Interest of the MGAP score on in-hospital trauma patients: Comparison with TRISS, ISS and NISS scores

Trauma scoring systems were created to predict mortality and enhance triage capabilities. However, ef-ficacy of scoring systems to predict mortality and accuracy of originally reported severity thresholds re-mains uncertain. A single-center, retrospective study was conducted at University of Virginia (UVA), an American College of Surgeons verified Level I trauma center. We compared four scoring systems: MGAP (Mechanism, Glasgow Coma Scale, Age, and arterial pressure), Injury Severity Score (ISS), New Injury Severity Score (NISS), and Trauma Related Injury Severity Score (TRISS) to predict in-hospital mortal-ity and disposition from the emergency department to higher acuity level of care including mortality (i.e. operating room, intensive care unit, morgue) versus standard floor admission using area under the curve (AUC) for receiver operating characteristic analysis. Second, we examined sensitivity of these scores at standard thresholds to determine if adjustments were needed to minimize under-triage (sensitivity >= 95%). TRISS was the best predictor of mortality in a cohort of n = 16,265 with AUC of 0.920 (95% CI: 0.911-0.929, p < 0.0 0 01), followed by MGAP with AUC of 0.900 (95% CI: 0.889-0.911, p < 0.0 0 01), and fi-nally ISS and NISS (0.830 (95% CI: 0.814-0.847) and 0.827 (95% CI: 0.809-0.844) respectively). NISS was the best predictor of high acuity disposition with an AUC of 0.729 (95% CI: 0.721-0.736, p < 0.0 0 01), fol-lowed by ISS with AUC of 0.714 (95% CI: 0.707-0.722, p < 0.0 0 01), and finally TRISS and MGAP (0.673 (95% CI: 0.665-0.682) and 0.613 (95% CI: 0.604-0.621) respectively ( p < 0.0 0 01). At historic thresholds, no scor-ing system displayed adequate sensitivity to predict mortality, with values ranging from 73% for ISS to 80% for NISS. In conclusion, in the reported study cohort, TRISS was the best predictor of mortality while NISS was the best predictor of high acuity disposition. We also stress updating scoring system thresholds to achieve ideal sensitivity, and investigating how scoring systems derived to predict mortality perform when predicting indicators of morbidity such as disposition from the emergency department. (c) 2022 Elsevier Ltd. All rights reserved.