An Assessment of Clinical Accuracy of Vital Sign–based Triage Tools Among U.S. and Coalition Forces

Abstract

ABSTRACT Introduction Early appropriate allocation of resources for critically injured combat casualties is essential. This is especially important when inundated with an overwhelming number of casualties where limited resources must be efficiently allocated, such as during mass casualty events. There are multiple scoring systems utilized in the prehospital combat setting, including the shock index (SI), modified shock index (MSI), simple triage and rapid treatment (START), revised trauma score (RTS), new trauma score (NTS), Glasgow Coma Scale + age + pressure (GAP), and the mechanism + GAP (MGAP) score. The optimal score for application to the combat trauma population remains unclear. Materials and Methods This is a secondary analysis of a previously described dataset from the Department of Defense Trauma Registry from January 1, 2007 through March 17, 2020. We constructed univariable analyses to determine the area under the receiving operator characteristic (AUROC) for the scoring systems of interest. Our primary outcomes were early death (within 24 hours) or early massive transfusion, as defined by ≥3 units. Results There were 12,268 casualties that met inclusion criteria. There were 168 (1%) who died within the first 24 hours and 2082 (17%) that underwent significant transfusion within the first 24 hours. When assessing the predictive capabilities for death within 24 hours, the AUROCs were 0.72 (SI), 0.69 (MSI), 0.89 (START), 0.90 (RTS), 0.83 (NTS), 0.90 (GAP), and 0.91 (MGAP). The AUROCs for massive transfusion were 0.89 (SI), 0.89 (MSI), 0.82 (START), 0.81 (RTS), 0.83 (NTS), 0.85 (MGAP), and 0.86 (GAP). Conclusions This study retrospectively applied seven triage tools to a database of 12,268 cases from the Department of Defense Trauma Registry to evaluate their performance in predicting early death or massive transfusion in combat. All scoring systems performed well with an AUROC >0.8 for both outcomes. Although the SI and MSI performed best for predicting massive transfusion (both had an AUROC of 0.89), they ranked last for assessment of mortality within 24 hours, with the other tools performing well. START, RTS, NTS, MGAP and GAP reliably identified early death and need for massive transfusion, with MGAP and GAP performing the best overall. These findings highlight the importance of assessing triage tools to best manage resources and ultimately preserve lives of traumatically wounded warfighters. Further studies are needed to explain the surprising performance discrepancy of the SI and MSI in predicting early death and massive transfusion.