A. Laytin1, D. Clarke2, M. Gerdin Wärnberg3, V. Kong2, J. Bruce2, G. Laing2, C. Juillard4 3Karolinska Institutet,Solna, STOCKHOLM, Sweden 4University Of California – San Francisco,San Francisco, CA, USA 1Hospital Of The University Of Pennsylvania,Philadelphia, PA, USA 2University of KwaZulu-Natal,Pietermaritzburg, KWAZULU-NATAL, South Africa
Introduction: Injury care is a global health priority with 5 million deaths due to injury per year worldwide, and the burden of injury is especially high in low- and middle-income countries (LMIC). Efforts to strengthen injury care in LMIC benefit from accurate injury scores to quantify injury severity and predict a patient’s likelihood of mortality. The Injury Severity Score (ISS) and Trauma Score-Injury Severity Score (TRISS) are widely used in the US, but require comprehensive anatomic injury data collection that is often impractical in LMIC. We hypothesized that a simple injury score appropriate for resource-limited settings could achieve discrimination and strength of association with in-hospital mortality similar to resource-intensive injury scores.
Methods: This study uses data collected in a regional trauma registry in KwaZulu-Natal, South Africa. Data from 2012-2017 were used to compare the discrimination and strength of association with in-hospital mortality of two comprehensive anatomic injury scores—ISS and TRISS—with those of four relatively simple injury scores that rely primarily on physiologic data—Shock Index (SI), Glasgow Coma Score (GCS), Revised Trauma Score (RTS) and Kampala Trauma Score (KTS). Discrimination was assessed with ROC curve analysis. Strength of association with in-hospital mortality was assessed with standardized regression coefficients (β). KTS was developed in Kampala, Uganda in 2000. While uncommon in the US, it has been used as an injury score in several sub-Saharan African countries.
Results: Trauma registry data were reviewed for 4,179 patients, disclosing a median age of 30 years, a male preponderance of 84% and a 48% prevalence of penetrating injury mechanisms. Median time from injury to presentation was 13 hours, with an in-hospital mortality rate of 2.5%. TRISS, ISS and KTS had similar discrimination and strength of association with in-hospital mortality, while the other injury scores demonstrated weaker discrimination and strength of association, especially among patients presenting more than 6 hours post-injury.
Conclusion: In searching for a robust injury score to deploy in LMIC, KTS evidenced discrimination and strength of association with in-hospital mortality similar to the gold-standard injury scores ISS and TRISS. Using KTS can help to measure changes in outcomes over time, to compare outcomes between LMIC medical centers and to evaluate the impact of performance improvement efforts when calculating ISS or TRISS is not feasible. Presentation delay degraded the utility of the other injury scores that principally rely on physiologic data and may reflect survival bias in that patient population.
