The utility of the Kampala trauma score as a triage tool in a sub-Saharan African trauma cohort

Prevalence and levels of disability post road traffic orthopaedic injuries in Rwanda

Background

Prolonged disability resulting from road traffic injuries (RTIs) contributes significantly to morbidity and disease burden. A good understanding of the prevalence and the level of disability of orthopaedic injuries in developing countries is crucial for improvement; however, such data are currently lacking in Rwanda.

Objectives

To determine the prevalence and levels of disability of 2 years post-road traffic orthopaedic injuries in Rwanda.

Method

A multicentre, cross-sectional study from five Rwandan referral hospitals of 368 adult RTI victims' sustained from accidents in 2019. Between 02 June 2022, and 31 August 2022, two years after the injury, participants completed the World Health Organization Disability Assessment Schedule (WHODAS 2.0) Questionnaire for the degree of impairment and the Upper Extremity Functional Scale and Lower-Extremity Functional Scale forms for limb functional evaluation. Descriptive, inferential statistics Chi-square and multinomial regression models were analysed using R Studio.

Results

The study's mean age of the RTOI victims was 37.5 (±11.26) years, with a sex ratio M: F:3: 1. The prevalence of disability following road traffic orthopedic injury (RTOI) after 2 years was 36.14%, with victims having WHODAS score > 25.0% and 36.31% were still unable to return to their usual activities. Age group, Severe Kampala Trauma Score and lack of rehabilitation contributed to disability. The most affected WHODAS domains were participation in society (33%) and life activities (28%).

Conclusion

The prevalence and levels of disability because of RTOI in Rwanda are high, with mobility and participation in life being more affected than other WHODAS domains. Middle-aged and socio-economically underprivileged persons are the most affected.

Contribution

This study showed that a good rehabilitation approach and economic support for the RTI victims would decrease their disabilities in Rwanda.

The Bidirectionality of Global Surgical Research: The Utility of the Malawi Trauma Score in the United States Trauma Population

INTRODUCTION

Trauma scoring systems provide valuable risk stratification of injured patients. Trauma scoring systems developed in resource-limited settings, such as the Malawi Trauma Score (MTS), are based on readily available clinical information. This study sought to test the performance of the MTS in a United States trauma population.

MATERIALS AND METHODS

We analyzed the United States National Trauma Data Bank during 2017-2020. MTS uses alertness score: alert, responds to verbal or painful stimuli, or unresponsive (AVPU), age, sex, presence of a radial pulse, and primary anatomic injury location. MTS and an age-adjusted version reflective of the US age distribution, was evaluated for its performance in predicting crude mortality in the National Trauma Data Bank using receiver operating characteristic analysis. We utilized logistic regression to model the odds ratio of death at a particular MTS cutoff.

RESULTS

A total of 3,833,929 patients were included. The mean age was 49.3 y (sandard deviation 24.4), with a male preponderance (61.1%). Crude mortality was 3.4% (n = 131,452/3,833,929). The area under the curve for the MTS in predicting mortality was 0.87 (95% CI 0.87, 0.88). The area under the curve for a cutoff of 15 was 0.83 (95% CI 0.83, 0.83). An MTS of 15 higher had an odds ratio of death of 46.5 (95% CI 45.9, 47.1), compared to those with a score of 14 or lower.

CONCLUSIONS

MTS has excellent performance as a predictor of mortality in a US trauma population. MTS is simple to calculate and can be estimated in the prehospital setting or the emergency department. Consequently, it may have utility as a triage tool in both high-income trauma systems and resource-limited settings.

Use of a trauma registry to drive improvement in the regional trauma network systems in Hawassa, Ethiopia

AIM

Our aim is to establish and analyse the first year of trauma registry data from Hawassa University Comprehensive Specialised Hospital (HUCSH)-an Ethiopian tertiary referral centre. We plan to identify possible trends in injury patterns, access to health care and referral pathways and establish if our observations are in keeping with data published from other sub-Saharan LMIC's.

METHODS

Prospective data collection using the WHO trauma registry dataset. All trauma patients presenting to HUCSH between November 2019 and November 2020 were included. Military patients were excluded.

DATASET

Age, sex, region of residence, mode of transport to hospital, referral centre, time from injury to arrival in HUCSH, arrival triage category, Kampala Trauma Score (KTS), mechanism of injury, injury type, closed/open fracture and 24 h outcomes. Data statistical analysis was performed to calculate frequencies of the above variables.

RESULTS

There were a total of 1919 cases. Fifty-three per cent were caused by road traffic collision and 49% were fracture/dislocations. Public transport was the most common mode to hospital-40%. Seventy-seven per cent of all trauma admissions were referred from other centres, 69% had a mild KTS. A total of 376 patients presented with an open fracture-76% had a low KTS and 67% remained in ED for > 24 h. Sixty-five per cent of ambulances were utilised for mild KTS patients, only 25% of ambulances transported moderate and severe injuries.

CONCLUSION

We have shown that a 'one size fits all approach' should not be adopted for LMIC's as trends vary between regions and countries. Improvements are needed in ambulance utilisation, the use of appropriate triaging tools to facilitate initial basic trauma care and appropriate, timely referrals and the management of open fractures.

Malawi Trauma Score is Predictive of Mortality at a District Hospital: A Validation Study

BACKGROUND

Trauma scoring systems can identify patients who should be transferred to referral hospitals, but their utility in LMICs is often limited. The Malawi Trauma Score (MTS) reliably predicts mortality at referral hospitals but has not been studied at district hospitals. We sought to validate the MTS at a Malawi district hospital and evaluate whether MTS is predictive of transfer to a referral hospital.

METHODS

We performed a retrospective study using trauma registry data from Salima District Hospital (SDH) from 2017 to 2021. We excluded patients brought in dead, discharged from the Casualty Department, or missing data needed to calculate MTS. We used logistic regression modeling to study the relationship between MTS and mortality at SDH and between MTS and transfer to a referral hospital. We used receiver operating characteristic analysis to validate the MTS as a predictor of mortality.

