Archives of Academic Emergency Medicine. 2020; 8(1): e65 OR I G I N A L RE S E A RC H Incidence and Risk Factors of Clinical Deterioration dur- ing Inter-Facility Transfer of Critically Ill Patients; a Co- hort Study Kannapatch Srithong1, Siriorn Sindhu2∗, Napaporn Wanitkun3, Chukiat Viwatwongkasem4 1. DNS candidate, Faculty of Graduate Studies, Mahidol University, Thailand. 2. RN, PhD, Associate Professor, Department of Surgical nursing, Faculty of Nursing, Mahidol University, Thailand. 3. RN, PhD, Assistant Professor, Department of Surgical nursing, Faculty of Nursing, Mahidol University, Thailand. 4. PhD, Associate Professor, Biostatistics Department, Faculty of Public Health, Mahidol University, Thailand. Received: June 2020; Accepted: June 2020; Published online: 4 July 2020 Abstract: Introduction: Critically ill and injured patients are at a higher risk of developing clinical deterioration during inter-facility transfers. This study aimed to determine the incidence rate and risk factors of clinical deterioration among critically ill patients during inter-facility transfers in Thailand. Methods: The present cohort study was conducted in 22 referring hospitals and 7 receiving hospitals under the supervision of Ministry of Public Health, Thailand, between March 15 and December 31, 2018. The subjects were comprised of 839 critically ill patients aged 18 and over, 63 coordinator nurses in referral centers, and 312 referral team leaders. Data collected included pre-transfer risk score, clinical data of patient during transfer, characteristics of referral team leader, ambulance type, preparation time, time to definitive care, transfer distance, and National Early Warning Score (NEWS) (clin- ical deterioration). Multilevel mixed-effects regression analysis was performed. Results: The incidence rate of clinical deterioration was 28.69%. The most common types of clinical deterioration were hemodynamic instabil- ity, respiratory instability, and neurological alteration. Time between 31-45 minutes was significantly associated with clinical deterioration (β 0.133, P value 0.027). The following illnesses were associated with higher proba- bility of clinical deterioration: body region injuries/head injury/burn/ingested poison (β 0.670, P value 0.030), respiratory distress/convulsion (β 0.919, P value 0.001), shock/ arrhythmias/chest pain/hemorrhage (β 1.134, P value <0.001), comatose/alteration of consciousness/syncope (β 1.343, P value <0.001), and post-cardiac arrest (β 2.251, P value <0.001). Patients with unstable conditions (β 1.689, P value 0.001) and pre-transfer risk score of 8 or higher (β 0.625, P value 0.001) had a higher rate of deterioration. Transfer by non- emergency room (ER) nurses (β 0.495, P value 0.008) and transportation in a mobile intensive care unit (ICU) were associated with a higher rate of deterioration (β 0.848, P value 0.001). Conclusion: The incidence of clinical deterioration during inter-facility transfer in Thailand was high. Illnesses involving circulatory, respiratory, and neurological systems, clinical instability, high pre-transfer risk score, transport time of 31-45 minutes, transportation by non-ER nurse, and mobile ICU were associated with a higher rate of clinical deterioration. Keywords: Patient transfer; critical illness; clinical deterioration; Thailand Cite this article as: Srithong K, Sindhu S, Wanitkun N, Viwatwongkasem Ch. Incidence and Risk Factors of Clinical Deterioration during Inter-Facility Transfer of Critically Ill Patients; a Cohort Study. Arch Acad Emerg Med. 2020; 8(1): e65. ∗Corresponding Author: Siriorn Sindhu; Faculty of Nursing, Mahidol Univer- sity, Bangkok, Thailand. No. 2 Wang Lang Road, Siriraj, Bangkoknoi, Bangkok 10700, Thailand, Email: siriorn.sin@mahidol.edu, Phone: +668-1817-6060, Fax: +662-412-8415. 1. Introduction Clinical deterioration, including cardiac arrest, can occur during inter-facility transfer of critically ill patients. Recent evidence on the incidence rate of clinical deterioration dur- ing inter-facility transfer is quite limited. Two studies con- ducted in the Netherlands reported incidence rates between 15.2 and 34.0% (1, 2). Two studies in Canada and Saudi Ara- bia reported the incidence rates of 6.5% and 13.6%, respec- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem K. Srithong et al. 2 tively (3, 4). A number of risk factors for inter-facility clini- cal deterioration have been examined in international litera- ture. These factors can be classified into patient-related fac- tors and service-related factors (1-5). Clinical deteriorations have been found to be associated with age, sex, clinical conditions, and level of patient acuity (4, 5). Increasing age, being male, sustaining illnesses involv- ing circulatory, respiratory, and neurological systems, and having high acuity increase the risk of clinical deteriorations. The service factors included pre-transfer risk score, during- transfer care, characteristics of referral team leader, ambu- lance type, preparation time, time to definitive care, and transfer distance (1-4, 6). Higher pre-transfer risk score predicts higher probability of clinical deterioration (3, 6). Patients receiving appropri- ate management until achieving clinical stabilization before transfer are less likely to experience clinical deterioration (4, 5, 7). During-transfer care includes assessment and moni- toring of the patient as well as interventions given to