somewhat more variation. For example, the increasing number of heterosexual transmission cases is reflected in a marked pae- diatric epidemic in Puerto Rico (Löytönen & Ar- bona 1996). As the HIV infection continues to penetrate the poor and less advantaged people of the region, the future epidemic here might soon resemble the situation in sub-Saharan Africa as regards the predominant modes of transmission – but not necessarily as regards the rates of infect- ed people (Pan American... 1997). On the other hand, countries such as Brazil, Argentina, and Fig. 2. Global estimate of people living with HIV/AIDS, cu- mulative number of HIV/AIDS deaths and cumulative number of children orphaned by AIDS, end of 1999. FENNIA 181: 1 (2003) 5AIDS and Africa – when will the epidemic level off Mexico have increased their efforts in providing proper care for those infected and in intervening the epidemic by education. As a consequence, statistics are showing decrease in AIDS mortality in the aforementioned countries. Looking at the world map of the HIV pandem- ic, one can without doubt state that the situation is clearly the worst of all regions and continents in sub-Saharan Africa. The UNAIDS and WHO have estimated that there are currently 29.4 mil- lion adolescent and adult people infected with HIV. The geographical distribution of the cases is rather uneven throughout the continent (Figs. 3 and 4). While the first major epidemics occurred in central and eastern African countries, the epi- demic is now the worst in the southern part of the continent. During the 1990s in South Africa, infection rates among the adult population in- creased from less than 1% to ca. 20% – an in- crease in its own class. Looking at the infection rates on the map, other countries that have equal- ly high rates are Namibia, Lesotho, Swaziland, Zambia, and Zimbabwe. The situation is almost as bad in Ethiopia, Uganda, Kenya, Tanzania, Mozambique, Central African Republic, Togo, and Fig. 3. Estimated number of people living with HIV/AIDS in Africa, end of 1999. Fig. 4. Estimated HIV/AIDS prevalence in Africa, end of 1999. Ivory Coast. With two exceptions, the prevalence rates are on a clearly lower level in West Africa. In sub-Saharan Africa the HI virus is spread pre- dominantly by heterosexual intercourse. The male to female ratio is about 1:1. The high incidence of infection among women of reproductive age has created a paediatric epidemic with far-reach- ing consequences (The status... 2000; Matshala- ga & Powell 2002; Morgan et al. 2002). Forecasting – how and what Epidemiological forecasting and spatial analysis of the diffusion of contagious diseases are essen- tial methods for monitoring and controlling of epidemics. It involves both geographical and mathematical modelling. In addition to different geographic and demographic variables, the data needed in such modelling deals with disease-spe- cific factors providing fundamental information for estimating the necessary parameters. In most diseases these parameters are well known, or can be reliably estimated from existing data or can be obtained by analysing a representative sample of 6 FENNIA 181: 1 (2003)Markku Löytönen case histories. Based on such data, modelling of the growth curves and the spatial diffusion proc- esses can be done with a high degree of accura- cy. When attempting to model the HIV epidemic, however, the situation is somewhat more compli- cated. While clinical, virological, and immuno- logical research has progressed rapidly, epidemi- ological research on HIV has been riddled with data-related problems. Estimates of the growth of the HIV epidemic have proved to be far less suc- cessful than one might wish. The conclusion is practically the same whether one looks at fore- casts provided by applying explanatory (multivar- iate) models or autoregressive (univariate) mod- els. Some attempts have been made to estimate the growth curve of AIDS cases or HIV carriers in a given population by using simulation techniques. Although these studies have provided valuable information about the epidemiological parameters needed in the modelling, they have mostly fo- cused on limited geographical areas or strictly defined subgroups of the population and thus not succeeded in providing very reliable forecasts of the future course (geographically or demographi- cally) of the epidemic at large. The problems in forecasting the HIV epidemic are mostly due to the virological features of the virus, and its connection to various forms of hu- man sexual behaviour i.e., to some