- Email: [email protected]
Injury prevalence and causality in developing nations: Results from a countrywide population-based survey in Nepal
Injury prevalence and causality in developing nations: Results from a countrywide population-based survey in Nepal Shailvi Gupta, MD, MPH,a,h Evan G. Wong, MD, MPH,b,h Sarthak Nepal, MBBS,c Sunil Shrestha, MBBS,d Adam L. Kushner, MD, MPH, FACS,e,h Benedict C. Nwomeh, MD, MPH, FACS,f,h and Sherry M. Wren, MD, FACS,g Oakland and Stanford, CA, Montreal, Quebec, Canada, Pokhara and Kathmandu, Nepal, Baltimore, MD, Columbus, OH, and New York, NY
Background. Traumatic injury affects nearly 5.8 million people annually and causes 10% of the world’s deaths. In this study we aimed to estimate injury prevalence, to describe risk-factors and mechanisms of injury, and to estimate the number of injury-related deaths in Nepal, a low-income South Asian country. Methods. A cluster randomized, cross-sectional nationwide survey using the Surgeons OverSeas Assessment of Surgical Need tool was conducted in Nepal in 2014. Questions were structured anatomically and designed around a representative spectrum of operative conditions. Two-stage cluster sampling was performed: 15 of 75 districts were chosen randomly proportional to population; within each district, after stratification for urban and rural populations, 3 clusters were randomly chosen. Injury-related results were analyzed. Results. A total of 1,350 households and 2,695 individuals were surveyed verbally, with a response rate of 97%. A total of 379 injuries were reported in 354 individuals (13.1%, 95% confidence interval 11.9–14.5%), mean age of 32.6. The most common mechanism of injury was falls (37.5%), road traffic injuries (19.8%), and burns (14.2%). The most commonly affected anatomic site was the upper extremity (42.0%). Of the deaths reported in the previous year, 16.3% were injury-related; 10% of total deaths may have been averted with access to operative care. Conclusion. This study provides baseline data on the epidemiology of traumatic injuries in Nepal and is the first household-based countrywide assessment of injuries in Nepal. These data provide valuable information to help advise policymakers and government officials for allocation of resources toward trauma care. (Surgery 2015;157:843-9.) From the University of California San Francisco, East Bay,a Oakland, CA; McGill University Centre for Global Surgery,b Montreal, Quebec, Canada; Manipal College of Medical Sciences,c Pokhara, Nepal; Department of Surgery,d Nepal Medical College, Sinamangal, Kathmandu, Nepal; Department of International Health,e Johns Hopkins Bloomberg School of Public Health, Baltimore, MD; Nationwide Children’s Hospital,f Columbus, OH; Stanford School of Medicine,g Stanford, CA; and Surgeons Overseas,h New York, NY
IT IS ESTIMATED THAT 1 PERSON IN THE WORLD DIES EVERY 5 SECONDS as a result of a traumatic injury, which amounts to more than 5.8 million victims a year or 10% of the world’s deaths---32% more than malaria, tuberculosis, and HIV/AIDS combined. The Association for Academic Surgery 2014 Global Surgery Research Fellowship Award and Surgeons OverSeas provided funding for logistics and transportation. Accepted for publication December 3, 2014. Reprint requests: Shailvi Gupta, MD, MPH, University of California San Francisco, East Bay, 1411 East 31st Street, Oakland, CA 94602. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.12.020
Injuries affect younger individuals disproportionately, as well as citizens of low- and middle-income countries (LMICs), with 90% of injury-related deaths occurring in these settings.1 Efforts have been made to quantify the number and impact of injuries across the globe, notably the Global Burden of Diseases, Injuries, and Risk Factors Study 2010.2 This project has been instrumental in highlighting the widespread prevalence and burden of injuries and has provided data to inform policymakers, despite being based primarily on countryspecific data gathered by Ministries of Health, which may underestimate the true disease prevalence in LMICs because of a lack of access to care. The Surgeons OverSeas Assessment of Surgical Need (SOSAS) is a household-based survey of SURGERY 843
844 Gupta et al
operative need that has been implemented successfully in 3 countries: Rwanda, Sierra Leone, and Nepal.3-5 The prevalence of injury using the SOSAS survey has been reported for Rwanda and Sierra Leone previously, revealing that 7.2% and 12.0% of respondents endorsed an injury-related condition in the past year, respectivel6,7; however, countrywide community-level data on injuries outside of Africa is sparse. Thus, the objectives of this study are as follows: (1) to provide an estimate of injury prevalence; (2) to describe injury risk factors and mechanisms; and (3) to estimate the number of injury-related deaths at the community level in Nepal. Nepal is a landlocked, low-income South Asian country (147,181 km2) with a population of 27.8 million.8 The varied terrain of Nepal, which includes the Himalayas in the north, allows only 43% of the population access to all-weather roads. This inability to accessible transportation results in delays in health care within an alreadyoverwhelmed government health care system.8,9 METHODS The SOSAS survey was executed countrywide in Nepal from May 25 to June 12, 2014. SOSAS is a cluster randomized cross-sectional survey, described in more detail previously.10 The SOSAS survey is divided into 2 sections. The first section collects demographic data regarding the household’s access to health care and recent deaths in the household. If a household member died within the past year, further questions regarding health care received, access to operative care, barriers to operative care, and specifics of the death were asked. The second section selects 2 household members randomly; each member undergoes a verbal head-to-toe examination of 6 anatomic regions: (1) face, head, and neck; (2) chest and breast; (3) abdomen; (4) groin and genitalia; (5) back; and (6) extremities. Verbally, each respondent elicits symptoms or experiences associated with a general spectrum of operative conditions, including wounds, masses, deformities, burns, and injuries. If the problem was considered a ‘‘wound due to injury,’’ further information of the type of injury was sought. The choices of injuries in the SOSAS survey included: car/truck/bus collision, motorcycle collision, pedestrian/bicycle crash, gunshot wound, stab or slash, house or work accident, bite or animal attack, fall, or burn due to open fire, hot water, or hot object. Further details, including the timing and mechanism of an injury, were asked, as well as the health-seeking behavior
Surgery May 2015