RESULTS

We included 2196 patients (84.3% discharged, 12.7% transferred, 3.0% died). These groups had similar ages, sex, and admission vitals. Mean (SD) MTS was 7.9(3.0) among discharged patients, 8.4(3.9) among transferred patients, and 14.2(8.0) among patients who died (p < 0.001). Higher MTS was associated with increased odds of mortality at SDH (OR 1.21, 95% CI 1.14-1.29, p < 0.001) but was not related to transfer. ROC area for mortality was 0.73 (95% CI 0.65-0.80).

CONCLUSIONS

MTS is predictive of district hospital mortality but not inter-facility transfer. We suggest that MTS be used to identify patients with severe trauma who are most likely to benefit from transfer to a referral hospital.

The Kampala Trauma Score: A 20-year track record

ABSTRACT

Quantifying the severity of traumatic injury has been foundational for the standardization of outcomes, quality improvement research, and health policy throughout the evolution of trauma care systems. Many injury severity scores are difficult to calculate and implement, especially in low- and middle-income countries (LMICs) where human resources are limited. The Kampala Trauma Score (KTS)-a simplification of the Trauma Injury Severity Score-was developed in 2000 to accommodate these settings. Since its development, numerous instances of KTS use have been documented, but extent of adoption is unknown. More importantly, does the KTS remain useful for determining injury severity in LMICs? This review aims to better understand the legacy of the KTS and assess its strengths and weaknesses. Three databases were searched to identify scientific papers concerning the KTS. Google Scholar was searched to identify grey literature. The search returned 357 papers, of which 199 met inclusion criteria. Eighty-five studies spanning 16 countries used the KTS in clinical settings. Thirty-seven studies validated the KTS, assessing its ability to predict outcomes such as mortality or need for admission. Over 80% of these studies reported the KTS equalled or exceeded more complicated scores at predicting mortality. The KTS has stood the test of time, proving itself over the last twenty years as an effective measure of injury severity across numerous contexts. We recommend the KTS as a means of strengthening trauma systems in LMICs and suggest it could benefit high-income trauma systems that do not measure injury severity.

Evaluation of the Revised Trauma Score, MGAP, and GAP scoring systems in predicting mortality of adult trauma patients in a low-resource setting

Background

Numerous trauma scoring systems have been developed in an attempt to accurately and efficiently predict the prognosis of emergent trauma cases. However, it has been questioned as to whether the accuracy and pragmatism of such systems still hold in lower-resource settings that exist in many hospitals in lower- and middle-income countries (LMICs). In this study, it was hypothesized that the physiologically-based Revised Trauma Score (RTS), Mechanism/Glasgow Coma Scale/Age/Pressure (MGAP) score, and Glasgow Coma Scale/Age/Pressure (GAP) score would be effective at predicting mortality outcomes using clinical data at presentation in a representative LMIC hospital in Upper Egypt.

Methods

This was a retrospective analysis of the medical records of trauma patients at Beni-Suef University Hospital. Medical records of all trauma patients admitted to the hospital over the 8-month period from January to August 2016 were reviewed. For each case, the RTS, MGAP, and GAP scores were calculated using clinical data at presentation, and mortality prediction was correlated to the actual in-hospital outcome.

Results

The Area Under the Receiver Operating Characteristic (AUROC) was calculated to be 0.879, 0.890, and 0.881 for the MGAP, GAP, and RTS respectively, with all three scores showing good discriminatory ability. With regards to prevalence-dependent statistics, all three scores demonstrated efficacy in ruling out mortality upon presentation with negative predictive values > 95%, while the MGAP score best captured the mortality subgroup with a sensitivity of 94%. Adjustment of cutoff scores showed a steep trade-off between optimizing the positive predictive values versus the sensitivities.

Conclusion

The RTS, MGAP, and GAP all showed good discriminatory capabilities per AUROC. Given the relative simplicity and potentially added clinical benefit in capturing critically ill patients, the MGAP score should be further studied for stratifying risk of incoming trauma patients to the emergency department, allowing for more efficacious triage of patients in lower-resource healthcare settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12873-022-00653-1.

Predicting surgical resource consumption and in-hospital mortality in resource-scarce conflict settings: a retrospective study

Background

In armed conflicts, civilian health care struggles to cope. Being able to predict what resources are needed is therefore vital. The International Committee of the Red Cross (ICRC) implemented in the 1990s the Red Cross Wound Score (RCWS) for assessment of penetrating injuries. It is unknown to what extent RCWS or the established trauma scores Kampala trauma Score (KTS) and revised trauma score (RTS) can be used to predict surgical resource consumption and in-hospital mortality in resource-scarce conflict settings.

Methods

A retrospective study of routinely collected data on weapon-injured adults admitted to ICRC’s hospitals in Peshawar, 2009–2012 and Goma, 2012–2014. High resource consumption was defined as ≥3 surgical procedures or ≥ 3 blood-transfusions or amputation. The relationship between RCWS, KTS, RTS and resource consumption, in-hospital mortality was evaluated with logistic regression and adjusted area under receiver operating characteristic curves (AUC). The impact of missing data was assessed with imputation. Model fit was compared with Akaike Information Criterion (AIC).

Results

A total of 1564 patients were included, of these 834 patients had complete data. For high surgical resource consumption AUC was significantly higher for RCWS (0.76, 95% CI 0.74–0.78) than for KTS (0.53, 95% CI 0.50–0.56) and RTS (0.51, 95% CI 0.48–0.54) for all patients. Additionally, RCWS had lower AIC, indicating a better model fit. For in-hospital mortality AUC was significantly higher for RCWS (0.83, 95% CI 0.79–0.88) than for KTS (0.71, 95% CI 0.65–0.76) and RTS (0.70, 95% CI 0.63–0.76) for all patients, but not for patients with complete data.