the pa- tient. In the Netherlands, Ligtenberg et al. (2006) reckoned that the lack of during-transfer care protocols contributed to high incidence of clinical deterioration (1). A subsequent Dutch study, which incorporated a during-transfer protocol, reported a much lower incidence rate for clinical deteriora- tion (2). During the transfer, patients managed by doctors, nurses, and paramedics who underwent critical care trans- port training were less likely to experience clinical deteriora- tion (2, 3). Patients being transported in an ambulance fully equipped with critical care technologies (i.e., mobile inten- sive care unit (ICU)) were less likely to experience clinical de- terioration (2, 7). Increasing preparation time, time to defini- tive care, and transfer distance increase clinical deterioration incidence (4). Recent studies have not found an association between clinical deterioration and post-transfer outcomes such as ICU admission and death (3, 8). This may be due to limitations related to sampling and outcome measurement. Nonetheless, in hospitalized patients, clinical deterioration is often associated with more serious health outcomes such as ICU admission and death. Clinical deterioration is still an issue of great concern. Approximately 52-91% of clinical de- terioration incidents are preventable. Clinical deterioration can be prevented via early detection (9). Early detection of clinical deterioration leads to timely management of deteri- orating changes and therefore, prevents progression to fatal events such as cardiopulmonary arrest. In developed countries, inter-facility transfer of critically ill patients is carried out either by well-equipped ambulance centers or tertiary care hospitals (2, 7, 10-12). Staff respon- sible for transporting the patients in these countries must undergo critical care transport training to ensure suitable knowledge and skills in managing patients’ conditions and clinical deterioration (2, 3). In Thailand, transfers of criti- cally ill patients is carried out by resource-limited commu- nity hospitals. These resource-deprived community hospi- tals rely on emergency room (ER) nurses and non-ER nurses with varying qualifications and experiences when transport- ing the patients. The majority of these nurses do not have for- mal training in inter-facility transfer. Lack of knowledge and skills among the referring nurses can compromise the abil- ity of nurses in detecting early warning signs, managing de- terioration, and preventing and managing these conditions during the transfer. Variations in patient transfer structures, practices and outcomes can be expected due to varying ma- terial and human resources as well as the lack of consensus practice guidelines among these hospitals. Since clinical deterioration during the transfer can be harm- ful, data about its incidence and risk factors are important for health policymakers and practitioners to understand the situation and note gaps in existing inter-facility transfer ser- vices. Such knowledge is needed for evidence-based decision making to improve regulations, structural support, and hu- man resource development to improve early detection, pre- vention, and management of patient deterioration during the transfer. This cohort study aimed to determine the in- cidence and risk factors of inter-facility transfer clinical dete- rioration. 2. Methods 2.1. Study design and setting This cohort study took place in Thailand, between March 15 and December 31, 2018. Stratified random sampling was used. Seven health service networks were randomly selected from 12 networks (excluding Bangkok Metropolitan Adminis- tration). Seven provinces were then selected from these net- works. Based on the proportion, two to six community hos- pitals within these provinces were then randomly recruited into the study as referring hospitals. In case there were more than one receiving hospitals in the province, the most ad- vanced hospital was selected as receiving hospital. The study involved 22 referring hospitals and 7 receiving hospitals un- der the supervision of Ministry of Public Health across Thai- land. This study concerned only inter-facility transfers that took place within the province (i.e., intra-province transfers). Transfers between provinces were not included in this study. The study protocol was approved by the Institutional Re- view Board, Faculty of Nursing, Mahidol University (No: IRB- NS2017/423.2512). To comply with local human research ethics requirements, additional ethics approvals were also sought from seven receiving hospitals before the commence- ment of the study. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 3 Archives of Academic Emergency Medicine. 2020; 8(1): e65 2.2. Participants Critically ill patients, referring nurses, and coordinator nurses comprised the study samples. The patient sample size calculation was based on power of analysis, level of signifi- cance, effect size, intraclass correlation (ICC) coefficient, and possible sample size of the participants (13). First, we de- termined the effect size (d) based on data from the previous study on inter-facility transfer (14) with power of analysis and level of significance of 0.80 and 0.05, respectively. Second, we determined the sample size of the referring hospitals based on the desired intra-class correlation (ICC) of 0.30 and the expected sample size of 30 patients per hospital. The calcu- lation resulted in the referring hospital sample size of 22. The patient sample size was, therefore, 660 (i.e., 30 x 22). To deal with potential loss of data, we increased the sample size by 30% or 198 patients. After rounding up, the final sample size was 880. 