of the individual during the injury. Questions regarding disability from each injury sustained were asked; forms of perceived disability included: (1) the injury is not disabling; (2) the individual feels ashamed after sustaining the injury; (3) the individual is unable to work as he or she used to; (4) the individual needs help with transport; and (5) the individual requires help with daily living. In addition, a visual physical examination by a Nepali physician was performed on all areas except the groin in both sexes and anterior chest in women. Nepal consists of 75 administrative districts. Two-stage, cluster sampling was performed. First, 15 of the 75 districts were selected randomly proportional to population, after which 45 Village Development Committees (VDCs) were selected randomly, 3 per each district, after stratification for urban and rural population distribution, 2 rural to 1 urban (Fig); this methodology was similar to that used by the Demographic and Health Surveys in Nepal.11 Interviewers began at a central location and sampled every 5th household within a selected VDC; 30 households per VDC were sampled, with a total sample size of 1,350 households countrywide. Estimation of sample size was calculated from a prevalence of unmet operative need of 5%, reported in a pilot study of SOSAS in Nepal in January 2014.12 The surveys were conducted orally by a total of 100 Nepali medical interns and students. All surveys were administered in Nepali, and the responses recorded in English via paper surveys. Because the surveys were verbally conducted, literacy was not a prerequisite to participate in the survey. Institutional Review Board approval was obtained from the Nepal Health Research Council in Kathmandu, Nepal, and Nationwide Children’s Hospital in Columbus, Ohio. Verbal consent was obtained from all respondents before the survey was administered; parental consent, oral assent, and/or parental permission was obtained for individuals younger than 18 years. Individuals noted to be impaired cognitively by the other household members were excluded from the study; household members of all ages were included. Data were analyzed in STATA 130 (StataCorp, College Station, TX). Estimated proportions of injuries were reported with respective 95% confidence intervals (95% CIs) and were compared by sex, age, village type (rural or urban), occupation, education, and literacy via univariate analyses performed with v2 tests. Odds ratios of injuries among various covariates were calculated with a logistic regression model.
Gupta et al 845
Surgery Volume 157, Number 5
Fig. Clusters surveyed in Nepal.
RESULTS A total of 1,350 households and 2,695 individuals were surveyed, with a response rate of 97%. A total of 379 injuries were reported in 354 individuals (13.1%, 95% CI 11.9–14.5%), with a mean age of 32.6 (95% CI 30.7–34.6). Of these 379 injuries, 76 were reported within the past year (20.1%, 95% CI 16.1–24.4%). Demographic data of respondents and individuals who reported at least 1 injury in his or her lifetime are shown in Table I. Using logistic regression, we found that injuries in Nepal were associated with the male sex, advanced age, and an urban village type, though only the association with the male sex was statistically significant. Those individuals with a primary or secondary education had greater odds of sustaining an injury than those with a tertiary or graduate-level education. The most common mechanism of injury was falls (37.5%), followed by road traffic injuries (19.8%) and burns (14.2%). The most commonly affected anatomic site secondary to an injury were the extremities (68.6%); 159 injuries reported in the upper extremity (42%) and 101 injuries reported in the lower extremity (27%) (Table II). Among the 354 individuals who experienced an injury, 42 reported a current injury concurrent with the inability to access operative care (11.9%, 95% CI 8.7–15.7%). Barriers to care included: (1) personnel/facility/equipment are unavailable (n = 21); (2) no money for health care or transportation (n = 15); and (3) fear/no trust (n = 6). With the addition of the visual physical examination, an
observed agreement between the verbal response and visual physical examination findings was 94.6%; such high agreement helps to validate the SOSAS tool. Of the 379 injuries sustained, disability was obtained for 373 injuries (98.4%). Of these, 15.8% of the injuries caused the individual to not be able to work as he or she had before the injury, 3.2% felt ashamed after sustaining the injury, and 2.7% needed help with daily living after sustaining the injury; the injury was not disabling in 76.9% of individuals surveyed. Of the 1,350 households surveyed, 13 of 80 total deaths within the past year were caused by injury (16.3%). Seven deaths were caused by a fall injury, 3 by motorcycle crashes, and 3 by a car/truck/bus collision. Of these 13 deaths caused by injury, 8 deaths may have been averted with access to operative care (10% of the total deaths [n = 80], reported within the past year by head of household interview). Reasons for barriers to operative care included: (1) no time (n = 6); (2) personnel/facility/equipment were not available (n = 1); and (3) fear/no trust (n = 1). DISCUSSION This study is the first nationwide, populationbased assessment of the incidence of injuries in Nepal. We show a nonfatal injury prevalence of 13.1% and a fatal injury prevalence of 16.3% of total deaths. Extrapolating the aforementioned results to the country’s population, we estimate the prevalence of nonfatal injuries to be 3.6
846 Gupta et al
Surgery May 2015
Table I. Demographics of respondents who reported at least one injury in their lifetime Demographic Sex Male Female Total Age category, y 0–14 15–64 $65 Total Village type Rural Urban Total Occupation Unemployed Homemaker Domestic helper Farmer Self-employed Government Nongovernment Total Education None Primary Secondary Tertiary Graduate Total Literacy Illiterate Literate Total
P value
Frequency
Percent
Proportion
Crude OR (95% CI)
227 127 354
64% 36% 100%
15.8 (14.0, 17.8) 10.1 (8.5, 11.9)
1 (Reference) 0.6 (0.5–0.8)
63 268 23 354
18% 76% 6% 100%
10.9 (8.5, 13.7) 13.7 (12.2, 15.3) 14.6 (9.5, 21.2)
1 (Reference) 1.29 (1.0–1.7) 1.40 (0.8–2.4)
.1770
231 123 354
65% 35% 100%
12.9 (11.3, 14.5) 13.7 (11.5, 16.1)
1 (Reference) 1.08 (0.9–1.4)
.542
145 47 11 49 66 12 23 353 (1 missing)
41% 13% 3% 14% 19% 3% 7% 100%
13.1 8.5 17.7 16.5 15.3 8.8 21.9
(11.1, 15.2) (6.3, 11.2) (9.2, 29.5) (12.5, 21.2) (12.1, 19.1) (4.6, 14.8) (14.4, 31.0)
1 0.62 1.43 1.31 1.20 0.64 1.86
(Reference) (0.44–0.88) (0.73–2.81) (0.92–1.86) (0.88–1.65) (0.3–1.2) (1.1–3.1)
.0003
87 104 110 43 10 354
25% 29% 31% 12% 3% 100%
10.4 14.8 15.1 12.0 16.7
(8.4, 12.6) (12.2, 17.6) (12.6, 17.9) (8.8, 15.8) (8.3, 28.5)
1 1.50 1.54 1.18 1.74
(Reference) (1.1–2.0) (1.1–2.1) (0.8–1.7) (0.9–3.6)
.025
94 260 354
73% 27% 100%
11.0 (9.0, 13.3) 14.1 (12.6, 15.8)
1 (Reference) 1.32 (1.0–1.7)
.026
.00
CI, Confidence interval; OR, odds ratio.