Conclusion

RCWS appears to predict surgical resource consumption better than KTS and RTS. RCWS may be a promising tool for planning and monitoring surgical care in resource-scarce conflict settings.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12873-021-00488-2.

The characteristics and outcomes of trauma admissions to an adult general surgery ward in a tertiary teaching hospital

Background

Traumatic injuries are proportionally higher in low- and middle-income countries (LMICs) than high-income counties. Data on trauma epidemiology and patients' outcomes are limited in LMICs.

Methods

A retrospective review of medical records was performed for trauma admissions to the Princess Marina Hospital general surgical (GS) wards from August 2017 to July 2018. Data on demographics, mechanisms of injury, body parts injured, Revised Trauma Score, surgical procedures, hospital stay, and outcomes were analysed.

Results

During the study period, 2610 patients were admitted to GS wards, 1307 were emergency admissions. Trauma contributed 22.1% (576) of the total and 44.1% of the emergency admissions. Among the trauma admissions, 79.3% (457) were male. The median[interquartile range(IQR)](range) age in years was 30[24–40](13–97). The main mechanisms of injury were interpersonal violence (IPV), 53.1% and road traffic crashes (RTCs), 23.1%. More females than males suffered animal bites (5.9% vs. 0.9%), and burns (8.4% vs. 4.2%), while more males than females were affected by IPV (57.8% vs. 35.3%) and self-harm (5.5% vs. 3.4%). Multiple body parts were injured in 6.6%, mainly by RTCs. Interpersonal violence (IPV) and RTCs resulted in significant numbers of head and neck injuries, 57.3% and 22.2% respectively. More females than males had multiple body-parts injury 34.5% vs. 18.5%. Revised Trauma Score (RTS) of ≤11 was recorded in IPV, 38.4% and RTCs, 33.6%. Surgical procedures were performed on 44.4% patients. The most common surgical procedures were laparotomy (27.8%), insertion of chest tube (27.8%), and craniotomy/burr hole(25.1%). Complications were recorded in 10.1% of the patients(58) including 39 deaths, 6.8% of the 576.

Conclusion

Trauma contributed significantly to the total GS and emergency admissions. The most common mechanism of injury was IPV with head and neck the most frequently injured body part. Further studies on IPV and trauma admissions involving paediatric and orthopaedic patients are warranted.

Comparison of the Pediatric Resuscitation and Trauma Outcome (PRESTO) Model and Pediatric Trauma Scoring Systems in a Middle-Income Country

BACKGROUND

The pediatric resuscitation and trauma outcome (PRESTO) model was developed to aid comparisons of risk-adjusted mortality after injury in low- and middle-income countries (LMICs). We sought to validate PRESTO using data from a middle-income country (MIC) trauma registry and compare its performance to the Pediatric Trauma Score (PTS), Revised Trauma Score, and pediatric age-adjusted shock index (SIPA).

METHODS

We included children (age < 15 years) admitted to a single trauma center in South Africa from December 2012 to January 2019. We excluded patients missing variables necessary for the PRESTO model-age, systolic blood pressure, pulse, oxygen saturation, neurologic status, and airway support. Trauma scores were assigned retrospectively. PRESTO's previously high-income country (HIC)-validated optimal threshold was compared to MIC-validated threshold using area under the receiver operating characteristic curves (AUROC). Prediction of in-hospital death using trauma scoring systems was compared using ROC analysis.

RESULTS

Of 1160 injured children, 988 (85%) had complete data for calculation of PRESTO. Median age was 7 (IQR: 4, 11), and 67% were male. Mortality was 2% (n = 23). Mean predicted mortality was 0.5% (range 0-25.7%, AUROC 0.93). Using the HIC-validated threshold, PRESTO had a sensitivity of 26.1% and a specificity of 99.7%. The MIC threshold showed a sensitivity of 82.6% and specificity of 89.4%. The MIC threshold yielded superior discrimination (AUROC 0.86 [CI 0.78, 0.94]) compared to the previously established HIC threshold (0.63 [CI 0.54, 0.72], p < 0.0001). PRESTO showed superior prediction of in-hospital death compared to PTS and SIPA (all p < 0.01).

CONCLUSION

PRESTO can be applied in MIC settings and discriminates between children at risk for in-hospital death following trauma. Further research should clarify optimal decision thresholds for quality improvement and benchmarking in LMIC settings.

Performance of Prognostication Scores for Mortality in Injured Patients in Rwanda

INTRODUCTION

While trauma prognostication and triage scores have been designed for use in lower-resourced healthcare settings specifically, the comparative clinical performance between trauma-specific and general triage scores for risk-stratifying injured patients in such settings is not well understood. This study evaluated the Kampala Trauma Score (KTS), Revised Trauma Score (RTS), and Triage Early Warning Score (TEWS) for accuracy in predicting mortality among injured patients seeking emergency department (ED) care at the Centre Hospitalier Universitaire de Kigali (CHUK) in Rwanda.

METHODS

A retrospective, randomly sampled cohort of ED patients presenting with injury was accrued from August 2015-July 2016. Primary outcome was 14-day mortality and secondary outcome was overall facility-based mortality. We evaluated summary statistics of the cohort. Bootstrap regression models were used to compare areas under receiver operating curves (AUC) with associated 95% confidence intervals (CI).

RESULTS

Among 617 cases, the median age was 32 years and 73.5% were male. The most frequent mechanism of injury was road traffic incident (56.2%). Predominant anatomical regions of injury were craniofacial (39.3%) and lower extremities (38.7%), and the most common injury types were fracture (46.0%) and contusion (12.0%). Fourteen-day mortality was 2.6% and overall facility-based mortality was 3.4%. For 14-day mortality, TEWS had the highest accuracy (AUC = 0.88, 95% CI, 0.76-1.00), followed by RTS (AUC = 0.73, 95% CI, 0.55-0.92), and then KTS (AUC = 0.65, 95% CI, 0.47-0.84). Similarly, for facility-based mortality, TEWS (AUC = 0.89, 95% CI, 0.79-0.98) had greater accuracy than RTS (AUC = 0.76, 95% CI, 0.61-0.91) and KTS (AUC = 0.68, 95% CI, 0.53-0.83). On pairwise comparisons, RTS had greater prognostic accuracy than KTS for 14-day mortality (P = 0.011) and TEWS had greater accuracy than KTS for overall (P = 0.007) mortality. However, TEWS and RTS accuracy were not significantly different for 14-day mortality (P = 0.864) or facility-based mortality (P = 0.101).