of the most hidden sides of human life. Many of the funda- mental disease-specific parameters are still only partly known and understood. These include, among others, the mean and group-specific incu- bation time, the rate of risk associated with the different ways of transmission, the number of part- ners and the frequency of sexual intercourse and its variation in the heterosexual population, and the exact role of transmission cofactors such as genital ulcers. The data problem is also due to the fact that in most countries there are no systematically com- piled data available about the true HIV preva- lence, although the need for such data is gener- ally recognised. As a result, most attempts at fore- casting the HIV epidemic are based on data cov- ering only symptomatic AIDS patients. From the forecasting point of view, the primary focus should always be on analysing the spread of the virus, not merely the statistics of the clinical de- velopment of the infection – which again varies much depending on several factors. These in- clude, among others, the availability of clinical treatment, the socio-economic status of the pa- tients, and the quality of the health care system in general. Attempts to use the traditional epidemiological models has led us to a dilemma in which the ac- curacy of the data does not meet the requirements set by the models. This is not to say that work done in this field is not valuable. Quite the con- trary. Every bit of new information brings us clos- er to the date when we will able to produce reli- able forecasts using even the most demanding and complex methods (for a review, please see Ander- son 1989; Bremerman & Anderson 1990; Löytö- nen 1991, 1994a, 1994b; Thomas 1992, 1996, 1999; Low-Beer & Stoneburner 1997; Auvert et al. 2000). When will the growth level off in Africa As regards the HIV epidemics in Africa, the lack of reliable data is a key problem when planning of modelling and forecasting the detailed geo- graphical diffusion or distribution of HIV/AIDS cases. The recent estimates on country-level pro- duced by UNAIDS and WHO (The status and... 2000; AIDS epidemic... 2003) are without doubt the best and most reliable ones currently availa- ble and in my view they should be used as the base for making plans of how to most efficiently intervene the epidemics and of how to best or- ganise the scarce resources available for health care. The numbers are based on several small sets of data obtained from many independent studies and surveys throughout the continent and ana- lysed by experts (the Delphi method) to deliver as reliable forecasts on country level as possible given the conditions. Geographically speaking the picture is very broad and provides little if any advice of what are the spatial diffusion patterns within these coun- tries not to mention even more detailed spatial units. Should such modelling and forecasting be possible, the results would provide one step fur- ther planning of how to intervene the epidemics. On the other hand, one could also ask whether it is reasonable even to attempt such modelling and forecasting with the lack of sufficiently good data. Even the simplest autoregressive forecasting meth- ods would be based on estimated parameters if applied in sub-Saharan African countries. FENNIA 181: 1 (2003) 7AIDS and Africa – when will the epidemic level off There is, however, one important question that should be addressed from the forecasting point of view. Let us return to the opening point of my paper regarding the basic population dynamics. In the industrial countries, there are already clear signs that the growth curves of the HIV epidem- ics are beginning to level off. In other words, giv- en the socio-economic and other environmental conditions in most industrial countries of the world, a balance between the HIV epidemics and the population seems to be emerging. The key question, thus, is what will be the prevalence rate level once the balance has been achieved in those conditions prevailing in each region or country. If – as it seems – the prevalence rate levels in the industrial countries remain more or less be- low 1%, the epidemics can be seen to be under control to a certain extent at population level. Subgroups of known high risk behaviour such as iv drug users and men who have sex with men without protection will of course demonstrate much higher prevalence rates. But in most coun- tries these subgroups are marginal and will have lesser impact on how the epidemics of the over- all population will develop. In