Trauma kills as many people a year in the world as HIV/AIDS, tuberculosis, malaria, and breast, prostate, and colorectal cancer combined.1 Injuries and violence are often neglected according to the World Health Organization, despite being predictable and largely preventable.2 South Africa is experiencing a high level of deaths due to interpersonal violence and road injuries.3,4 Homicides and deaths as a result of road traffic injuries are 4 times higher in South Africa than globally.4 The incidence of trauma is so high that it attracts surgeons from many other countries to travel to South Africa, where they can receive training in trauma surgery.
Early deaths due to hemorrhage have been reduced following the introduction of Advanced Trauma Life Support courses, damage control surgery, computed tomography, and the organization of trauma centers.5 The in-hospital trauma mortality rate has also been reduced worldwide, but it has varied substantially from 5.1% to 21.4% in different countries during the past 20 years.6–18
In response to the excessive trauma burden in South Africa, many trauma centers have adopted a systematic trauma care strategy as advocated by Advanced Trauma Life Support. In spite of this, the in-hospital trauma mortality rate remains unknown, and it is unclear whether deficits in hospital care are contributing to the high level of trauma-related mortality when compared with other countries. It was hypothesized that the severity of trauma is the most important factor for trauma-related mortality. We aimed to review the in-hospital trauma mortality rate, to analyze the cause and risk factors of death, and to estimate the proportion of avoidable and unavoidable in-hospital deaths using a trauma database at the Pietermaritzburg Metropolitan Trauma Service, KwaZulu-Natal Province, South Africa.
The study retrospectively analyzed prospectively captured data on all consecutive patients hospitalized for trauma at the Department of Surgery at Grey’s Hospital, Pietermaritzburg, South Africa, during a period of 6 years from January 1, 2013, to January 1, 2019. The Pietermaritzburg Metropolitan Trauma Service serves a catchment population of more than 4 million people and includes 2 public hospitals, Grey’s hospital and Edendale hospital. Grey’s hospital is the tertiary trauma center of the Western area of KwaZulu-Natal. The Department of Surgery at Grey’s Hospital has 70 beds and the intensive care unit have 10 beds to serve the whole hospital. It is the largest academic surgical trauma center in the western KwaZulu-Natal Province. The trauma service adheres to Advanced Trauma Life Support and the doctors have attended basic and advanced courses in trauma care.19 The emergency medical services system includes many providers who retrieve patients from the scene to hospitals in the area. These retrievals are coordinated via the Pietermaritzburg Metropolitan Trauma Service, and patients are triaged to hospitals based on acuity, severity, and actual resources. There was no neurosurgical service in Pietermaritzburg, and all patients requiring a neurosurgical procedure were transferred to the department of neurosurgery in the city of Durban, 80 kilometers away from Grey’s hospital.
The Hybrid Electronic Medical Registry and the study protocols were approved according to the Declaration of Helsinki by the Biomedical Research Ethics Committee at the University of KwaZulu-Natal (ethics number BCA 207/09 and BCA 221/13).
The primary outcome was in-hospital mortality among patients hospitalized because of trauma. The secondary outcome was the classification of avoidable and unavoidable deaths. All cases of in-hospital mortality were presented at a weekly morbidity and mortality conference and classified into avoidable or unavoidable deaths using the Delphi technique by an expert panel of attending surgeons.20 Deaths were reviewed independently by each panelist and were then classified as avoidable or unavoidable according to the majority.
The Hybrid Electronic Medical Registry is a validated trauma database at Pietermaritzburg Metropolitan Trauma Service.21 Patient data are entered prospectively by a surgeon at the emergency department on every trauma admission and at discharge. All data are quality controlled every day by 2 senior trauma surgeons. They also work out the Abbreviated Injury Scale and the Injury Severity Score. Data on all surgical trauma patients were collected from the Hybrid Electronic Medical Registry starting on January 1, 2013, and ending in January 2019. Cases that were dead on arrival at the hospital and cases found not to be related to trauma were excluded from analysis.