CONCLUSION

In this cohort of emergently injured patients in Rwanda, the TEWS demonstrated the greatest accuracy for predicting mortality outcomes, with no significant discriminatory benefit found in the use of the trauma-specific RTS or KTS instruments, suggesting that the TEWS is the most clinically useful approach in the setting studied and likely in other similar ED environments.

Defining High-risk Emergency Chief Complaints: Data-driven Triage for Low- and Middle-income Countries

OBJECTIVES

Emergency medicine in low- and middle-income countries (LMICs) is hindered by lack of research into patient outcomes. Chief complaints (CCs) are fundamental to emergency care but have only recently been uniquely codified for an LMIC setting in Uganda. It is not known whether CCs independently predict emergency unit patient outcomes.

METHODS

Patient data collected in a Ugandan emergency unit between 2009 and 2018 were randomized into validation and derivation data sets. A recursive partitioning algorithm stratified CCs by 3-day mortality risk in each group. The process was repeated in 10,000 bootstrap samples to create an averaged risk ranking. Based on this ranking, CCs were categorized as "high-risk" (>2× baseline mortality), "medium-risk" (between 2 and 0.5× baseline mortality), and "low-risk" (<0.5× baseline mortality). Risk categories were then included in a logistic regression model to determine if CCs independently predicted 3-day mortality.

RESULTS

Overall, the derivation data set included 21,953 individuals with 7,313 in the validation data set. In total, 43 complaints were categorized, and 12 CCs were identified as high-risk. When controlled for triage data including age, sex, HIV status, vital signs, level of consciousness, and number of complaints, high-risk CCs significantly increased 3-day mortality odds ratio (OR = 2.39, 95% confidence interval [CI] = 1.95 to 2.93, p < 0.001) while low-risk CCs significantly decreased 3-day mortality odds (OR = 0.16, 95% CI = 0.09 to 0.29, p < 0.001).

CONCLUSIONS

High-risk CCs were identified and found to predict increased 3-day mortality independent of vital signs and other data available at triage. This list can be used to expand local triage systems and inform emergency training programs. The methodology can be reproduced in other LMIC settings to reflect their local disease patterns.

Validation of the PRESTO score in injured children in a South-African quaternary trauma center

INTRODUCTION

The Pediatric RESuscitation and Trauma Outcome (PRESTO) model was developed for standardized risk-adjustment in pediatric trauma and is adapted to low-resource settings. It includes easily-accessible demographic and physiologic variables that are available at point of care in virtually any setting. The purpose of this study was to evaluate the PRESTO model's ability to predict in-hospital mortality in a South African pediatric trauma unit by comparing it to the widely used Injury Severity Score (ISS).

METHODS

Data prospectively collected between 2007 and 2017 in the Inkosi Albert Luthuli Central Hospital Trauma Registry were retrospectively reviewed. Injured children younger than 14 years were included if they were admitted to hospital or died as a result of their injury. We excluded patients with minor injuries who were treated and discharged home and patients with incomplete hospital disposition data. Receiver-Operating Characteristic (ROC) curves were constructed for PRESTO and ISS, and the areas under the curve (AUCs) were compared using Delong's test. The sensitivity and specificity of PRESTO were calculated at different prognostic threshold values identified through literature review.

RESULTS

We identified 419 patients; 67 died in hospital (16%). The AUCs for PRESTO and ISS were 0.82 (95% confidence interval CI [0.76-0.87]) and 0.75 (CI [0.68-0.81]), respectively. This difference trended towards statistical significance (p = 0.07). Using the optimal threshold of 0.13 described in the original publication, PRESTO had a 97% sensitivity and 37% specificity, while a threshold of 0.50 yielded 90% sensitivity and 54% specificity. The mean predicted probability of in-hospital death among patients who died was 0.79. Using this value as a threshold yielded the 57% sensitivity and 85% specificity.

CONCLUSION

This analysis has demonstrated the validity of the PRESTO model for in-hospital mortality prediction for pediatric trauma patients in the setting of a dedicated high-complexity trauma unit in a South African trauma referral center.

LEVEL OF EVIDENCE

Level III: Case-control.

Thefeasibility, appropriateness, and applicability of trauma scoring systems in low and middle-income countries: a systematic review

Background

About 5.8 million people die each year as a result of injuries, and nearly 90% of these deaths occur in low and middle-income countries (LMIC). Trauma scoring is a cornerstone of trauma quality improvement (QI) efforts, and is key to organizing and evaluating trauma services. The objective of this review was to assess the appropriateness, feasibility, and QI applicability of traditional trauma scoring systems in LMIC settings.

Materials and methods

This systematic review searched PubMed, Scopus, CINAHL, and trauma-focused journals for articles describing the use of a standardized trauma scoring system to characterize holistic health status. Studies conducted in high-income countries (HIC) or describing scores for isolated anatomic locations were excluded. Data reporting a score’s capacity to discriminate mortality, feasibility of implementation, or use for QI were extracted and synthesized.

Results

Of the 896 articles screened, 336 were included. Over half of studies (56%) reported Glasgow Coma Scale, followed by Injury Severity Score (ISS; 51%), Abbreviated Injury Scale (AIS; 24%), Revised Trauma Score (RTS; 19%), Trauma and Injury Severity Score (TRISS; 14%), and Kampala Trauma Score (7%). While ISS was overwhelmingly predictive of mortality, 12 articles reported limited feasibility of ISS and/or AIS. RTS consistently underestimated injury severity. Over a third of articles (37%) reporting TRISS assessmentsobserved mortality that was greater than that predicted by TRISS. Several articles cited limited human resources as the key challenge to feasibility.