million, approximately 31,720 deaths annually are caused by injury, and that 19,460 injury-related deaths could be averted with access to operative care annually. There are similarities and differences in our findings compared with those reported in a 2009 systematic review of all studies reporting injuries or violence in Nepal.13 In this report, men between the ages of 20 and 50 years were the victims of injuries most commonly, and road traffic injuries, especially when motorcycle-related, were the most common mechanism. In our study, road traffic incidents were the second most common cause for injury, and falls were the leading cause of neurotrauma. Our reported results mirror more closely the results in the 2011–2012 Annual Health Report from the Nepal Department of Health Services, in which falls, injuries, and fractures were recorded as a total of 578,568 cases
and were nearly 8 times more prevalent than road traffic incidents, which only accounted for 75,366 cases.11 The 2 referenced Nepalese publications illustrate some of the difficulties in obtaining accurate data on injury prevalence. Both data from the Department of Health and the review article by Joshi and Shrestha13 depend on various data sources, such as surveys, hospital admissions, census, and police reports. Data sources can have variable accuracy, which can influence prevalence data and confound results. For example, in a review of police and mortuary data sets in Western Cape Province, South Africa, substantial problems with data quality were exposed, including duplication, missing data, and substantial underreporting. The estimated completeness of mortuary and police data sets was only 57.6% and 46.4%, respectively.14
Gupta et al 847
Surgery Volume 157, Number 5
Table II. Mechanism of injury by anatomic location Mechanism of injury
Face
Chest
Back
Abdomen
Groin
Extremity
Total
Fall Stab/slash/cut/crush Burn Hot liquid/hot object Open fire/explosion Traffic related Motorcycle crash Car, truck, bus crash Pedestrian, bike Bite or animal attack Gunshot House or work accident Total
31 9
2 0
4 0
1 1
0 0
104 25
142 (38%) 35 (9%)
7 3
2 1
2 2
1 1
2 0
18 15
32 (8%) 22 (6%)
6 5 2 5 0 15 83 (22%)
0 2 0 0 0 2 9 (2%)
1 1 2 0 0 4 16 (4%)
0 0 0 1 1 0 6 (2%)
0 0 0 0 0 0 2 (1%)
29 12 15 8 0 37 260 (69%)
A second important factor in the accuracy of data collection is sampling bias. Because substantial barriers to access medical care exist in LMICs, hospital-based trauma registries underestimate likely the actual prevalence of traumatic injuries, because unknown numbers of injured patients never reach a health center and therefore are not accounted for; only those able to access medical care are accounted for in trauma registries. Therefore, household-based community surveys are complimentary to other modalities of data collection and should be considered additive to other modalities in obtaining valid estimates. It is estimated that Nepal has more than 95 hospitals, including 8 tertiary care centers, 205 primary health care centers, and thousands of primary health posts.11 Yet, despite these health facilities, 11.1% of injuries reported were considered as a current unmet operative need, with the most common barrier to care being that a facility, equipment, or personnel were unavailable. Although many health facilities in Nepal exist, many qualified health professionals are unwilling to work in such low-resource settings given lack of incentive, further contributing to this particular barrier to care.15 To date, however, no comprehensive assessment of Nepal’s trauma system has been performed. With the development of hospital-based tools, such as the Guidelines for Essential Trauma Care by the World Health Organization and index of the International Assessment of the Capacity for Trauma (INTACT), as well as the system-wide Global Trauma System Evaluation Tool (G-TSET), our study suggests a formal assessment of trauma capacity in Nepal is a logical next step and could use these tools.16-18 Even without a formal assessment of the system, the incidence and type of injuries found in our
36 20 19 14 1 58 379
(10%) (5%) (5%) (4%) (0.3%) (15%) (100%)
study have important implications for trauma care in Nepal. First and foremost, injury-prevention programs may have the most impact. On the basis of the findings of this study, targeted interventions for working age men and all those susceptible to falls and road traffic injuries should be prioritized. The prevalence of falls as the leading cause of injury needs to be investigated further as to their location and etiology. Without understanding further the mechanisms of this common injury, it is difficult to target them for possible prevention trainings. To improve access to care, trauma systems should be matured, because this approach has been shown to improve outcomes both in highand low-resource settings.19-22 Specifically, improvements in prehospital transport may break down certain barriers to care. Because these improvements may be resourceintensive, some programs have been successful in using laypersons as first-responders.23-25 Hospitalbased capacities to provide trauma care, notably operative capacity, should be strengthened. As trauma systems are developed, resources should be allocated to health posts and district hospitals for patient stabilization before transport, because supplies have been shown to be particularly tenuous at the primary level in these settings.26 Ultimately, because injury remains an important source of disability, rehabilitation services should be developed. In Nepal, however, only 43% of the population has access to all-weather roads, making travel to health facilities quite a challenge.27 Attempts have been made to propose an emergency medical service plan that advocates for the provision of navigable roads to increase access to mountainous regions, but such construction will take time and strong support from government officials.28
848 Gupta et al
This study inevitably has limitations. As not all districts were sampled directly, and findings were extrapolated to the entire population; therefore, a degree of uncertainty arising from the sampling mechanism exists. Although this survey incorporated a visual physical examination component, because of ethical and cultural reasons, the chest and groin areas, both potentially relevant areas of injury, were not examined. Nevertheless, because the verbal survey alone has been shown to have excellent validity, we are confident of our findings.5 Because the enumeration of deaths from injury relied on the recollection of household members of events, recall bias remains a possibility. The exact timeframe of deaths may have been skewed, certain deaths may have been omitted or forgotten, and without actual autopsies, it is impossible to determine the actual causes of death. Furthermore, the severity of the injuries sustained was not assessed with the SOSAS survey. Even an injury severity score designed for developing countries like Nepal, notably the Kampala Trauma Score, could not be assessed. The Kampala Trauma Score requires 5 key components: (1) systolic blood pressure (mmHg); (2) respiratory rate (breaths per minute); (3) neurologic status (AVPU system); (4) number of serious injuries requiring hospital admission; and (5) age (years).29 Given these parameters are mainly data points only at the hospital level, our study was unable to assess this. To assist in this limitation, we reported perceived disability. All in all, because of the robustness of our sampling mechanism and the use of a verbal, validated survey, we believe that our results are a valid representation of the current state of injuries in Nepal. To conclude, we found that the unmet burden of injuries in this setting remains high and, therefore, provide specific alleys for future assessments of trauma capacity and improvements in injury care. The Nepali enumerators graciously donated their time and expertise.