Age, gender, race, date and time of admission, mechanism of injury, vital signs, Glasgow coma scale score, systolic blood pressure, respiratory rate, and admission notes were recorded on arrival at the hospital. Date of discharge or death, regions injured, cause of death, clinical course notes, surgical notes, and radiological notes were recorded during admission. Mechanisms of injury were categorized as road traffic accidents including motor vehicle accidents, pedestrian vehicle accidents, and motorcycle accidents, assaults, falls, gunshot wounds, stab wounds, others, and unknown. The Abbreviated Injury Scale and the Injury Severity Score were calculated and recorded prospectively in the database.22,23
Categorical variables are described as frequency and percent and analyzed with the χ2 test. Interval variables are described as the median and interquartile range (IQR) and different groups are compared with the Mann-Whitney U test. Clopper-Pearson confidence intervals (CIs) were used for death rates. Odds ratios and 95% CIs were calculated for in-hospital death in a crude and adjusted logistic regression model with variables of mechanism of trauma, Injury Severity Score, age, and gender. The statistical tests were 2-sided and a significance level of 5% was used. IBM Corp. Released 2017. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY: IBM Corp was used for statistical analysis.
A total of 9795 patients were hospitalized because of trauma during a period of 6 years, from January 1, 2013, to January 1, 2019. There were 8119 men (83%) and 1676 women, their mean (SD) age was 30 (15) years, and they had a mean Injury Severity Score of 10.4 (9.2). Four hundred twelve of them (4.2%; 95% CI, 3.8–4.6) died during in-hospital care (Fig. 1). A total of 7062 adult men were hospitalized and 332 of them died (4.7%), whereas there were 1223 adult females of whom 54 died (4.4%), 1061 boys younger than 18 years of age, 18 of whom died (1.7%), and 449 girls, of whom 8 died (1.8%).
Fifty percent of patients (n = 205) died because of traumatic brain injury, 20% (n = 83) because of multiple organ failure, 13% (n = 53) because of hemorrhagic shock, and 8% (n = 33) because of respiratory failure. The cause of death was unknown in 9%. Patients who died in hospital had a higher median Injury Severity Score of 24.5 (IQR, 16.0–32.0) at admission compared with 9.0 (IQR, 4.0–13.0) in survivors (P < 0.001). Sixty-six percent of the patients who died had experienced a blunt trauma versus 58% of survivors (P = 0.002). Patients who died were older, with a median age of 32 years (IQR, 26–45 years) versus 28 years (IQR, 22–37 years) in survivors (P < 0.001) (Table 1).
TABLE 1. -
Basal Characteristic of Patients Who Died In-Hospital and Survivors
||In-Hospital Deaths, n = 412
||Survivors, n = 9383
|Blunt/penetrating trauma (%)
|Age, years, median (IQR)
|Injury Severity Score, median (IQR)
†Mann-Whitney U test.
Forty-six percent of patients (n = 190) died after road traffic accidents, 19% (n = 78) after gunshot wounds, 13% (n = 53) after stab wounds, 10% (n = 41) after assaults, and 5% (n = 19) after falls. The remainder were due to other or unknown trauma (Fig. 2 and Table 2). The death rate was highest after road traffic accidents (7.0%) and after gunshot wounds (6.6%), but, despite this, the Injury Severity Score among them was similar to other mechanisms of trauma (Table 2). In a multivariable analysis, it was found that the mechanism of trauma, the Injury Severity Score, and age were independently related to increased risk of death, while gender was not (Table 3). An increase of 10 degrees in the Injury Severity Score was related to a threefold increase in the odds of death, while an increase in 10 years increased this odds by 1.3.