Conclusions

The findings of this review reveal that implementing systems designed for HICs may not be relevant to the burden and resources available in LMICs. Adaptations or alternative scoring systems may be more effective.

PROSPERO registration number

CRD42017064600.

Epidemiology of paediatric injuries in Rwanda using a prospective trauma registry

BACKGROUND

Child survival initiatives historically prioritized efforts to reduce child morbidity and mortality from infectious diseases and maternal conditions. Little attention has been devoted to paediatric injuries in resource-limited settings. This study aimed to evaluate the demographics and outcomes of paediatric injury in a sub-Saharan African country in an effort to improve prevention and treatment.

METHODS

A prospective trauma registry was established at the two university teaching campuses of the University of Rwanda to record systematically patient demographics, prehospital care, initial physiology and patient outcomes from May 2011 to July 2015. Univariable analysis was performed for demographic characteristics, injury mechanisms, geographical location and outcomes. Multivariable analysis was performed for mortality estimates.

RESULTS

Of 11 036 patients in the registry, 3010 (27·3 per cent) were under 18 years of age. Paediatric patients were predominantly boys (69·9 per cent) and the median age was 8 years. The mortality rate was 4·8 per cent. Falls were the most common injury (45·3 per cent), followed by road traffic accidents (30·9 per cent), burns (10·7 per cent) and blunt force/assault (7·5 per cent). Patients treated in the capital city, Kigali, had a higher incidence of head injury (7·6 per cent versus 2·0 per cent in a rural town, P < 0·001; odds ratio (OR) 4·08, 95 per cent c.i. 2·61 to 6·38) and a higher overall injury-related mortality rate (adjusted OR 3·00, 1·50 to 6·01; P = 0·019). Pedestrians had higher overall injury-related mortality compared with other road users (adjusted OR 3·26, 1·37 to 7·73; P = 0·007).

CONCLUSION

Paediatric injury is a significant contributor to morbidity and mortality. Delineating trauma demographics is important when planning resource utilization and capacity-building efforts to address paediatric injury in low-resource settings and identify vulnerable populations.

The Malawi trauma score: A model for predicting trauma-associated mortality in a resource-poor setting

BACKGROUND

Globally, traumatic injury is a leading cause of morbidity and mortality in low-income countries. Current tools for predicting trauma-associated mortality are often not applicable in low-resource environments due to a lack of diagnostic adjuncts. This study sought to derive and validate a model for predicting mortality that requires only a history and physical exam.

METHODS

We conducted a retrospective analysis of all patients recorded in the Kamuzu Central Hospital trauma surveillance registry in Lilongwe, Malawi from 2011 through 2014. Using statistical randomization, 80% of patients were used for derivation and 20% were used for validation. Logistic regression modeling was used to derive factors associated with mortality and the Malawi Trauma Score (MTS) was constructed. The model fitness was tested.

RESULTS

62,354 patients are included. Patients are young (mean age 23.0, SD 15.9 years) with a male preponderance (72%). Overall mortality is 1.8%. The MTS is tabulated based on initial mental status (alert, responds to voice, responds only to pain or worse), anatomical location of the most severe injury, the presence or absence of a radial pulse on examination, age, and sex. The score range is 2-32. A mental status exam of only responding to pain or worse, head injury, the absence of a radial pulse, extremes of age, and male sex all conferred a higher probability of mortality. The ROC area under the curve for the derivation cohort and validation cohort were 0.83 (95% CI 0.78, 0.87) and 0.83 (95% CI 0.75, 0.92), respectively. A MTS of 25 confers a 50% probability of death.

CONCLUSIONS

The MTS provides a reliable tool for trauma triage in sub-Saharan Africa and helps risk stratify patient populations. Unlike other models previously developed, its strength is its utility in virtually any environment, while reliably predicting injury- associated mortality.

Machine learning without borders? An adaptable tool to optimize mortality prediction in diverse clinical settings

BACKGROUND

Mortality prediction aids clinical decision making and is necessary for quality improvement initiatives. Validated metrics rely on prespecified variables and often require advanced diagnostics, which are unfeasible in resource-constrained contexts. We hypothesize that machine learning will generate superior mortality prediction in both high-income and low- and middle-income country cohorts.

METHODS

SuperLearner, an ensemble machine-learning algorithm, was applied to data from three prospective trauma cohorts: a highest-activation cohort in the United States, a high-volume center cohort in South Africa (SA), and a multicenter registry in Cameroon. Cross-validation was used to assess model discrimination of discharge mortality by site using receiver operating characteristic curves. SuperLearner discrimination was compared with standard scoring methods. Clinical variables driving SuperLearner prediction at each site were evaluated.

RESULTS

Data from 28,212 injured patients were used to generate prediction. Discharge mortality was 17%, 1.3%, and 1.7% among US, SA, and Cameroonian cohorts. SuperLearner delivered superior prediction of discharge mortality in the United States (area under the curve [AUC], 94-97%) and vastly superior prediction in Cameroon (AUC, 90-94%) compared with conventional scoring algorithms. It provided similar prediction to standard scores in the SA cohort (AUC, 90-95%). Context-specific variables (partial thromboplastin time in the United States and hospital distance in Cameroon) were prime drivers of predicted mortality in their respective cohorts, whereas severe brain injury predicted mortality across sites.

CONCLUSIONS

Machine learning provides excellent discrimination of injury mortality in diverse settings. Unlike traditional scores, data-adaptive methods are well suited to optimizing precise site-specific prediction regardless of diagnostic capabilities or data set inclusion allowing for individualized decision making and expanded access to quality improvement programming.