Surgery May 2015
4.
5.
6.
7.
8.
9.
10.
11.
12.
13. 14.
15. 16. 17.
18.
19.
REFERENCES 1. World Health Organization. Injuries and violence: the facts. Geneva: World Health Organization; 2010. 2. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380: 2095-128. 3. Petroze RT, Groen RS, Niyonkuru F, et al. Estimating operative disease prevalence in a low-income country: results of
20.
21.
22.
a nationwide population survey in Rwanda. Surgery 2013; 153:457-64. Groen RS, Samai M, Stewart KA, et al. Untreated surgical conditions in Sierra Leone: a cluster randomised, crosssectional, countrywide survey. Lancet 2012;380:1082-7. Gupta S, Shrestha S, Ranjit A, Nagarajan N, Groen RS, Kushner AL, Nwomeh BC. Surgical care in Nepal: Conditions, preventable deaths, procedures and validation of a countrywide survey. British Journal of Surgery 2015. In press. Stewart KAA, Groen RS, Kamara TB, et al. Traumatic injuries in developing countries: report from a nationwide cross-sectional survey of Sierra Leone. JAMA Surgery 2013; 148:463-9. Petroze RT, Joharifard S, Groen RS, et al. Injury, Disability and Access to Care in Rwanda: Results of a Nationwide Cross-Sectional Population Study. World J Surg 2015;39: 62-9. The World Bank. Nepal Data. 2014. Available from: http:// data.worldbank.org/country/nepal. Accessed September 22, 2014. Spiegel DA, Shrestha OP, Rajbhandary T, Bijukachhe B, Sitoula P, Banskota B, et al. Epidemiology of surgical admissions to a children’s disability hospital in Nepal. World J Surg 2010;34:954-62. Groen RS, Samai M, Petroze RT, Kamara TB, Yambasu SE, Calland JF, et al. Pilot testing of a population based survey tool in Sierra Leone. World J Surg 2010;36:771-4. Department of Health Services. Ministry of Health and Population, Nepal. Annual Report. 2012. Available from: http:// phpnepal.org/index.php?listId=453-.U6R7iJSSy2E. Accessed August 4, 2014 Gupta S, Ranjit A, Shrestha S, Wong E, Robinson WC, Nwomeh BC, et al. Surgical needs of Nepal: pilot study of population based survey in Pokhara, Nepal. World J Surg 2014; 38:3041-6. Joshi SK, Shrestha S. A study of injuries and violence related articles in Nepal. JNMA J Nepal Med Assoc 2009;48:209-16. Chokotho LC, Matzopoulos R, Myers JE. Assessing quality of existing data sources on road traffic injuries (RTIs) and their utility in informing injury prevention in the Western Cape Province, South Africa. Traffic Inj Prev 2013;14: 267-73. Ghimire P. Nepal may have enough doctors but they’re in the wrong place. BMJ 2014;349:4913. Mock C, Lormand JD, Goosen J, et al. Guidelines for essential trauma care. Geneva: World Health Organization; 2004. Wong EG, Gupta S, Deckelbaum DL, et al. The International Assessment of Capacity for Trauma (INTACT): an index for trauma capacity in low-income countries. J Surg Res 2014;190:522-7. Remick KN, Wong EG, Chep CC, Morton RT, Monsour A, Fisher D, et al. Development of a novel global trauma system evaluation tool and initial results of implementation in the Republic of South Sudan. Injury 2014;45:1731-5. Jurkovich GJ, Mock C. Systematic review of trauma system effectiveness based on registry comparisons. J Trauma 1999;47:S46-55. Mann NC, Mullins RJ, MacKenzie EJ, et al. Systematic review of published evidence regarding trauma system effectiveness. J Trauma 1999;47:S25-33. Nathens AB, Jurkovich GJ, Rivara FP, et al. Effectiveness of state trauma systems in reducing injury-related mortality: a national evaluation. J Trauma 2000;48:25-31. Mock CN, Jurkovich GJ, nii-Amon-Kotei D, et al. Trauma mortality patterns in three nations at different economic
Surgery Volume 157, Number 5
23.
24.
25.
26.
levels: implications for global trauma system development. J Trauma 1998;44:804-12; discussion 812-4. Callese TE, Richards CT, Shaw P, et al. Layperson trauma training in low- and middle-income countries: a review. J Surg Res 2014;190:104-10. Jayaraman S, Mabweijano JR, Lipnick MS, et al. Current patterns of prehospital trauma care in Kampala, Uganda and the feasibility of a lay-first-responder training program. World J Surg 2009;33:2512-21. Murad MK, Husum H. Trained lay first responders reduce trauma mortality: a controlled study of rural trauma in Iraq. Prehosp Disaster Med 2010;25:533-9. Wong EG, Gupta S, Deckelbaum DL, Razek T, Kushner AL. Prioritizing injury care: a review of trauma capacity
Gupta et al 849
in low and middle-income countries. J Surg Res 2015; 193:217-22. 27. The World Bank: working for a world free of poverty. Nepal transport sector. Washington (DC): The World Bank; 2013. Available from: http://go.worldbank.org/I99TRS72B0. Accessed November 13, 2014. 28. Nepal Community Emergency Preparedness Group, Karmacharya PC, Singh GK, Singh MP, Gautam VG, Par A, Banskota AK, et al. Managing the injury burden in Nepal. Clin Orthop Relat Res 2008;466:2343-9. 29. Haac B, Varela C, Geyer A, Cairns B, Charles AG. The utility of the kampala trauma score as a triage tool in SubSaharan African trauma cohort. World J Surg 2015;39: 356-62.