2.3. Inclusion and exclusion criteria Patients To be included in the study, the patients had to be at least 18 years of age, treated in the emergency room at referring hospitals, and critically ill as per patient acuity classification. Patients diagnosed with psychiatric and obstetric emergen- cies were excluded. Referring and coordinators nurses As per the national clinical practice guideline, at least two health practitioners are assigned to the transfer of critically ill patients. Only the nurses who were designated as team leader of the inter-facility transfer team were included in the study regardless of their experience and qualifications. Co- ordinator nurses who were on duty at the receiving hospitals during the transfer were included. 2.4. Variables Independent variables measurement Independent variables included patient sex and age, level of patient acuity, clinical signs and symptoms, and pre-transfer risk score, referring nurse’s qualification, referring nurse’s experience, preparation time, type of ambulance, during- transfer reassessment, transfer distance, transfer time, and time to definitive treatment. Pre-transfer risk score was as- sessed using Risk Score of Interhospital Transport of Criti- cally Ill Patients (RSTP) (6). An RSTP score of 8 or higher was considered as an indication of high risk for clinical de- terioration. Inter-rater reliability of RSTP among the re- searcher and 14 research assistants was 0.83 based on Kappa coefficient (15). Nurse’s qualifications were described as emergency nurse practitioner (ENP), ER nurse, and non-ER nurse. Nurse’s inter-facility transfer experience was mea- sured as number of years involved in interfacility transfer (IFT). Preparation time was the duration between the time when the patient first presented to ER to the time when re- ferral decision was made. Time to definitive treatment was the duration between the time when the referral decision was made to the time when the patient received treatment at the receiving hospital. These times were measured in minutes. Transfer distance was the distance between referring hospi- tal and receiving hospital and was measured in kilometers. Content validity of other instruments was obtained through a panel of experts. These data were collected from inter-facility transfer records from referring and receiving hospitals as well as on the ambulance. Outcome measurement The studied outcome was clinical deterioration. The follow- ing clinical conditions occurring during the transfer was de- fined as incidence of clinical deterioration: systolic blood pressure (SBP) < 90 mmHg or > 160 mmHg, SpO2 < 92% for intubated patients or < 90% or baseline reduction ≥ 5% in adults, respiratory rate ≥ 26 bpm or ≤ 10 bpm, heart rate ≥ 140 bpm or ≤ 40 bpm, increased chest pain, arrhythmias, car- diac arrest, a drop in Glasgow Coma Score ≥ 2, and convul- sion. Patients who experienced any of the above-mentioned conditions at any given time during the transfer were re- garded as having clinical deterioration. The National Early Warning Score (NEWS) was used to assess changes in clini- cal deterioration score. Internationally, clinical deterioration can be detected using different tools such as Rapid Emer- gency Medicine Score (REM), Early Warning Score (EWS), Risk Score for Transfer Patient (RSTP), Modified Early Warn- ing Score (MEWS), Therapeutic Intervention Scoring System (TISS-28), and National Early Warning Score (NEWS) (6, 16- 19). Among these instruments, NEWS is the most effective in predicting serious clinical deterioration and was found to be sensitive to cardiopulmonary arrest (19). This study, there- fore, used NEWS to assess patient deterioration. The score was derived from the assessment of 7 physiological param- eters as follows: respiratory rate, oxygen saturation, supple- mental oxygen, temperature, systolic blood pressure, heart rate, and level of consciousness (19, 20). NEWS parame- ters were recorded by referring nurse in the patient trans- fer form at six different time intervals throughout the trans- fer, namely, 1-15 minutes, 16-30 minutes, 31-45 minutes, 46-60 minutes, 61-75 minutes, and 76-90 minutes. The re- searcher and research assistants then retrieved the data from the record form. 2.5. Statistical Analysis Multilevel mixed-effects linear regression analysis was done using STATA statistical software version 16.0. Findings were presented as a model of factors predicting clinical deteriora- tion during inter-facility transfer. P values less than 0.05 were considered statistically significant. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem K. Srithong et al. 4 Table 1: Baseline characteristics of studied patients Variable Values Patient demographics Sex Male 509 (60.67) Female 330 (39.33) Age (year) Mean ± SD (range) 59.57±18.00 (18-99) Clinical signs and symptoms of illness (a) Hemiplegia/Paraplegia/Severe pain/High fever 110 (13.11) (b) Body region injuries/Head injury/Burn/Ingested poison 113 (13.47) (c) Respiratory distress/Convulsion 280 (33.37) (d) Shock/Arrhythmia/Chest pain/Hemorrhage 202 (24.08) (e) Comatose/Alteration of consciousness/Syncope 95 (11.32) (f ) Post cardiac arrest 39 (4.65) Patient acuity level Stable with medium risk of deterioration 34 (4.05) Stable with high risk of deterioration 707 (84.27) Unstable with clinical deterioration 98 (11.68) Pre-transfer risk score Low risk (≤7 scores) 330 (39.33) High risk (≥8 scores) 509 (60.67) Mean ± SD (range) 8.00±2.61 (0-20) Inter-facility service demographics Transfer team leader’ qualification (n=312) Non-ER nurse 98 (31.41) ER nurse 173 (55.45) ENPs or ENs 41 (13.14) Transfer team leader’s experience (year) Mean ± SD (range) 9.51±7.54 (1-36) Ambulance types EMS ambulance 622 (74.14) Mobile ICU ambulance 217 (25.86) During transfer reassessment More than 15 minutes 306 (36.47) Every 5-15 minutes 533 (63.53) Preparation time (minutes) Mean ± SD (range) 72.10±28.35 (12-120) Time to definitive treatment (minute) Mean ± SD (range) 52.38±21.97 (16-170) Transfer distance (kilometer) Mean ± SD (range) 59.97±28.31 (8.9-110.0) Data are presented as mean ± standard deviation (SD) or frequency (%). ER: emergency room, ENP: emergency nurse practitioner, EN: emergency nurse, EMS: emergency medical service, ICU: intensive care unit. 3. Results 3.1. Baseline characteristics of inter-facility transfer system 839 critically ill patients comprised the final patient sample in this study. The majority were male (60.67%). The mean age was 59.57 ± 18.00 (18 - 99) years. The majority suffered from critical illnesses involving circulatory and respiratory systems (57.45%) and were classified as stable with high risk of deterioration (84.27%). 60.67% of the patients were con- sidered as having a high risk for clinical deterioration. The majority of referring team leaders were ER nurses (55.45%). The mean inter-facility transfer experiences were 9.51 ± 7.54 (<1-36) years. The mean preparation time and time to defini- tive treatment were 72.10 ± 28.35 (12-120) and 52.38 ± 21.97 (16-170) minutes, respectively. Transfer distance was 52.97 ± 28.31 (8.9-110.0) kilometers. 3.2. Clinical deterioration The incidence of clinical deterioration during inter-facility transfer was 28.37% (n = 238/839). Unstable patients had a higher rate of clinical deterioration (63.27%; n = 62) com- pared to stable patients with a high risk of deterioration (24.33%; n = 172) and stable patients with a medium risk of This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 5 Archives of Academic Emergency Medicine. 2020; 8(1): e65 deterioration (11.76%; n = 4). Overall, the average NEWS of critically ill patient sample during each timeframe was be- tween 4.57 to 4.90. The NEWS of individual patients ranged from 0 to 20. The highest average score was 4.90 ± 2.77, which was measured during 31-45 minutes. Considering clinical stability, stable patients with high risk of deterioration had an average NEWS between 4.69 and 4.92 with the highest aver- age score of 4.92 ± 2.80 at 76-90 minutes. High average NEWS (between 6.35 and 7.12) was observed among unstable pa- tients with the highest average score of 7.12 ± 3.20 at 46-60 minutes. 3.3. Predictors of clinical deterioration during transfer The following variables, which were found to have significant association (P value < 0.25) with clinical deterioration in a univariate analysis, were included in subsequent multilevel mixed-effects linear regression analysis: patient sex, clinical signs and symptoms of illnesses, patient acuity level, pre- transfer risk score, transfer team leader’s qualification, trans- fer team leader’s experience, during transfer reassessment, ambulance types, preparation time, time to definitive treat- ment, and transfer distance. In this study, individual patients with critical illness had an average NEWS of 2.403 throughout the transfer. Clinical signs and symptoms involving circula- tory, respiratory, and neurological systems, unstable condi- tions, pre-transfer high risk score, transfer by non-ER nurses, transport by Mobile ICU, and 31-45 minutes transfer time- frame were found to be associated with a significant increase in NEWS (Table 3). 4. Discussion This study examined clinical deterioration of critically ill pa- tients during inter-facility transfer. The inter-facility clinical deterioration rate was 28.37% (n = 238). Symptoms of clinical deterioration could be classified into hemodynamic instabil- ity, respiratory instability, and neurological alteration. The classification resembled previous studies in Canada, Saudi Arabia, and Australia (3, 4, 10). The major results described patient, service, and time- related factors associated with patient deterioration as mea- sured by NEWS. Illnesses involving circulatory, respira- tory, and neurological systems significantly increased clin- ical deterioration incidences when compared with motor deficit/severe pain/high fever. Unstable conditions signifi- cantly increased NEWS by 1.689 compared with stable condi- tions with a medium risk of deterioration. High pre-transfer risk score (≥ 8) significantly increased the NEWS by 0.625 compared with low pre-transfer risk score (≤ 7). A significant increase of NEWS by 0.133 was found 31-45 minutes after transfer when compared with NEWS before transfer. Transfer by non-ER nurses increased the NEWS by 0.495 when com- pared to the transfer by ER nurses. Transfer by mobile ICU increased NEWS by 0.848 when compared with transfers by standard ambulance. In our study, it is clear that patients having circulatory, res- piratory, and neurological illnesses, unstable clinical condi- tions and pre-transfer risk