LEVEL OF EVIDENCE

Prognostic and therapeutic, level II and III.

Trauma registry implementation and operation in low and middle income countries: A scoping review

Injury is a major public health crisis contributing to more than 4.48 million deaths annually. Trauma registries have proven highly effective in reducing injury morbidity and mortality rates in high income countries. They are a critical source of information for injury prevention, benchmarking care, quality improvement, and resource allocation. Historically, low and middle income countries (LMICs) have largely been excluded from trauma registry development due to limited resources. Recently, this has begun to change with low-resource hospitals adopting innovative strategies to implement trauma registries. Nonetheless, dissemination of these strategies remains fragmented. Hospitals looking to develop their own trauma registries have no current, comprehensive resource that summarises the implementation decisions of other registries in similar contexts. This scoping review aims to identify where trauma registries are located in LMICs, bringing up to date previous estimates, and to identify the most common approaches to registry implementation and operation in these settings.

Machine Learning Models of Survival Prediction in Trauma Patients

BACKGROUND

We aimed to build a model using machine learning for the prediction of survival in trauma patients and compared these model predictions to those predicted by the most commonly used algorithm, the Trauma and Injury Severity Score (TRISS).

METHODS

Enrolled hospitalized trauma patients from 2009 to 2016 were divided into a training dataset (70% of the original data set) for generation of a plausible model under supervised classification, and a test dataset (30% of the original data set) to test the performance of the model. The training and test datasets comprised 13,208 (12,871 survival and 337 mortality) and 5603 (5473 survival and 130 mortality) patients, respectively. With the provision of additional information such as pre-existing comorbidity status or laboratory data, logistic regression (LR), support vector machine (SVM), and neural network (NN) (with the Stuttgart Neural Network Simulator (RSNNS)) were used to build models of survival prediction and compared to the predictive performance of TRISS. Predictive performance was evaluated by accuracy, sensitivity, and specificity, as well as by area under the curve (AUC) measures of receiver operating characteristic curves.

RESULTS

In the validation dataset, NN and the TRISS presented the highest score (82.0%) for balanced accuracy, followed by SVM (75.2%) and LR (71.8%) models. In the test dataset, NN had the highest balanced accuracy (75.1%), followed by the TRISS (70.2%), SVM (70.6%), and LR (68.9%) models. All four models (LR, SVM, NN, and TRISS) exhibited a high accuracy of more than 97.5% and a sensitivity of more than 98.6%. However, NN exhibited the highest specificity (51.5%), followed by the TRISS (41.5%), SVM (40.8%), and LR (38.5%) models.

CONCLUSIONS

These four models (LR, SVM, NN, and TRISS) exhibited a similar high accuracy and sensitivity in predicting the survival of the trauma patients. In the test dataset, the NN model had the highest balanced accuracy and predictive specificity.

Dynamic multi-outcome prediction after injury: Applying adaptive machine learning for precision medicine in trauma

OBJECTIVE

Machine learning techniques have demonstrated superior discrimination compared to conventional statistical approaches in predicting trauma death. The objective of this study is to evaluate whether machine learning algorithms can be used to assess risk and dynamically identify patient-specific modifiable factors critical to patient trajectory for multiple key outcomes after severe injury.

METHODS

SuperLearner, an ensemble machine-learning algorithm, was applied to prospective observational cohort data from 1494 critically-injured patients. Over 1000 agnostic predictors were used to generate prediction models from multiple candidate learners for outcomes of interest at serial time points post-injury. Model accuracy was estimated using cross-validation and area under the curve was compared to select among predictors. Clinical variables responsible for driving outcomes were estimated at each time point.

RESULTS

SuperLearner fits demonstrated excellent cross-validated prediction of death (overall AUC 0.94-0.97), multi-organ failure (overall AUC 0.84-0.90), and transfusion (overall AUC 0.87-0.9) across multiple post-injury time points, and good prediction of Acute Respiratory Distress Syndrome (overall AUC 0.84-0.89) and venous thromboembolism (overall AUC 0.73-0.83). Outcomes with inferior data quality included coagulopathic trajectory (AUC 0.48-0.88). Key clinical predictors evolved over the post-injury timecourse and included both anticipated and unexpected variables. Non-random missingness of data was identified as a predictor of multiple outcomes over time.

CONCLUSIONS

Machine learning algorithms can be used to generate dynamic prediction after injury while avoiding the risk of over- and under-fitting inherent in ad hoc statistical approaches. SuperLearner prediction after injury demonstrates promise as an adaptable means of helping clinicians integrate voluminous, evolving data on severely-injured patients into real-time, dynamic decision-making support.

Potential benefits of triage for the trauma patient in a Kenyan emergency department

BACKGROUND

Improved trauma management can reduce the time between injury and medical interventions, thus decreasing morbidity and mortality. Triage at the emergency department is essential to ensure prioritization and timely assessment of injured patients. The aim of the present study was to investigate how a lack of formal triage system impacts timely intervention and mortality in a sub-Saharan referral hospital. Further, the study attempts to assess potential benefits of triage towards efficient management of trauma patients in one middle income country.

METHODS

A prospective descriptive study was conducted. Adult trauma patients admitted to the emergency department during an 8-month period at Moi Teaching and Referral Hospital in Eldoret, Kenya, were included. Mode of arrival and vital parameters were registered. Variables included in the analysis were Injury Severity Score, time before physician's assessment, length of hospital stay, and mortality. The patients were retrospectively categorized according to the Rapid Emergency Triage and Treatment System (RETTS) from patient records.

RESULTS

A total of 571 patients were analyzed, with a mean Injury Severity Score of 12.2 (SD 7.7) with a mean length of stay of 11.6 (SD 18.3) days. The mortality rate was 1.8%. The results obtained in this study illustrate that trauma patients admitted to the emergency department at Eldoret are not assessed in a timely fashion, and the time frame recommendations postulated by RETTS are not adhered to. Assessment of patients according to the triage algorithm used revealed a significantly higher average Injury Severity Score in the red category than in the other color categories.

