Comparative performance of reverse shock index multiplied by simplified motor score (rSI-sMS) and conventional indices for predicting outcomes in trauma patients: a retrospective analysis of a 13-year multicenter prospectively maintained trauma regis

BACKGROUND: The reverse shock index multiplied by the simplified motor score (rSI-sMS) can rapidly reflect hemorrhagic shock and neurological status for rapid injury severity assessment in trauma patients during the emergency department (ED) triage, even when patient information is incomplete. However, its predictive and discriminatory performance for short- and long-term trauma outcomes remains to be validated as superior to conventional physiological indicators. METHODS: This study was a 13-year multicenter validation cohort study conducted across three geographically distinct campuses of Tzu Chi Hospital within the same healthcare system network. We retrospectively analyzed data from a prospectively maintained multicenter trauma registry to compare the discriminatory performance of the rSI-sMS, shock index (SI), modified shock index (mSI), reverse shock index multiplied by the Glasgow Coma Scale motor subscale (rSI-GCSM), and reverse shock index multiplied by the Glasgow Coma Scale (rSI-GCS) in predicting 3- and 7-day mortality, in-hospital mortality, ≥14-day intensive care unit (ICU) stay, and ≥30-day total hospital stay among trauma patients. RESULTS: This study included 26,606 trauma patients. The rSI-sMS's discriminatory performance for in-hospital mortality, 3-day mortality, 7-day mortality, ≥14-day ICU stay, and ≥30-day total hospital stay (areas under the receiver operating characteristic curves [AUROCs]: 0.722, 0.793, 0.774, 0.605, and 0.639, respectively) was superior to those of SI (AUROCs: 0.540, 0.580, 0.557, 0.513, and 0.576, respectively), mSI (AUROCs: 0.551, 0.584, 0.567, 0.522, and 0.577, respectively), and rSI-GCS (AUROCs: 0.707, 0.773, 0.758, 0.601, and 0.637 respectively; all P < 0.001). rSI-sMS was also superior at discriminating clinical outcomes in patients with injury severity scores ≥ 16, chronic diseases, cardiovascular diseases, and traumatic brain injuries, as well as geriatric and nongeriatric patients. The optimal rSI-sMS cutoff of 4 demonstrated 79.5%, 79.4%, 79.5%, 66.6%, and 78.7% overall accuracy for 3-day mortality, 7-day mortality, in-hospital mortality, ≥14-day ICU stay, and ≥30-day total hospital stay, respectively. CONCLUSIONS: Compared with SI, mSI, and rSI-GCSM, rSI-sMS demonstrated modestly improved discriminatory performance across several clinical outcomes, including mortality, ICU admission, and prolonged hospitalization, with discriminatory ability comparable to that of rSI-GCS. However, its performance for long-term hospitalization outcomes remained limited, and its clinical impact and practical utility require further confirmation through prospective studies and validation.