score were more likely to expe- rience clinical deterioration during the transfer. In a previ- ous study by Alabdali et al. (2017), cardiac patients were less likely to develop clinical deterioration (adjusted OR: 0.117, 95% CI: 0.02 to 0.52 and p value <0.01) (3). In another study, cardiac patients (crude OR: 0.6, 95% CI: 0.4 to 0.8), neuro- logic patients (crude OR: 0.4, 95% CI: 0.3 to 0.6), and trauma patients (crude OR: 0.5, 95% CI: 0.3 to 0.8) were less likely to develop clinical deterioration (4). Our findings appear to contradict with previous studies. It is, however, of note that the way we categorized patients’ illness in this study differed from that of the previous two studies. Our study used clini- cal signs and symptoms as a basis for illness categorization whereas the previous studies relied on medical diagnoses. Clinical signs and symptoms and medical diagnoses are not always the same. Patients with cardiac diagnosis, for exam- ple, do not always present with cardiac signs and symptoms. This also applies to patients with neurologic and trauma di- agnoses. We believe that clinical signs and symptoms bet- ter reflect patient conditions and needs, since they change during critical illness unlike medical diagnosis. Clinical signs and symptoms offered a better basis for illness categoriza- tion as they were more sensitive to clinical deterioration as evident in this study; whereas, medical diagnoses were not. In addition, clinical signs and symptoms help guide appro- priate interventions related to physiological changes, which are common in critically ill patients. Patient acuity and pre-transfer risk score could predict pa- tient deterioration during the transfer. This is not surprising as these variables were derived from data concerning clini- cal deterioration, for example, low blood pressure, low heart rate, decreased level of consciousness, increased respiratory rate, oxygen desaturation, and cardiac arrest. Our findings were consistent with previous studies in Canada, Saudi Ara- bia, and Hong Kong (3-5). Lee et al. (2010) reported that physiological instability before the transfer was a significant risk factor of clinical deterioration (P value 0.004). Singh et al. (2014) reported that patients with baseline hemodynamic instability were more likely to develop clinical deterioration during the transfer (crude OR: 3.9, 95% CI: 3.1 to 4.9). Al- abdali et al. (2017) reported that patients with pre-transfer risk score of 6 or higher were 1.3 times more likely to expe- rience clinical deterioration. Our study has confirmed that patient acuity and pre-transfer risk score were predictive of clinical deterioration, and thus can be used as tools to antic- ipate changes in patient conditions and plan for proper pre- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem K. Srithong et al. 6 Table 2: National Early Warning Score (NEWS) classified by time of measurement and patient acuity level NEWS (0-20) Time of measurement Time 0 before transfer Time 1 1-15 minutes Time 2 16-30 minutes Time 3 31-45 minutes Time 4 46-60 minutes Time 5 61-75 minutes Time 6 76-90 minutes Total critically ill patients (n=839) (n=839) (n=776) (n=657) (n=497) (n=258) (n=80) Mean ± SD 4.70±2.66 4.74±2.75 4.73±2.76 4.90±2.77 4.84±2.85 4.57±2.70 4.86±3.02 Range (0 to 14) (0 to 18) (0 to 18) (0 to 20) (0 to 20) (0 to 15) (0 to 13) Stable with medium risk of deterioration (n=34) (n=34) (n=32) (n=25) (n=25) (n=16) (n=4) Mean ± SD 0.00±0.00 0.21±0.64 0.31±1.00 0.32±1.14 0.24±1.01 0.00±0.00 0.00±0.00 Range (0) (0 to 3) (0 to 4) (0 to5) (0 to 5) (0) (0) Stable with high risk of deterioration (n=707) (n=707) (n=662) (n=564) (n=422) (n=222) (n=66) Mean ± SD 4.64±2.43 4.69±2.05 4.74±2.59 4.89±2.56 4.84±2.53 4.74±2.44 4.92±2.80 Range (3 to 13) (2 to 13) (3 to 17) (3 to 17) (3 to 17) (2 to 15) (3 to 13) Unstable with clinical deterioration (n=98) (n=98) (n=82) (n=68) (n=50) (n=20) (n=10) Mean ± SD 6.67±2.56 6.72±2.90 6.35±2.74 6.68±2.91 7.12±3.29 6.40±2.87 6.40±3.20 Range (3 to 14) (3 to 18) (3 to 18) (4 to 20) (4 to 20) (3 to 11) (3 to 13) SD: standard deviation. vention and management during the transfer. Only one study identified the effect of time on clinical de- terioration. Clinical deterioration increased 1.15 times for each additional 10 minutes of transport time (4). Longer ride leads to accumulative effects of physiological changes asso- ciated with vibrations and inertial force during ambulance moving and these effects can contribute to the occurrence of deterioration (11). Acceleration, deceleration, and vibrations in a moving car are associated with hemodynamic changes. These changes include low blood pressure, interference with blood flow to the brain, and increased heart rate and respira- tory rate (12). Our study also found the effect of time on clinical deterio- ration. However, a significant increase of NEWS was found only 31-45 minutes after departure when compared with de- parture time. Upon close examination, we found that the increase in NEWS 31-45 minutes after departure was due to the occurrence of serious clinical deterioration due to cardiac arrest, hypotension, oxygen desaturation, and alteration of consciousness in this period. Why these serious events oc- curred more frequently during this period warrants explana- tion. Due to limited knowledge about this occurring, we at- tempt to hypothesize this for future research. First, the ad- verse effects of certain medical