TABLE 2. -
Age, Injury Severity Score, and Death Rate in Relation to Mechanism of Injury
|Mechanisms of trauma
||All Admitted Patients
||Death Rate% (95% CI)
||Age, Median (IQR)
||Injury Severity Score, Median (IQR)
||Age, Median (IQR)
||Injury Severity Score, Median (IQR)
|Road traffic accidents
TABLE 3. -
Risk Factors for In-Hospital Death
||Crude OR (95% CI)
||Adjusted OR* (95% CI)
| Road traffic accidents
| Stab wounds
| Gunshot wounds
| Other mechanisms
|Injury Severity Score (10 degrees)
|Age (10 years)
*Adjusted logistic regression model with variables of mechanism, Injury Severity Score, age, and gender.
OR indicates odds ratio.
Thirty-two percent of the deaths occurred on the wards, 29% at the emergency department, 25% at the intensive care unit, 6% in the operating theater, and 8% had a missing place of death. Deaths on the ward and at the emergency department were mainly due to severe traumatic brain injury. At the intensive care unit, deaths were mainly due to multiple organ failure and, in the operating theater, they were due to hemorrhagic shock, in all but 1 patient.
Avoidable and Unavoidable In-Hospital Deaths
According to the Delphi system, unavoidable deaths were considered in 72% (n = 296) of the 412 patients who died in hospital and 16% of deaths were avoidable (n = 67). No decision was made in 12% (n = 49) of deaths because they were either lost for discussion at a morbidity and mortality conference or because no consensus was reached at the conference. Deaths due to traumatic brain injury were avoidable in 5% and unavoidable in 80%. Deaths due to hemorrhagic shock were avoidable in 23% and unavoidable in 72%. Deaths due to multiple organ failure were avoidable in 25% and unavoidable in 67%. Deaths due to respiratory failure were avoidable in 48% and unavoidable in 45% (Fig. 3). Some 47% of avoidable deaths due to respiratory failure and 25% of deaths from multiple organ failure occurred at the ward and the remainder died at the intensive care unit.
Cases of unavoidable death had higher median Injury Severity Scores of 25.0 (IQR, 16.0–34.0) at admission, compared with avoidable cases that had a score of 17.0 (IQR, 9.0–25.0) (P < 0.001), while there was no difference in age. Unavoidable deaths occurred after a median of 1.0 day (IQR, 0.0–5.0 days), which was earlier than avoidable deaths, occurring after 5.0 days (IQR, 2.0–14.8 days) (P < 0.001) (see Figure, Supplemental Digital Content 1, https://links.lww.com/AOSO/A172, which illustrates time from admission to death for avoidable and unavoidable deaths).
The 67 avoidable deaths were due to several factors and were often combined. The main factor was regarded as medical failure or inappropriate care in 37 cases, resource limitations in 28 cases, and lack of data in 2 cases. Resource limitations included dialysis in 9 cases, lack of advanced respiratory care in 7, lack of intensive care unit for different reasons in 4, lack of a local neurosurgeon for onsite neurosurgical service in 4 cases, lack of blood in 3, and no immediate access to the operating theater in 1.
Deaths due to multiple organ failure were avoidable in 25% (n = 21). They were due to resource limitations of dialysis at the intensive care unit in 9 patients, intensive unit care for several reasons including dialysis in one, incorrect surgical decision or delay in repeat operation in 6 patients and failure to recognize a deterioration and to resuscitate in another 5 patients.
Deaths due to respiratory failure were avoidable in 48% (n = 16). This included lack of ventilators and antibiotics at the intensive care unit in 7 patients, most often for hospital-acquired pneumonia, no immediate access to the operating theater in 1 patient, failure to recognize life-threatening conditions in 6 patients, incorrect surgical decision in 1 patient, and inappropriate medical care in 1 patient.