sub-Saharan Africa, we are already hitting prevalence rate levels that are horribly high com- pared with virtually any other part of the world. There are, furthermore, no significant signs of the growth levelling off in most of the countries. The few exceptions are Uganda that has managed to bring down its estimated prevalence rate from 14% to 8% with a strong prevention campaign, and Zambia where there are some encouraging signs of alike development. If the growth in sub- Saharan Africa continues with current speed – and this is most likely based on what we know of the predominant modes of transmission, how limited resources are in Africa, and by looking at the most recent estimates of the current trends of the epi- demics – then the levelling off will happen at a level leaving little if any hope to many of the countries in sub-Saharan Africa. Undoubtedly, there will be great geographical variation in Afri- ca from country to country, within countries, and between different groups of people according to their socio-economic, ethnic and behavioural sta- tus. But even if some countries and regions suc- ceed in intervening in the epidemic and forcing the growth to level off at lower level, continent- wide the level will be much higher than else- where (Ojo & Delaney 1997; Gilks et al. 1997; Tarantola & Schwartlander 1997; Bouckenooghe & Shandera 1999) (Fig. 5). Is there hope for sub-Saharan Africa In industrial countries, the use of highly active anti-retroviral therapies since 1995 have reduced AIDS mortality significantly (Fig. 6). Such treat- ment of HIV infected individuals is very expen- sive, especially since monitoring and therapy is required throughout the germination period which with currently available drugs may be tens of years. The rich industrial countries may well be able to cope with the costs of the HIV epidemic Fig. 5. HIV prevalence by age, South African antenatal clin- ics, 1993–1999. Fig. 6. AIDS cases and AIDS deaths in the WHO European region since the beginning of the epidemic. The use of highly active anti-retroviral therapies since 1995 have sig- nificantly reduced AIDS mortality. 8 FENNIA 181: 1 (2003)Markku Löytönen especially now when the growth curves are lev- elling off at a manageable level. But the situation in Africa is far more difficult. The treatment of HIV infected people is already exhausting the limited resources available for health care. Little is left for any attempts at controlling of the epidemic by means of educating people. It is already clear that AIDS mortality in Africa will soon rise to meas- ures never seen before (Fig. 7). This will not only leave millions of orphans, but strike hard and deep in the societies through loss of work force in their best age (Figs. 8 and 9). The gap is enor- mous between the poor and rich countries of their chances of overcoming the economic and social problems raised by the HIV pandemic (De Kock et al. 2002; Rankin 2002). The worldwide campaign to eliminate small- pox provides an interesting point of comparison for the HIV pandemic. The campaign was suc- cessful because the basic costs were low, the project gained broad international acceptance and the vaccination procedure did not require any advanced professional skills. Such a cam- paign could in principle be mounted against the HI virus once a vaccine has been developed. Despite significant efforts such vaccine is still far away in the future. And current information nev- ertheless suggests that any vaccine that might become available is likely to be an expensive preparation even by the standards of the indus- trial countries and it may require advanced med- ical techniques and highly skilled health care staff for its administration. Fig. 7. Estimated number of AIDS deaths in Africa in 1999. Fig. 9. Projected population structure with and without the AIDS epidemic, Botswana, 2020. Fig. 8. Cumulative number of children orphaned by AIDS, end of 1999. FENNIA 181: 1 (2003) 9AIDS and Africa – when will the epidemic level off Would it be perceivable to think that the indus- trial countries financed a campaign of the same magnitude as the smallpox campaign? This is hardly the future unless the costs of treatment per individual HIV carrier come down to a managea- ble level. With currently available highly complex and expensive treatment such a campaign is not a likely option. Could the worsening HIV epidemic in Africa and perhaps in other developing areas – especial- ly in the densely populated countries such as In- dia – pose a serious threat to the populations in