treatments received at the re- ferring hospital may fall within this period. Such treatments are, for example, thrombolytic agent and bronchodilator. Peak of action of bronchodilator is between 30-60 minutes. The adverse reactions that can present during this time frame include palpitation, hypertension, tachycardia, chest pain, dyspnea, and paradoxical bronchospasm. Streptokinase, is known to induce arrhythmias and hypotension, which nor- mally occur during the first 15-30 minutes after infusion. In Thailand, streptokinase is infused for patients with myocar- dial infarction at the referring community hospital before de- parture. Currently there are two common practices regard- ing streptokinase infusion. First, the referring hospital com- pletes the infusion before commencing the transfer. Second, the referring hospital commences the transfer immediately after starting the infusion. The latter practice might lead to more frequent occurrence of clinical deterioration due to un- managed adverse drug reactions during the early stages of transfer. Another explanation concerns the sustainability of therapeu- tic effects of medical treatments received at the referring hos- pital. For example, optimal blood pressure in patients with septic shock and trauma can be achieved through fluid resus- citation in ER. However, this resuscitative effect is best seen during the first 30 minutes and only lasts for 60 minutes. This may reintroduce clinical instability (i.e., hypotension) during transfer, especially when there is no or inadequate interven- tion to maintain optimal hemodynamics. Our data showed that a number of critically ill patients developed new and re- current clinical deterioration throughout the transfer. The patients transferred by non-ER nurses were more likely to experience clinical deterioration compared to those trans- ferred by ER nurses and ENPs. As per the quality mandate in Thailand, all ER nurses must undergo compulsory ad- vanced life support annually. In-house trainings related to pre-hospital care, inter-facility transfer and mass casualty management were also very common among these nurses. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 7 Archives of Academic Emergency Medicine. 2020; 8(1): e65 Table 3: Multilevel mixed-effects analysis to predict clinical deterioration Variables Multivariable analysis β 95%CI P value Fixed effect Intercept (β0 ) 2.403 (1.580–3.226) <0.001 NEWS Score during-transfer time Time 0 (Before transfer) ref. Time 1 (1-15 minutes) 0.037 (-0.071–0.145) 0.504 Time 2 (16-30 minutes) 0.104 (-0.007–0.216) 0.067 Time 3 (31-45 minutes) 0.133 (0.015–0.251) 0.027* Time 4 (46-60 minutes) 0.106 (-0.025–0.236) 0.112 Time 5 (61-75 minutes) 0.009 (-0.158–0.176) 0.918 Time 6 (76-90 minutes) 0.242 (-0.036 0.520) 0.087 Clinical signs and symptoms of illness (a) Hemiplegia/Paraplegia/Severe pain/High fever ref. (b) Body region injuries/Head injury/Burn/Ingested poison 0.670 (0.065–1.274) 0.030* (c) Respiratory distress/Convulsion 0.919 (0.371–1.467) 0.001* (d) Shock/Arrhythmias/Chest pain/Hemorrhage 1.134 (0.569–1.698) <0.001* (e) Comatose/Alteration of consciousness/Syncope 1.343 (0.661–2.026) <0.001* (f ) Cardiac arrest 2.251 (1.175–3.328) <0.001* Patient Acuity Level Stable with medium risk of deterioration ref. Stable with high risk of deterioration 0.651 (-0.172–1.475) 0.121 Unstable with clinical deterioration 1.689 (0.663–2.715) 0.001* Pre-transfer risk score RSTP ≤7 scores ref. RSTP ≥ 8 scores 0.625 (0.269–0.981) 0.001* Referral team leader ER nurses ref. Non-ER nurses 0.495 (0.132–0.857) 0.008* ENPs or ENs 0.064 (-0.519–0.647) 0.829 Ambulance type Standard ambulance ref. Mobile ICU 0.848 (0.349–1.347) 0.001* * Significant P value. NEWS: National Early Warning Score; RSTP: Risk Score for Transport Patients, ER: emergency room; ENP: emergency nurse practitioner, EN: emergency nurse, ICU: intensive care unit; ref.: reference. In addition, ER nurses were involved in inter-facility trans- fers on a much more regular basis than non-ER nurses. The ENPs, on the other hand, completed a four-month intensive training program plus continuing staff development activi- ties undertaken by ER nurses. This set of knowledge and skills acquired through trainings and experiences made the ER nurses and ENPs more efficient in managing critically ill patients during the transfer, compared to non-ER nurses who rarely underwent these trainings and infrequently at- tended to critically ill patients. Our finding was consistent with those described in previous studies. Patients transferred by paramedics with training in critical care transport ex- perienced clinical deterioration significantly less than those transferred by paramedics without critical care transport (3). This finding was indicative of the role of knowledge and skills of transferring personnel in prevention and management of clinical deterioration. This study was the first to examine the role of ambulance type on clinical deterioration. Our original hypothesis was that patients transferred by standard ambulance would be at higher risk of experiencing clinical deterioration than those transferred by a mobile ICU. To our surprise, the result was the opposite. Our study revealed that transfer by mobile ICU ambulance increased NEWS by 0.848 when compared to transfer by standard ambulance. We reckon that the increase in clinical deterioration rate in mobile ICU ambulance re- flected the detectability