Deaths due to hemorrhagic shock were avoidable in 23% (n = 12). No access to blood was the cause of death in 2 patients, while limited access to blood contributed to death in most of the patients who died because of hemorrhagic shock. One patient died on the table, as there no resuscitative endovascular balloon occlusion of the aorta catheter was available. Failure to recognize life-threatening conditions with failure to resuscitate was the primary cause in 5 cases. Two of them had pelvic bleeds and the decision to perform pre-pelvic packing was delayed inappropriately, failure to apply a tourniquet on arrival to a mangled limb contributed to mortality in 1 patient, 1 patient with a traumatic amputation was resuscitated in the emergency room instead of being taken directly to the operating theater and failure to be resuscitated on the table because of a stabbed heart was the cause in the fifth patient. Three patients died because of incorrect surgical decision-making and 1 patient died because he was given too high a dose of adrenaline after the repair of a stabbed heart injury.
Deaths due to traumatic brain injury were avoidable in 5% (n = 10). Four died because of no local neurosurgeon being available for surgery. In 3 of them, no bed was available in Durban and 1 patient had to wait unreasonably long for the transfer. Three patients died because of no bed at the intensive care unit. Two died because of medical problems with intubation failure in 1 patient and 1 patient fell out of bed, deteriorated and died. One patient died because of failure to recognize a life-threatening condition.
This is the first study to report in-hospital mortality after trauma in South Africa. During a period of 6 years, a total of 9795 patients were hospitalized after trauma at the Department of Surgery at Grey’s Hospital, Pietermaritzburg. The in-hospital trauma mortality rate was only 4.2%. Still, 16% of the death cases were regarded as avoidable, either due to resource limitations or inappropriate medical care. The mechanism of trauma, Injury Severity Score and age were independently related to an increased risk of death at the hospital. Road traffic accidents and gunshot wounds were the deadliest mechanisms of trauma. Traumatic brain injury was the single largest contributor to mortality with half of the deaths and 80% of them were regarded as unavoidable. Unavoidable deaths occurred at an early stage, after a median of 1 day, and patients who died from unavoidable causes had higher Injury Severity Scores at admission than patients whose deaths were classified as avoidable.
The present in-hospital trauma mortality rate at a single South African trauma center of 4.2% is lower than the in-hospital trauma mortality rate reported in 13 studies from other countries during the past 20 years (Fig. 4).6–18 The reported in-hospital trauma mortality rate was 5.1%–7.9% in the United States, Europe, and Malawi,6–8,12,13,16–18 9% in Japan,11 and 10.8%–12% in Canada and Malaysia.9,10,14 The highest in-hospital mortality rate of 21.4% was reported from 4 university hospitals in India without an established trauma system.15
The mechanism of trauma, Injury Severity Score on arrival and age were independently related to death in the present study, while gender was not. The severity of injury was a strong predictor of death and an increase of 10 degrees in the Injury Severity Score was associated with a threefold increase in the odds of death, while 10 years of age was associated with an increase in the odds of only 1.3 times. The present patients from South Africa were a mean of 30 years old, which was close to the age of patients from India15 but much lower than the mean age of 41–64 years in patients from high-income countries.7,9,11–13,17 The present mean Injury Severity Score of 10.4 was close to that in the largest study comprising 898,982 patients from 700 trauma centers in the United States, also reporting a low in-hospital mortality rate of 5.1%.7 A trauma center in Houston, Texas, reported a low mortality rate of 6.1%, despite a high mean Injury Severity Score of 26 on arrival.12 The highest in-hospital mortality of 24.1% was reported from 4 university hospitals in India, despite a low mean age of 31 years and a mean Injury Severity Score of 11.9 on arrival.15 It was suggested that this high mortality rate was a result of not using any established trauma system. Eighty-three percent of trauma victims in South Africa were men. Men also accounted for 65%–90% of the victims in Houston, Texas, Canada, London, United Kingdom, Japan, Malaysia, and India.8–12,14,15 An equal prevalence of men and women on trauma admission was only reported from France.17