the industrial countries and consequently force the rich countries to launch a campaign? Even a cautious evaluation of the situation suggests that it is quite possible that the growing population potential of the developing countries will even- tually result in increasing migratory flows exert- ing pressure on the borders of the industrial coun- tries – indeed there are already many signs of the attractiveness of these countries in the eyes of people living in the developing countries. As long as the costs of the campaign exceed the expect- ed benefits, political will needed for such a cam- paign is hardly achievable. Returning to the simple principle of population dynamics, the HIV epidemics in Africa and else- where will eventually level off and begin to os- cillate around a mean value depending of the so- cio-economic and other environmental factors. We already now that this mean will be significant- ly higher in sub-Saharan African countries than e.g., in industrial countries. Unless significant sci- entific breakthrough in developing new cheaper drugs or a working easily administered and cheap vaccination materialises or a global political will materialises quickly, no real options for interven- ing the HIV epidemics in Africa can be seen in the near future. This means that most African countries will suffer from colossal human, eco- nomic and intellectual losses in the near future (Fig. 10). What will happen after that in sub-Sa- haran African countries is beyond all reasonable forecasts and will remain to be seen by the next generation – with or without hope. REFERENCES AIDS epidemic update 2002 (2003). 18.2.2003. Anderson RM (1989). Mathematical and statistical studies of the epidemiology of HIV. AIDS 3, 333– 346. Arbona SI & M Löytönen (1997). AIDS epidemic in Puerto Rico – ideas for geography teaching. Fen- nia 175, 125–138. Auvert B, G Buonamico, E Lagarde & B William (2000). Sexual behavior, heterosexual transmis- sion, and the spread of HIV in sub-Saharan Afri- ca: a simulation. Computers and Biomedical Re- search 33, 84–96. Barnett T & P Blaikie (1992). AIDS in Africa: its present and future impact. 193 p. John Wiley & Sons, Chichester. Bond GC, J Kreniske, I Susser & J Vincent (1997). AIDS in Africa and the Caribbean. 234 p. West- View Press, Boulder CO. Bouckenooghe A & W Shandera (1999). HIV trends in African blood donors. Journal of Infection 39, 122–128. Bremermann HJ & RW Anderson (1990). Mathemat- ical models of HIV infection. I. Threshold condi- tions for transmission and host survival. Journal of Acquired Immune Deficiency Syndrome 3, 1129–1134. Chin J & JM Mann (1988). The global patterns of AIDS and HIV infection. AIDS 2, 247–252. Cliff AD & P Haggett (1988). Atlas of disease distri- butions: analytic approaches to epidemiological data. 300 p. Blackwell, Oxford. Cliff AD & P Haggett (1989). Spatial aspects of epi- demic control. Progress in Human Geography 13, 315–347. Fig. 10. Projected loss in life expectancy in selected coun- tries due to HIV/AIDS in children born in 2000. ACKNOWLEDGEMENTS I kindly thank Ms. Carolyn Allsopp of the World Health Organisation for revising the language. 10 FENNIA 181: 1 (2003)Markku Löytönen Cliff AD, P Haggett & JK Ord (1986). Spatial aspects of influenza epidemics. 280 p. Pion, London. Cliff AD, P Haggett, JK Ord & GR Versey (1981). Spa- tial diffusion. An historical geography of epidem- ics in an island community. 238 p. Cambridge University Press, Cambridge. De Kock KM, D Mbori-Ngacha & E Marum (2002). Shadow on the continent: public health and HIV/ AIDS in Africa in the 21st century. The Lancet 360, 67–72. Epstein PR (1995). Emerging diseases and ecosystem instability: new threats to public health. Ameri- can Journal of Public Health 85, 168–172. European Centre for the Epidemiological Monitoring of AIDS (2002). HIV/AIDS Surveillance in Europe. Report No. 61. Saint-Maurice. Flahault A & A-J Valleron (1992). A method for as- sessing the global spread of HIV-1 infection based on air travel. Mathematical Population Studies 3, 161–171. Gao F, E Bailes, DL Robertson, Y Chen, CM Roden- burg, SF Michael, LB Cummins, LO Arthur, M Peeters, GM Shaw, PM Sharp & BH Hahn (1999). Origin of HIV-1 in the chimpanzee Pan troglo- dytes troglodytes. Nature 397, 385–386. Gilks CF, E Katabira & KM De Kock (1997). The chal- lenge of providing effective care for HIV/AIDS in Africa. AIDS 11 (Suppl B), s99-s106. Gould P (1989). Geographic dimensions of the AIDS epidemic. The Professional Geographer 41, 71–78. Gould P (1993). The slow plague: a geography of the AIDS pandemic. 