Background. Traumatic injury affects nearly 5.8 million people annually and causes 10% of the world’s deaths. In this study we aimed to estimate injury prevalence, to describe risk-factors and mechanisms of injury, and to estimate the number of injury-related deaths in Nepal, a low-income South Asian country. Methods. A cluster randomized, cross-sectional nationwide survey using the Surgeons OverSeas Assessment of Surgical Need tool was conducted in Nepal in 2014. Questions were structured anatomically and designed around a representative spectrum of operative conditions. Two-stage cluster sampling was performed: 15 of 75 districts were chosen randomly proportional to population; within each district, after stratification for urban and rural populations, 3 clusters were randomly chosen. Injury-related results were analyzed. Results. A total of 1,350 households and 2,695 individuals were surveyed verbally, with a response rate of 97%. A total of 379 injuries were reported in 354 individuals (13.1%, 95% confidence interval 11.9–14.5%), mean age of 32.6. The most common mechanism of injury was falls (37.5%), road traffic injuries (19.8%), and burns (14.2%). The most commonly affected anatomic site was the upper extremity (42.0%). Of the deaths reported in the previous year, 16.3% were injury-related; 10% of total deaths may have been averted with access to operative care. Conclusion. This study provides baseline data on the epidemiology of traumatic injuries in Nepal and is the first household-based countrywide assessment of injuries in Nepal. These data provide valuable information to help advise policymakers and government officials for allocation of resources toward trauma care. (Surgery 2015;157:843-9.) From the University of California San Francisco, East Bay,a Oakland, CA; McGill University Centre for Global Surgery,b Montreal, Quebec, Canada; Manipal College of Medical Sciences,c Pokhara, Nepal; Department of Surgery,d Nepal Medical College, Sinamangal, Kathmandu, Nepal; Department of International Health,e Johns Hopkins Bloomberg School of Public Health, Baltimore, MD; Nationwide Children’s Hospital,f Columbus, OH; Stanford School of Medicine,g Stanford, CA; and Surgeons Overseas,h New York, NY
IT IS ESTIMATED THAT 1 PERSON IN THE WORLD DIES EVERY 5 SECONDS as a result of a traumatic injury, which amounts to more than 5.8 million victims a year or 10% of the world’s deaths---32% more than malaria, tuberculosis, and HIV/AIDS combined. The Association for Academic Surgery 2014 Global Surgery Research Fellowship Award and Surgeons OverSeas provided funding for logistics and transportation. Accepted for publication December 3, 2014. Reprint requests: Shailvi Gupta, MD, MPH, University of California San Francisco, East Bay, 1411 East 31st Street, Oakland, CA 94602. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2014.12.020
Injuries affect younger individuals disproportionately, as well as citizens of low- and middle-income countries (LMICs), with 90% of injury-related deaths occurring in these settings.1 Efforts have been made to quantify the number and impact of injuries across the globe, notably the Global Burden of Diseases, Injuries, and Risk Factors Study 2010.2 This project has been instrumental in highlighting the widespread prevalence and burden of injuries and has provided data to inform policymakers, despite being based primarily on countryspecific data gathered by Ministries of Health, which may underestimate the true disease prevalence in LMICs because of a lack of access to care. The Surgeons OverSeas Assessment of Surgical Need (SOSAS) is a household-based survey of SURGERY 843
844 Gupta et al
operative need that has been implemented successfully in 3 countries: Rwanda, Sierra Leone, and Nepal.3-5 The prevalence of injury using the SOSAS survey has been reported for Rwanda and Sierra Leone previously, revealing that 7.2% and 12.0% of respondents endorsed an injury-related condition in the past year, respectivel6,7; however, countrywide community-level data on injuries outside of Africa is sparse. Thus, the objectives of this study are as follows: (1) to provide an estimate of injury prevalence; (2) to describe injury risk factors and mechanisms; and (3) to estimate the number of injury-related deaths at the community level in Nepal. Nepal is a landlocked, low-income South Asian country (147,181 km2) with a population of 27.8 million.8 The varied terrain of Nepal, which includes the Himalayas in the north, allows only 43% of the population access to all-weather roads. This inability to accessible transportation results in delays in health care within an alreadyoverwhelmed government health care system.8,9 METHODS The SOSAS survey was executed countrywide in Nepal from May 25 to June 12, 2014. SOSAS is a cluster randomized cross-sectional survey, described in more detail previously.10 The SOSAS survey is divided into 2 sections. The first section collects demographic data regarding the household’s access to health care and recent deaths in the household. If a household member died within the past year, further questions regarding health care received, access to operative care, barriers to operative care, and specifics of the death were asked. The second section selects 2 household members randomly; each member undergoes a verbal head-to-toe examination of 6 anatomic regions: (1) face, head, and neck; (2) chest and breast; (3) abdomen; (4) groin and genitalia; (5) back; and (6) extremities. Verbally, each respondent elicits symptoms or experiences associated with a general spectrum of operative conditions, including wounds, masses, deformities, burns, and injuries. If the problem was considered a ‘‘wound due to injury,’’ further information of the type of injury was sought. The choices of injuries in the SOSAS survey included: car/truck/bus collision, motorcycle collision, pedestrian/bicycle crash, gunshot wound, stab or slash, house or work accident, bite or animal attack, fall, or burn due to open fire, hot water, or hot object. Further details, including the timing and mechanism of an injury, were asked, as well as the health-seeking behavior
Surgery May 2015
of the individual during the injury. Questions regarding disability from each injury sustained were asked; forms of perceived disability included: (1) the injury is not disabling; (2) the individual feels ashamed after sustaining the injury; (3) the individual is unable to work as he or she used to; (4) the individual needs help with transport; and (5) the individual requires help with daily living. In addition, a visual physical examination by a Nepali physician was performed on all areas except the groin in both sexes and anterior chest in women. Nepal consists of 75 administrative districts. Two-stage, cluster sampling was performed. First, 15 of the 75 districts were selected randomly proportional to population, after which 45 Village Development Committees (VDCs) were selected randomly, 3 per each district, after stratification for urban and rural population distribution, 2 rural to 1 urban (Fig); this methodology was similar to that used by the Demographic and Health Surveys in Nepal.11 Interviewers began at a central location and sampled every 5th household within a selected VDC; 30 households per VDC were sampled, with a total sample size of 1,350 households countrywide. Estimation of sample size was calculated from a prevalence of unmet operative need of 5%, reported in a pilot study of SOSAS in Nepal in January 2014.12 The surveys were conducted orally by a total of 100 Nepali medical interns and students. All surveys were administered in Nepali, and the responses recorded in English via paper surveys. Because the surveys were verbally conducted, literacy was not a prerequisite to participate in the survey. Institutional Review Board approval was obtained from the Nepal Health Research Council in Kathmandu, Nepal, and Nationwide Children’s Hospital in Columbus, Ohio. Verbal consent was obtained from all respondents before the survey was administered; parental consent, oral assent, and/or parental permission was obtained for individuals younger than 18 years. Individuals noted to be impaired cognitively by the other household members were excluded from the study; household members of all ages were included. Data were analyzed in STATA 130 (StataCorp, College Station, TX). Estimated proportions of injuries were reported with respective 95% confidence intervals (95% CIs) and were compared by sex, age, village type (rural or urban), occupation, education, and literacy via univariate analyses performed with v2 tests. Odds ratios of injuries among various covariates were calculated with a logistic regression model.