rather than the actual occurrence of clinical deterioration. In a standard ambulance with no con- tinuous monitoring, it was likely that fewer signs of clinical deterioration were detected and recorded than they actually occurred. Continuous monitoring in mobile ICU ambulance facilitated rapid detection and documentation of abnormal changes in physiologic parameters such as heart rate, respi- ration, blood pressure and oxygen saturation. Sex and age of the patients, preparation or resuscitation time, frequency of patient reassessment during transfer, inter- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem K. Srithong et al. 8 facility transfer experience of nurses, transfer distance, and time to definitive treatment were not found to be associated with clinical deterioration. We expected to find the associ- ation between age and clinical deterioration as in a previ- ous study (4), however, our result was different. Sex was not also associated with clinical deterioration. These variables might have served as confounders in this study. Referral ex- perience of nurses, which was measured as number of years involved in inter-facility patient transfer was not associated with clinical deterioration. Preparation time was not associ- ated with the increase in NEWS despite our previous under- standing that preparation time could suggest patient sever- ity, which would result in higher probability of subsequent deterioration. In the Thai contexts, referring hospitals are ex- pected to stabilize the patient before they initiate the transfer. This preparation period may be short or long, depending on whether the patients achieve stabilization goals. Preparation time was, therefore, not an important factor in determining patient safety (i.e., clinical deterioration) but achievement of stable conditions before departure was. For us, the number of years in inter-facility transfer did not equate expertise in inter-facility transfer, which was found to affect clinical dete- rioration incidence. This is especially true for non-ER nurses who did not have enough training and were put in charge of patient transfer once in a while, even though they had been involved in inter-facility transfer for a number of years. It is, therefore, important that we consider expertise but not the number of practice years when choosing nurses for patient transfer. The Canadian study revealed the effect of transfer distance on clinical deterioration (4). In our study, the distance was not associated with clinical deterioration. The longest dis- tance in this study, which concerned only within-provincial transport, was 110 KMs; whereas, the shortest was 8.9 KMs. From the result, we could only say that the increase in dis- tance did not increase clinical deterioration incidence, when the distance was between 8.9-110 KMs. It is of practical note to point out that patient transport involving distances longer than 110 KMs, as in inter-provincial transport, may not be as safe. The longer distance might produce different results (i.e., increase in clinical deterioration) and system redesign may be needed to ensure safety during transport in such sit- uations. This is, however, subject to further investigation. The increase in time to definitive treatment did not increase NEWS. All the patients in this study reached the receiving hospitals within 170 minutes since the referring hospital con- tacted the receiving hospital for referral. Again, we could only say that this timeframe did not result in a higher rate of clinical deterioration. Further study is required to determine whether it is still as safe beyond this timeframe, or not. Clinical deterioration during inter-facility transfer of criti- cally ill patients can be better prevented and managed by pre-departure assessment of clinical deterioration risks us- ing associated clinical parameters such as patient acuity and signs and symptoms, adequate stabilization of the patients, designation of ENP or ER nurses as transfer leader, close monitoring of deteriorating signs through mobile ICU, and continuing management during transport. 5. Limitation Limitations concerning the completeness of clinical data recorded during transfer might have been present. Despite our briefing on documentation requirements, variations in nurses’ knowledge and skills could have impacted the qual- ity of documentation and thus, quality of the data collected. 6. Conclusion The incidence of clinical deterioration during inter-facility transfer of critically ill patients between community hospi- tals and provincial hospitals in Thailand was 28.7%. Patients with signs and symptoms involving circulatory, respiratory, and neurologic systems, unstable clinical conditions, and high pre-transfer risk scores were at risk of experiencing clini- cal deterioration. Achievement of stabilization before depar- ture, regardless of preparation time, was a safe practice. Pa- tients transferred by ENPs did not experience an increase in clinical deterioration rate; whereas, those managed by non- ER nurses did. Mobile ICU detected more deterioration inci- dents than standard ambulances. An increase in clinical de- terioration rate could be expected 31-45 minutes after depar- ture from the referring hospital. A timeframe between 16-170 minutes and transfer distance between 8.9-110 KMs did not cause any differences in clinical deterioration rate. Within the context of our study, these time and distance limits were, therefore, considered reasonably safe for patient transport. 