228 p. Blackwell, Oxford. Gould P & R Wallace (1994). Spatial structures and scientific paradoxes in the AIDS pandemic. Geo- grafiska Annaler 76B, 105–116. Haggett P (1994). Geographical aspects of the emer- gence of infectious diseases. Geografiska Anna- ler 76B, 91–104. Lam NS & KB Liu (1994). Spread of AIDS in rural America, 1982–1990. Journal of Acquired Im- mune Deficiency Syndrome 7, 485–490. Lamptey PR (2002). Reducing heterosexual transmis- sion of HIV in poor countries. British Medical Journal 324, 207–211 Low-Beer D & RL Stoneburner (1997). An age- and sex-structured HIV epidemiological model: fea- tures and applications. Bulletin of the World Health Organization 75, 213–221. Löytönen M & SI Arbona (1996). Forecasting the AIDS epidemic in Puerto Rico. Social Science and Medicine 42, 997–1010. Löytönen M (1991). Spatial diffusion of human im- munodeficiency virus type I in Finland, 1982– 1997. Annals of the Association of American Ge- ographers 80, 127–151. Löytönen M (1994a). Growth models and the HIV epidemic in Finland. Social Science and Medi- cine 38, 179–185. Löytönen M (1994b). The Box-Jenkins forecast of HIV-seropositive population in Finland, 1991– 1993. Geografiska Annaler 73B, 121–131. Löytönen M (1995). The effects of the HIV epidemic on the population of Europe. In Hall R & P White (eds). Europe’s population – towards the next cen- tury, 83–98. UCL Press. London. Löytönen M (1998). The AIDS epidemic in Russia. In Unwin T (ed). A European geography, 306– 309. Harlow, Longman. Matshalaga NR & G Powell (2002).Mass rophand- hood in the era of HIV/AIDS. British Medical Jour- nal 324, 185–186. McMichael AJ (1993). Planetary overload, global environmental change and the health of the hu- man species. 352 p. Cambridge University Press, Cambridge. Mertens TE & D Low-Beer (1996). HIV and AIDS: where is the epidemic going? Bulletin of the World Health Organization 74, 121–129. Morgan D, C Mahe, B Mayanja & JAG Whitworth (2002). Progression to symptomatic disease in people infected with HIV-1 in rural Uganda: pro- spective cohort study. British Medical Journal 324, 193–196. Ojo K & M Delaney (1997). Economic and demo- graphic consequences of AIDS in Namibia: rap- id assessment of the costs. International Journal of Health Planning and Management 12, 315- 326. Pan American Health Organization (1997). Health conditions in the Caribbean. PAHO Scientific Publication No. 561, Washington. Ramasundaram S (2002). Can India avoid being dev- astated by AIDS? British Medical Journal 324, 182–183. Rankin W (2002). AIDS and global justice. British Medical Journal 324, 181–182. Shannon GW & GF Pyle (1989). The origin and dif- fusion of AIDS: a view from medical geography. Annals of the Association of American Geogra- phers 79, 1–24. Shannon GW, GF Pyle & RL Bashbur (1991). The geography of AIDS: origins and course of an epi- demic. Guilford, New York NY. Smallman-Raynor MR & AD Cliff (1990). Acquired immune deficiency syndrome (AIDS): literature, geographical origins and global patterns. Progress in Human Geography 14, 157–213. Smallman-Raynor MR & AD Cliff (1991). Civil war and the spread of AIDS in Central Africa. Epide- miology and Infection 107, 69–80. Smallman-Raynor MR, AD Cliff & P Haggett (1992). London international atlas of AIDS. 430 p. Black- well, Oxford. Stoneburner RL, D Low-Beer, GS Tembo, TE Mertens & G Asiimwe-Okiror (1996). Human immunode- ficiency virus infection in East Africa deduced from surveillance data. American Journal of Epi- demiology 144, 682–695. Stoneburner RL, P Sato, A Burton & TE Mertens (1994). The global HIV pandemic. Acta Pediatri- ca 400 (Suppl.), s1–s4. Tarantola D & B Schwartlander (1997). HIV/AIDS FENNIA 181: 1 (2003) 11AIDS and Africa – when will the epidemic level off epidemics in sub-Saharan Africa: dynamism, di- versity and discrete declines. AIDS 11 (Suppl B), s5–s21. The status and trends of the HIV/AIDS epidemics in the world. XIII International AIDS Conference, Durban, South Africa, 5–7 July 2000 (2000). < h t t p : / / w w w. u n a i d s . o r g / wh a t s n e w / o t h e r s / MAP_Provisional.doc> 17.7.2000. Thomas RW (1992). Space–time interactions in mul- tiregion disease modelling. Environment and Planning A 24, 341–360. Thomas RW (1996). Alternative population dynam- ics in selected HIV/AIDS modeling systems: some cross-national comparisons. Geographical Anal- ysis 28, 108–125. Thomas RW (1999). Stability and mixing conditions for HIV/AIDS models with regional compart- ments. Journal of Geographical Systems 1, 1–19.