Gupta et al 845
Surgery Volume 157, Number 5
Fig. Clusters surveyed in Nepal.
RESULTS A total of 1,350 households and 2,695 individuals were surveyed, with a response rate of 97%. A total of 379 injuries were reported in 354 individuals (13.1%, 95% CI 11.9–14.5%), with a mean age of 32.6 (95% CI 30.7–34.6). Of these 379 injuries, 76 were reported within the past year (20.1%, 95% CI 16.1–24.4%). Demographic data of respondents and individuals who reported at least 1 injury in his or her lifetime are shown in Table I. Using logistic regression, we found that injuries in Nepal were associated with the male sex, advanced age, and an urban village type, though only the association with the male sex was statistically significant. Those individuals with a primary or secondary education had greater odds of sustaining an injury than those with a tertiary or graduate-level education. The most common mechanism of injury was falls (37.5%), followed by road traffic injuries (19.8%) and burns (14.2%). The most commonly affected anatomic site secondary to an injury were the extremities (68.6%); 159 injuries reported in the upper extremity (42%) and 101 injuries reported in the lower extremity (27%) (Table II). Among the 354 individuals who experienced an injury, 42 reported a current injury concurrent with the inability to access operative care (11.9%, 95% CI 8.7–15.7%). Barriers to care included: (1) personnel/facility/equipment are unavailable (n = 21); (2) no money for health care or transportation (n = 15); and (3) fear/no trust (n = 6). With the addition of the visual physical examination, an
observed agreement between the verbal response and visual physical examination findings was 94.6%; such high agreement helps to validate the SOSAS tool. Of the 379 injuries sustained, disability was obtained for 373 injuries (98.4%). Of these, 15.8% of the injuries caused the individual to not be able to work as he or she had before the injury, 3.2% felt ashamed after sustaining the injury, and 2.7% needed help with daily living after sustaining the injury; the injury was not disabling in 76.9% of individuals surveyed. Of the 1,350 households surveyed, 13 of 80 total deaths within the past year were caused by injury (16.3%). Seven deaths were caused by a fall injury, 3 by motorcycle crashes, and 3 by a car/truck/bus collision. Of these 13 deaths caused by injury, 8 deaths may have been averted with access to operative care (10% of the total deaths [n = 80], reported within the past year by head of household interview). Reasons for barriers to operative care included: (1) no time (n = 6); (2) personnel/facility/equipment were not available (n = 1); and (3) fear/no trust (n = 1). DISCUSSION This study is the first nationwide, populationbased assessment of the incidence of injuries in Nepal. We show a nonfatal injury prevalence of 13.1% and a fatal injury prevalence of 16.3% of total deaths. Extrapolating the aforementioned results to the country’s population, we estimate the prevalence of nonfatal injuries to be 3.6
846 Gupta et al
Surgery May 2015
Table I. Demographics of respondents who reported at least one injury in their lifetime Demographic Sex Male Female Total Age category, y 0–14 15–64 $65 Total Village type Rural Urban Total Occupation Unemployed Homemaker Domestic helper Farmer Self-employed Government Nongovernment Total Education None Primary Secondary Tertiary Graduate Total Literacy Illiterate Literate Total
P value
Frequency
Percent
Proportion
Crude OR (95% CI)
227 127 354
64% 36% 100%
15.8 (14.0, 17.8) 10.1 (8.5, 11.9)
1 (Reference) 0.6 (0.5–0.8)
63 268 23 354
18% 76% 6% 100%
10.9 (8.5, 13.7) 13.7 (12.2, 15.3) 14.6 (9.5, 21.2)
1 (Reference) 1.29 (1.0–1.7) 1.40 (0.8–2.4)
.1770
231 123 354
65% 35% 100%
12.9 (11.3, 14.5) 13.7 (11.5, 16.1)
1 (Reference) 1.08 (0.9–1.4)
.542
145 47 11 49 66 12 23 353 (1 missing)
41% 13% 3% 14% 19% 3% 7% 100%
13.1 8.5 17.7 16.5 15.3 8.8 21.9
(11.1, 15.2) (6.3, 11.2) (9.2, 29.5) (12.5, 21.2) (12.1, 19.1) (4.6, 14.8) (14.4, 31.0)
1 0.62 1.43 1.31 1.20 0.64 1.86
(Reference) (0.44–0.88) (0.73–2.81) (0.92–1.86) (0.88–1.65) (0.3–1.2) (1.1–3.1)
.0003
87 104 110 43 10 354
25% 29% 31% 12% 3% 100%
10.4 14.8 15.1 12.0 16.7
(8.4, 12.6) (12.2, 17.6) (12.6, 17.9) (8.8, 15.8) (8.3, 28.5)
1 1.50 1.54 1.18 1.74
(Reference) (1.1–2.0) (1.1–2.1) (0.8–1.7) (0.9–3.6)
.025
94 260 354
73% 27% 100%
11.0 (9.0, 13.3) 14.1 (12.6, 15.8)
1 (Reference) 1.32 (1.0–1.7)
.026
.00
CI, Confidence interval; OR, odds ratio.