Many subjects die at the scene in South Africa. A study from Pietermaritzburg recorded 1105 trauma victims during 2 years in 2010–2012 and 53% had died at the scene.24 In the present study, we found that 50% of patients died from traumatic brain injury. They died on the ward or were left at the emergency department to die because they could not be prioritized for the intensive care unit. There is a need for improved prehospital care and improved care after traumatic brain injury. Modern trauma centers and Advanced Trauma Life Support save many lives, especially in people with hemorrhagic shock and respiratory failure.6 Rescuing patients after a major head injury is, however, more complicated and resource consuming. These patients also run a high risk of a residual handicap, if they survive. While deaths due to hemorrhagic shock and respiratory failure are decreasing, traumatic brain injury now accounts for 72%–78% of in-hospital deaths at trauma centers and hospitals in France, The Netherlands, and Canada.10,16,17
The low in-hospital mortality rate can be explained by experienced staff who are accustomed to managing trauma, using the systematic trauma care approach of the Advanced Trauma Life Support program, including damage control surgery. The unavoidable deaths in the present study were early and mainly due to traumatic brain injury, while the avoidable deaths took place at a late stage and were mainly due to respiratory failure and multiple organ failure. The intensive care unit with only 10 beds at the hospital is heavily under resourced in terms of the burden of traumatized patients it is expected to manage. This is one explanation for avoidable deaths due to multiple organ failure and respiratory failure in the wards while patients are awaiting admission to intensive care unit. Looking after postoperative patients in a general ward is problematic and such patients are therefore not as closely monitored as needed. This means that it is difficult to recognize deteriorations in vital signs and physiological parameters. Other lack of resources included dialysis machines, ventilators, antibiotics, and blood for transfusion. It also included a need for more anesthesiologists and doctors and nurses specialized in traumatic brain injury, including a neurosurgeon to perform craniotomy.
Deaths from interpersonal violence were reduced in South Africa from 1997 to 2012, while deaths from road injuries were not.3 This study reports that the in-hospital trauma mortality rate is low at a South African trauma center and most deaths are early, unavoidable and due to traumatic brain injury. It is the excessively high level of gunshot wounds, stab wounds, assaults, and traffic accidents that explains the high trauma mortality and morbidity rate in South Africa compared with the rest of the world.25
One limitation is that the study has collected data from a single center, which limits its generalizability. Outcomes included in-hospital mortality and lacked important data on prehospital deaths and postdischarge outcome. Another limitation is the fact that the Delphi technique used to classify avoidable and unavoidable deaths is subjective and the routine care at the hospital may affect the classification.
In conclusion, the in-hospital trauma mortality rate at a South African trauma center using systematic trauma care is lower than that reported from other trauma centers in the world during the past 20 years. Nevertheless, 16% of death cases were assessed as avoidable if there had been better access to intensive care, dialysis, advanced respiratory care, blood for transfusion, and improvements in surgery and medical care.
E.S., V.K., G.L.L., M.S., and D.L.C. designed the study. M.S., V.C., G.L.L., J.L.B., W.B., and V.M. collected the data. E.S., M.S., J.S., and K.A.F. performed the data analysis. K.A.F., E.S., and D.L.C. drafted the article. V.K., G.L.L., M.S., J.L.B., W.B., J.S., and V.M. revised the article critically for important intellectual content. All the authors reviewed and approved the final article.
1. GBD 2017 Causes of Death Collaborators. Global, regional, and national age-sex-specific mortality for 282 causes of death in 195 countries and territories, 1980-2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet. 2018;392:1736–1788.
2. World Health Organization. Injuries and Violence: The Facts 2014. 2014. Available at: https://apps.who.int/iris/handle/10665/149798
. Accessed August 1, 2021.
3. Pillay-van Wyk V, Msemburi W, Laubscher R, et al. Mortality trends and differentials in South Africa from 1997 to 2012: second National Burden of Disease Study. Lancet Glob Health. 2016;4:e642–e653.
4. Matzopoulos R, Prinsloo M, Pillay-van Wyk V, et al. Injury-related mortality in South Africa: a retrospective descriptive study of postmortem investigations. Bull World Health Organ. 2015;93:303–313.