7. Declarations 7.1. Acknowledgements We thank all the study participants for taking part in this study as well as health professionals for their facilitation dur- ing data collection. 7.2. Author contribution All the authors met the criteria of authorship based on the recommendations of the international committee of medical journal editors. Authors ORCIDs Kannapatch Srithong: 0000-0003-1535-543X Siriorn Sindhu: 0000-0001-9326-757X Napaporn Wanitkun: 0000-0002-0512-1368 Chukiat Viwatwongkasem: 0000-0003-4918-2717 This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem 9 Archives of Academic Emergency Medicine. 2020; 8(1): e65 7.3. Funding/Support No funding and support were received for this study. 7.4. Conflict of interest The authors declared that they have no potential conflicts of interests with respect to the research, authorship, and/or publication of this article. References 1. Ligtenberg JJM, Arnold LG, Stienstra Y, van der Werf TS, Meertens JHJM, Tulleken JE, et al. Quality of interhospital transport of critically ill patients: a prospective audit. Crit Care. 2005;9(4):R446. 2. van Lieshout EJ, Binnekade J, Reussien E, Dongelmans D, Juffermans NP, de Haan RJ, et al. Nurses versus physician-led interhospital critical care transport: a randomized non-inferiority trial. Intensive Care Med. 2016;42(7):1146-54. 3. Alabdali A, Trivedy C, Aljerian N, Kimani PK, Lilford R. Incidence and predictors of adverse events and out- comes for adult critically ill patients transferred by paramedics to a tertiary care medical facility. J Health Spec. 2017;5(4):206. 4. Singh JM, MacDonald RD, Ahghari M. Critical events dur- ing land-based interfacility transport. Ann Emerg Med. 2014;64(1):9-15. e2. 5. Lee LLY, Lo WYL, Yeung KL, Kalinowski E, Tang SYH, Chan JTS. Risk stratification in providing inter-facility transport: experience from a specialized transport team. World J Emerg Med. 2010;1(1):49. 6. Markakis C, Dalezios M, Chatzicostas C, Chalkiadaki A, Politi K, Agouridakis PJ. Evaluation of a risk score for in- terhospital transport of critically ill patients. Emerg Med J. 2006;23(4):313-7. 7. Wiegersma JS, Droogh JM, Zijlstra JG, Fokkema J, Ligten- berg JJM. Quality of interhospital transport of the criti- cally ill: impact of a Mobile Intensive Care Unit with a specialized retrieval team. Crit Care. 2011;15(1):R75. 8. Strauch U, Bergmans DCJJ, Winkens B, Roekaerts PMHJ. Short-term outcomes and mortality after interhospital intensive care transportation: an observational prospec- tive cohort study of 368 consecutive transports with a mobile intensive care unit. BMJ open. 2015;5(4). 9. Droogh JM, Smit M, Absalom AR, Ligtenberg JJM, Zijlstra JG. Transferring the critically ill patient: are we there yet? Crit Care. 2015;19(1):62. 10. Fan E, MacDonald RD, Adhikari NK, Scales DC, Wax RS, Stewart TE, et al. Outcomes of interfacility critical care adult patient transport: a systematic review. Crit Care. 2005;10(1):R6. 11. ATOTW 330-Physiological effects of trans- fer for critically ill patients [Internet]. 2016 [cited 28th June 2017]. Available from: https://www.wfsahq.org/components/com_virtual_libr ary/media/e25fd136d2b7ebef4092cd6b949dab52- 330-Physiological-effects-of-transfer-for-critically-ill- patients.pdf. 12. Ambulance Services, Reliability Problems and Poten- tial Technologies [Internet]. 2010 [cited 28th Octpber 2018]. Available from: https://web.wpi.edu/Pubs/E- project/Available/E-project-042910- 085147/unrestricted/IQP_for_Submissionlin.pdf. 13. Scherbaum CA, Ferreter JM. Estimating statistical power and required sample sizes for organizational research us- ing multilevel modeling. Organizational Research Meth- ods. 2009;12(2):347-67. 14. Faine BA, Noack JM, Wong T, Messerly JT, Ahmed A, Fuller BM, et al. Interhospital transfer delays appropri- ate treatment for patients with severe sepsis and sep- tic shock: a retrospective cohort study. Crit Care Med. 2015;43(12):2589-96. 15. McHugh ML. Interrater reliability: the kappa statistic. Biochem Med (Zagreb). 2012;22(3):276-82. 16. Groarke JD, Gallagher J, Stack J, Aftab A, Dwyer C, Mc- Govern R, et al. Use of an admission early warning score to predict patient morbidity and mortality and treatment success. Emerg Med J. 2008;25(12):803-6. 17. Imhoff BF, Thompson NJ, Hastings MA, Nazir N, Mon- cure M, Cannon CM. Rapid Emergency Medicine Score (REMS) in the trauma population: a retrospective study. BMJ open. 2014;4(5):e004738. 18. Lee LLY, Yeung KL, Lo WYL, Lau YSC, Tang SYH, Chan JTS. Evaluation of a simplified therapeutic interven- tion scoring system (TISS-28) and the modified early warning score (MEWS) in predicting physiological de- terioration during inter-facility transport. Resuscitation. 2008;76(1):47-51. 19. Smith GB, Prytherch DR, Meredith P, Schmidt PE, Feath- erstone PI. The ability of the National Early Warning Score (NEWS) to discriminate patients at risk of early car- diac arrest, unanticipated intensive care unit admission, and death. Resuscitation. 2013;84(4):465-70. 20. Royal College of Physicians. National Early Warning Score (NEWS): Standardising the assessment of acute- illness severity in the NHS. Report of a working party. London: Royal College of Physicians; 2012. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: http://journals.sbmu.ac.ir/aaem Introduction Methods Results Discussion Limitation Conclusion Declarations References