million, approximately 31,720 deaths annually are caused by injury, and that 19,460 injury-related deaths could be averted with access to operative care annually. There are similarities and differences in our findings compared with those reported in a 2009 systematic review of all studies reporting injuries or violence in Nepal.13 In this report, men between the ages of 20 and 50 years were the victims of injuries most commonly, and road traffic injuries, especially when motorcycle-related, were the most common mechanism. In our study, road traffic incidents were the second most common cause for injury, and falls were the leading cause of neurotrauma. Our reported results mirror more closely the results in the 2011–2012 Annual Health Report from the Nepal Department of Health Services, in which falls, injuries, and fractures were recorded as a total of 578,568 cases
and were nearly 8 times more prevalent than road traffic incidents, which only accounted for 75,366 cases.11 The 2 referenced Nepalese publications illustrate some of the difficulties in obtaining accurate data on injury prevalence. Both data from the Department of Health and the review article by Joshi and Shrestha13 depend on various data sources, such as surveys, hospital admissions, census, and police reports. Data sources can have variable accuracy, which can influence prevalence data and confound results. For example, in a review of police and mortuary data sets in Western Cape Province, South Africa, substantial problems with data quality were exposed, including duplication, missing data, and substantial underreporting. The estimated completeness of mortuary and police data sets was only 57.6% and 46.4%, respectively.14
Gupta et al 847
Surgery Volume 157, Number 5
Table II. Mechanism of injury by anatomic location Mechanism of injury
Face
Chest
Back
Abdomen
Groin
Extremity
Total
Fall Stab/slash/cut/crush Burn Hot liquid/hot object Open fire/explosion Traffic related Motorcycle crash Car, truck, bus crash Pedestrian, bike Bite or animal attack Gunshot House or work accident Total
31 9
2 0
4 0
1 1
0 0
104 25
142 (38%) 35 (9%)
7 3
2 1
2 2
1 1
2 0
18 15
32 (8%) 22 (6%)
6 5 2 5 0 15 83 (22%)
0 2 0 0 0 2 9 (2%)
1 1 2 0 0 4 16 (4%)
0 0 0 1 1 0 6 (2%)
0 0 0 0 0 0 2 (1%)
29 12 15 8 0 37 260 (69%)
A second important factor in the accuracy of data collection is sampling bias. Because substantial barriers to access medical care exist in LMICs, hospital-based trauma registries underestimate likely the actual prevalence of traumatic injuries, because unknown numbers of injured patients never reach a health center and therefore are not accounted for; only those able to access medical care are accounted for in trauma registries. Therefore, household-based community surveys are complimentary to other modalities of data collection and should be considered additive to other modalities in obtaining valid estimates. It is estimated that Nepal has more than 95 hospitals, including 8 tertiary care centers, 205 primary health care centers, and thousands of primary health posts.11 Yet, despite these health facilities, 11.1% of injuries reported were considered as a current unmet operative need, with the most common barrier to care being that a facility, equipment, or personnel were unavailable. Although many health facilities in Nepal exist, many qualified health professionals are unwilling to work in such low-resource settings given lack of incentive, further contributing to this particular barrier to care.15 To date, however, no comprehensive assessment of Nepal’s trauma system has been performed. With the development of hospital-based tools, such as the Guidelines for Essential Trauma Care by the World Health Organization and index of the International Assessment of the Capacity for Trauma (INTACT), as well as the system-wide Global Trauma System Evaluation Tool (G-TSET), our study suggests a formal assessment of trauma capacity in Nepal is a logical next step and could use these tools.16-18 Even without a formal assessment of the system, the incidence and type of injuries found in our
36 20 19 14 1 58 379
(10%) (5%) (5%) (4%) (0.3%) (15%) (100%)
study have important implications for trauma care in Nepal. First and foremost, injury-prevention programs may have the most impact. On the basis of the findings of this study, targeted interventions for working age men and all those susceptible to falls and road traffic injuries should be prioritized. The prevalence of falls as the leading cause of injury needs to be investigated further as to their location and etiology. Without understanding further the mechanisms of this common injury, it is difficult to target them for possible prevention trainings. To improve access to care, trauma systems should be matured, because this approach has been shown to improve outcomes both in highand low-resource settings.19-22 Specifically, improvements in prehospital transport may break down certain barriers to care. Because these improvements may be resourceintensive, some programs have been successful in using laypersons as first-responders.23-25 Hospitalbased capacities to provide trauma care, notably operative capacity, should be strengthened. As trauma systems are developed, resources should be allocated to health posts and district hospitals for patient stabilization before transport, because supplies have been shown to be particularly tenuous at the primary level in these settings.26 Ultimately, because injury remains an important source of disability, rehabilitation services should be developed. In Nepal, however, only 43% of the population has access to all-weather roads, making travel to health facilities quite a challenge.27 Attempts have been made to propose an emergency medical service plan that advocates for the provision of navigable roads to increase access to mountainous regions, but such construction will take time and strong support from government officials.28
848 Gupta et al
This study inevitably has limitations. As not all districts were sampled directly, and findings were extrapolated to the entire population; therefore, a degree of uncertainty arising from the sampling mechanism exists. Although this survey incorporated a visual physical examination component, because of ethical and cultural reasons, the chest and groin areas, both potentially relevant areas of injury, were not examined. Nevertheless, because the verbal survey alone has been shown to have excellent validity, we are confident of our findings.5 Because the enumeration of deaths from injury relied on the recollection of household members of events, recall bias remains a possibility. The exact timeframe of deaths may have been skewed, certain deaths may have been omitted or forgotten, and without actual autopsies, it is impossible to determine the actual causes of death. Furthermore, the severity of the injuries sustained was not assessed with the SOSAS survey. Even an injury severity score designed for developing countries like Nepal, notably the Kampala Trauma Score, could not be assessed. The Kampala Trauma Score requires 5 key components: (1) systolic blood pressure (mmHg); (2) respiratory rate (breaths per minute); (3) neurologic status (AVPU system); (4) number of serious injuries requiring hospital admission; and (5) age (years).29 Given these parameters are mainly data points only at the hospital level, our study was unable to assess this. To assist in this limitation, we reported perceived disability. All in all, because of the robustness of our sampling mechanism and the use of a verbal, validated survey, we believe that our results are a valid representation of the current state of injuries in Nepal. To conclude, we found that the unmet burden of injuries in this setting remains high and, therefore, provide specific alleys for future assessments of trauma capacity and improvements in injury care. The Nepali enumerators graciously donated their time and expertise.