5. Pfeifer R, Tarkin IS, Rocos B, et al. Patterns of mortality and causes of death in polytrauma patients–has anything changed? Injury. 2009;40:907–911.
6. MacKenzie EJ, Rivara FP, Jurkovich GJ, et al. A national evaluation of the effect of trauma
-center care on mortality. N Engl J Med. 2006;354:366–378.
7. Valdez C, Sarani B, Young H, et al. Timing of death after traumatic injury–a contemporary assessment of the temporal distribution of death. J Surg Res. 2016;200:604–609.
8. Chalkley D, Cheung G, Walsh M, et al. Deaths from trauma
in London–a single centre experience. Emerg Med J. 2011;28:305–309.
9. Gomez D, Alali AS, Haas B, et al. Temporal trends and differences in mortality at trauma
centres across Ontario from 2005 to 2011: a retrospective cohort study. CMAJ Open. 2014;2:E176-E182.
10. Moore L, Evans D, Hameed SM, et al. Mortality in Canadian trauma
systems: a multicenter cohort study. Ann Surg. 2017;265:212–217.
11. Nagata I, Abe T, Uchida M, et al. Ten-year inhospital mortality trends for patients with trauma
in Japan: a multicentre observational study. BMJ Open. 2018;8:e018635.
12. Oyeniyi BT, Fox EE, Scerbo M, et al. Trends in 1029 trauma
deaths at a level 1 trauma
center: impact of a bleeding control bundle of care. Injury. 2017;48:5–12.
13. Girard E, Jegousso Q, Boussat B, et al.; TRENAU group. Preventable deaths in a French regional trauma
system: a six-year analysis of severe trauma
mortality. J Visc Surg. 2019;156:10–16.
14. Tan JH, Tan HCL, Noh NAM, et al. Validation of the trauma
mortality prediction scores from a Malaysian population. Burns Trauma
15. Roy N, Gerdin M, Ghosh S, et al. 30-day in-hospital trauma
mortality in four urban university hospitals using an Indian trauma
registry. World J Surg. 2016;40:1299–1307.
16. Jochems D, Leenen LPH, Hietbrink F, et al. Increased reduction in exsanguination rates leaves brain injury as the only major cause of death in blunt trauma
. Injury. 2018;49:1661–1667.
17. Bège T, Pauly V, Orleans V, et al. Epidemiology of trauma
in France: mortality and risk factors based on a national medico-administrative database. Anaesth Crit Care Pain Med. 2019;38:461–468.
18. Purcell LN, Mulima G, Reiss R, et al. Epidemiological comparisons and risk factors for pre-hospital and in-hospital mortality following traumatic injury in Malawi. World J Surg. 2020;44:2116–2122.
19. Rotondo M. About Advanced Trauma
Life Support. The American College of Surgeons. 2020. Available at: https://www.facs.org/quality-programs/trauma/atls/about
. Accessed August 1, 2021.
20. Dalkey NC, Helmer O. An experimental application of the Delphi method to the use of experts. Manage Sci. 1963;9:458–467.
21. Laing GL, Bruce JL, Skinner DL, et al. Development, implementation, and evaluation of a hybrid electronic medical record system specifically designed for a developing world surgical service. World J Surg. 2014;38:1388–1397.
22. Association for the Advancement of Automotive Medicine. The Abbreviated Injury Scale (AIS) 2005 Revision (Update 2008). Association for the Advancement of Automotive Medicine. 2008.
23. Baker SP, O’Neill B, Haddon W Jr, et al. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma
24. Moodley NB, Aldous C, Clarke DL. An audit of trauma
-related mortality in a provincial capital in South Africa. S Afr J Surg. 2014;52:101–104.
25. Haagsma JA, James SL, Castle CD, et al. Burden of injury along the development spectrum: associations between the socio-demographic index and disability-adjusted life year estimates from the Global Burden of Disease Study 2017. Inj Prev. 2020;26(supp 1):i12–i26.