CONCLUSION

The results from this study clearly illustrate a lack of correct prioritization of patients in relation to the need for timely assessment. This is further demonstrated by the retrospective triage classification of patients, which identified patients with high ISS as in urgent need of care. Since no significant difference in to time to assessment regardless of injury severity was observed, the need for a well-functioning triage system is apparent.

Exploring injury severity measures and in-hospital mortality: A multi-hospital study in Kenya

INTRODUCTION

Low- and middle-income countries (LMICs) have a disproportionately high burden of injuries. Most injury severity measures were developed in high-income settings and there have been limited studies on their application and validity in low-resource settings. In this study, we compared the performance of seven injury severity measures: estimated Injury Severity Score (eISS), Glasgow Coma Score (GCS), Mechanism, GCS, Age, Pressure score (MGAP), GCS, Age, Pressure score (GAP), Revised Trauma Score (RTS), Trauma and Injury Severity Score (TRISS) and Kampala Trauma Score (KTS), in predicting in-hospital mortality in a multi-hospital cohort of adult patients in Kenya.

METHODS

This study was performed using data from trauma registries implemented in four public hospitals in Kenya. Estimated ISS, MGAP, GAP, RTS, TRISS and KTS were computed according to algorithms described in the literature. All seven measures were compared for discrimination by computing area under curve (AUC) for the receiver operating characteristics (ROC), model fit information using Akaike information criterion (AIC), and model calibration curves. Sensitivity analysis was conducted to include all trauma patients during the study period who had missing information on any of the injury severity measure(s) through multiple imputations.

RESULTS

A total of 16,548 patients were included in the study. Complete data analysis included 14,762 (90.2%) patients for the seven injury severity measures. TRISS (complete case AUC: 0.889, 95% CI: 0.866-0.907) and KTS (complete case AUC: 0.873, 95% CI: 0.852-0.892) demonstrated similarly better discrimination measured by AUC on in-hospital deaths overall in both complete case analysis and multiple imputations. Estimated ISS had lower AUC (0.764, 95% CI: 0.736-0.787) than some injury severity measures. Calibration plots showed eISS and RTS had lower calibration than models from other injury severity measures.

CONCLUSIONS

This multi-hospital study in Kenya found statistical significant higher performance of KTS and TRISS than other injury severity measures. The KTS, is however, an easier score to compute as compared to the TRISS and has stable good performance across several hospital settings and robust to missing values. It is therefore a practical and robust option for use in low-resource settings, and is applicable to settings similar to Kenya.

Mortality prediction models in the general trauma population: A systematic review

BACKGROUND

Trauma is the leading cause of death in individuals younger than 40 years. There are many different models for predicting patient outcome following trauma. To our knowledge, no comprehensive review has been performed on prognostic models for the general trauma population. Therefore, this review aimed to describe (1) existing mortality prediction models for the general trauma population, (2) the methodological quality and (3) which variables are most relevant for the model prediction of mortality in the general trauma population.

METHODS

An online search was conducted in June 2015 using Embase, Medline, Web of Science, Cinahl, Cochrane, Google Scholar and PubMed. Relevant English peer-reviewed articles that developed, validated or updated mortality prediction models in a general trauma population were included.

RESULTS

A total of 90 articles were included. The cohort sizes ranged from 100 to 1,115,389 patients, with overall mortality rates that ranged from 0.6% to 35%. The Trauma and Injury Severity Score (TRISS) was the most commonly used model. A total of 258 models were described in the articles, of which only 103 models (40%) were externally validated. Cases with missing values were often excluded and discrimination of the different prediction models ranged widely (AUROC between 0.59 and 0.98). The predictors were often included as dichotomized or categorical variables, while continuous variables showed better performance.

CONCLUSION

Researchers are still searching for a better mortality prediction model in the general trauma population. Models should 1) be developed and/or validated using an adequate sample size with sufficient events per predictor variable, 2) use multiple imputation models to address missing values, 3) use the continuous variant of the predictor if available and 4) incorporate all different types of readily available predictors (i.e., physiological variables, anatomical variables, injury cause/mechanism, and demographic variables). Furthermore, while mortality rates are decreasing, it is important to develop models that predict physical, cognitive status, or quality of life to measure quality of care.

Diagnostic accuracy of the Kampala Trauma Score using estimated Abbreviated Injury Scale scores and physician opinion

BACKGROUND

The Kampala Trauma Score (KTS) has been proposed as a triage tool for use in low- and middle-income countries (LMICs). This study aimed to examine the diagnostic accuracy of KTS in predicting emergency department outcomes using timely injury estimation with Abbreviated Injury Scale (AIS) score and physician opinion to calculate KTS scores.

METHODS

This was a diagnostic accuracy study of KTS among injured patients presenting to Komfo Anokye Teaching Hospital A&E, Ghana. South African Triage Scale (SATS); KTS component variables, including AIS scores and physician opinion for serious injury quantification; and ED disposition were collected. Agreement between estimated AIS score and physician opinion were analyzed with normal, linear weighted, and maximum kappa. Receiver operating characteristic (ROC) analysis of KTS-AIS and KTS-physician opinion was performed to evaluate each measure's ability to predict A&E mortality and need for hospital admission to the ward or theatre.

RESULTS

A total of 1053 patients were sampled. There was moderate agreement between AIS criteria and physician opinion by normal (κ=0.41), weighted (κ_lin=0.47), and maximum (κ_max=0.53) kappa. A&E mortality ROC area for KTS-AIS was 0.93, KTS-physician opinion 0.89, and SATS 0.88 with overlapping 95% confidence intervals (95%CI). Hospital admission ROC area for KTS-AIS was 0.73, KTS-physician opinion 0.79, and SATS 0.71 with statistical similarity. When evaluating only patients with serious injuries, KTS-AIS (ROC 0.88) and KTS-physician opinion (ROC 0.88) performed similarly to SATS (ROC 0.78) in predicting A&E mortality. The ROC area for KTS-AIS (ROC 0.71; 95%CI 0.66-0.75) and KTS-physician opinion (ROC 0.74; 95%CI 0.69-0.79) was significantly greater than SATS (ROC 0.57; 0.53-0.60) with regard to need for admission.