Surgery May 2015
4.
5.
6.
7.
8.
9.
10.
11.
12.
13. 14.
15. 16. 17.
18.
19.
REFERENCES 1. World Health Organization. Injuries and violence: the facts. Geneva: World Health Organization; 2010. 2. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380: 2095-128. 3. Petroze RT, Groen RS, Niyonkuru F, et al. Estimating operative disease prevalence in a low-income country: results of
20.
21.
22.
a nationwide population survey in Rwanda. Surgery 2013; 153:457-64. Groen RS, Samai M, Stewart KA, et al. Untreated surgical conditions in Sierra Leone: a cluster randomised, crosssectional, countrywide survey. Lancet 2012;380:1082-7. Gupta S, Shrestha S, Ranjit A, Nagarajan N, Groen RS, Kushner AL, Nwomeh BC. Surgical care in Nepal: Conditions, preventable deaths, procedures and validation of a countrywide survey. British Journal of Surgery 2015. In press. Stewart KAA, Groen RS, Kamara TB, et al. Traumatic injuries in developing countries: report from a nationwide cross-sectional survey of Sierra Leone. JAMA Surgery 2013; 148:463-9. Petroze RT, Joharifard S, Groen RS, et al. Injury, Disability and Access to Care in Rwanda: Results of a Nationwide Cross-Sectional Population Study. World J Surg 2015;39: 62-9. The World Bank. Nepal Data. 2014. Available from: http:// data.worldbank.org/country/nepal. Accessed September 22, 2014. Spiegel DA, Shrestha OP, Rajbhandary T, Bijukachhe B, Sitoula P, Banskota B, et al. Epidemiology of surgical admissions to a children’s disability hospital in Nepal. World J Surg 2010;34:954-62. Groen RS, Samai M, Petroze RT, Kamara TB, Yambasu SE, Calland JF, et al. Pilot testing of a population based survey tool in Sierra Leone. World J Surg 2010;36:771-4. Department of Health Services. Ministry of Health and Population, Nepal. Annual Report. 2012. Available from: http:// phpnepal.org/index.php?listId=453-.U6R7iJSSy2E. Accessed August 4, 2014 Gupta S, Ranjit A, Shrestha S, Wong E, Robinson WC, Nwomeh BC, et al. Surgical needs of Nepal: pilot study of population based survey in Pokhara, Nepal. World J Surg 2014; 38:3041-6. Joshi SK, Shrestha S. A study of injuries and violence related articles in Nepal. JNMA J Nepal Med Assoc 2009;48:209-16. Chokotho LC, Matzopoulos R, Myers JE. Assessing quality of existing data sources on road traffic injuries (RTIs) and their utility in informing injury prevention in the Western Cape Province, South Africa. Traffic Inj Prev 2013;14: 267-73. Ghimire P. Nepal may have enough doctors but they’re in the wrong place. BMJ 2014;349:4913. Mock C, Lormand JD, Goosen J, et al. Guidelines for essential trauma care. Geneva: World Health Organization; 2004. Wong EG, Gupta S, Deckelbaum DL, et al. The International Assessment of Capacity for Trauma (INTACT): an index for trauma capacity in low-income countries. J Surg Res 2014;190:522-7. Remick KN, Wong EG, Chep CC, Morton RT, Monsour A, Fisher D, et al. Development of a novel global trauma system evaluation tool and initial results of implementation in the Republic of South Sudan. Injury 2014;45:1731-5. Jurkovich GJ, Mock C. Systematic review of trauma system effectiveness based on registry comparisons. J Trauma 1999;47:S46-55. Mann NC, Mullins RJ, MacKenzie EJ, et al. Systematic review of published evidence regarding trauma system effectiveness. J Trauma 1999;47:S25-33. Nathens AB, Jurkovich GJ, Rivara FP, et al. Effectiveness of state trauma systems in reducing injury-related mortality: a national evaluation. J Trauma 2000;48:25-31. Mock CN, Jurkovich GJ, nii-Amon-Kotei D, et al. Trauma mortality patterns in three nations at different economic
Surgery Volume 157, Number 5
23.
24.
25.
26.
levels: implications for global trauma system development. J Trauma 1998;44:804-12; discussion 812-4. Callese TE, Richards CT, Shaw P, et al. Layperson trauma training in low- and middle-income countries: a review. J Surg Res 2014;190:104-10. Jayaraman S, Mabweijano JR, Lipnick MS, et al. Current patterns of prehospital trauma care in Kampala, Uganda and the feasibility of a lay-first-responder training program. World J Surg 2009;33:2512-21. Murad MK, Husum H. Trained lay first responders reduce trauma mortality: a controlled study of rural trauma in Iraq. Prehosp Disaster Med 2010;25:533-9. Wong EG, Gupta S, Deckelbaum DL, Razek T, Kushner AL. Prioritizing injury care: a review of trauma capacity
Gupta et al 849
in low and middle-income countries. J Surg Res 2015; 193:217-22. 27. The World Bank: working for a world free of poverty. Nepal transport sector. Washington (DC): The World Bank; 2013. Available from: http://go.worldbank.org/I99TRS72B0. Accessed November 13, 2014. 28. Nepal Community Emergency Preparedness Group, Karmacharya PC, Singh GK, Singh MP, Gautam VG, Par A, Banskota AK, et al. Managing the injury burden in Nepal. Clin Orthop Relat Res 2008;466:2343-9. 29. Haac B, Varela C, Geyer A, Cairns B, Charles AG. The utility of the kampala trauma score as a triage tool in SubSaharan African trauma cohort. World J Surg 2015;39: 356-62.