CONCLUSIONS

KTS predicted mortality and need for admission from the ED well when early estimation of the number of serious injuries was used, regardless of method (i.e. AIS criteria or physician opinion). This study provides evidence for KTS to be used as a practical and valid triage tool to predict patient prognosis, ED outcomes and inform referral decision-making from first- or second-level hospitals in LMICs.

Mobile health technology transforms injury severity scoring in South Africa

BACKGROUND

The burden of data collection associated with injury severity scoring has limited its application in areas of the world with the highest incidence of trauma.

MATERIAL AND METHODS

Since January 2014, electronic records (electronic Trauma Health Records [eTHRs]) replaced all handwritten records at the Groote Schuur Hospital Trauma Unit in South Africa. Data fields required for Glasgow Coma Scale, Revised Trauma Score, Kampala Trauma Score, Injury Severity Score (ISS), and Trauma Score-Injury Severity Score calculations are now prospectively collected. Fifteen months after implementation of eTHR, the injury severity scores were compared as predictors of mortality on three accounts: (1) ability to discriminate (area under receiver operating curve, ROC); (2) ability to calibrate (observed versus expected ratio, O/E); and (3) feasibility of data collection (rate of missing data).

RESULTS

A total of 7460 admissions were recorded by eTHR from April 1, 2014 to July 7, 2015, including 770 severely injured patients (ISS > 15) and 950 operations. The mean age was 33.3 y (range 13-94), 77.6% were male, and the mechanism of injury was penetrating in 39.3% of cases. The cohort experienced a mortality rate of 2.5%. Patient reserve predictors required by the scores were 98.7% complete, physiological injury predictors were 95.1% complete, and anatomic injury predictors were 86.9% complete. The discrimination and calibration of Trauma Score-Injury Severity Score was superior for all admissions (ROC 0.9591 and O/E 1.01) and operatively managed patients (ROC 0.8427 and O/E 0.79). In the severely injured cohort, the discriminatory ability of Revised Trauma Score was superior (ROC 0.8315), but no score provided adequate calibration.

CONCLUSIONS

Emerging mobile health technology enables reliable and sustainable injury severity scoring in a high-volume trauma center in South Africa.

Perceptions of doctors and nurses at a Ugandan hospital regarding the introduction and use of the South African Triage Scale

Background

International Hospital Kampala (IHK) experienced a challenge with how to standardise the triaging and sorting of patients. There was no triage tool to help to prioritise which patients to attend to first, with very sick patient often being missed.

Aim and setting

To explore whether the introduction of the South African Triage Scale (SATS) was seen as valuable and sustainable by the IHK’s outpatient department and emergency unit (OPD and EU) staff.

Methods

The study used qualitative methods to introduce SATS in the OPD and EU at IHK and to obtain the perceptions of doctors and nurses who had used it for 3–6 months on its applicability and sustainability. Specific questions about challenges faced prior to its introduction, strengths and weaknesses of the triage tool, the impact it had on staff practices, and their recommendations on the continued use of the tool were asked. In-depth interviews were conducted with 4 doctors and 12 nurses.

Results

SATS was found to be necessary, applicable and recommended for use in the IHK setting. It improved the sorting of patients, as well as nurse-patient and nurse-doctor communication. The IHK OPD & EU staff attained new skills, with nurses becoming more involved in-patient care. It is possibly also useful in telephone triaging and planning of hospital staffing.

Conclusion

Adequate nurse staffing, a computer application for automated coding of patients, and regular training would encourage consistent use and sustainability of SATS. Setting up a hospital committee to review signs and symptoms would increase acceptability and sustainability. SATS is valuable in the IHK setting because it improved overall efficiency of triaging and care, with significantly more strengths than weaknesses.

Choice of injury scoring system in low- and middle-income countries: Lessons from Mumbai

INTRODUCTION

Injury is a major cause of morbidity and mortality in low- and middle-income countries. Effective trauma surveillance is imperative to guide research and quality improvement interventions, so an accurate metric for quantifying injury severity is crucial. The objectives of this study are (1) to assess the feasibility of calculating five injury scoring systems--ISS (injury severity score), RTS (revised trauma score), KTS (Kampala trauma score), MGAP (mechanism, GCS (Glasgow coma score), age, pressure) and GAP (GCS, age, pressure)--with data from a trauma registry in a lower middle-income country and (2) to determine which of these scoring systems most accurately predicts in-hospital mortality in this setting.

PATIENTS AND METHODS

This is a retrospective analysis of data from an institutional trauma registry in Mumbai, India. Values for each score were calculated when sufficient data were available. Logistic regression was used to compare the correlation between each score and in-hospital mortality.

RESULTS

There were sufficient data recorded to calculate ISS in 73% of patients, RTS in 35%, KTS in 35%, MGAP in 88% and GAP in 92%. ISS was the weakest predictor of in-hospital mortality, while RTS, KTS, MGAP and GAP scores all correlated well with in-hospital mortality (area under ROC (receiver operating characteristic) curve 0.69 for ISS, 0.85 for RTS, 0.86 for KTS, 0.84 for MGAP, 0.85 for GAP). Respiratory rate measurements, missing in 63% of patients, were a major barrier to calculating RTS and KTS.

CONCLUSIONS

Given the realities of medical practice in low- and middle-income countries, it is reasonable to modify the approach to characterising injury severity to favour simplified injury scoring systems that accurately predict in-hospital mortality despite limitations in trauma registry datasets.

Injury prevalence and causality in developing nations: Results from a countrywide population-based survey in Nepal

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.