Predictive accuracy of adding shock index to the American College of Surgeons' minimum criteria for full trauma team activation

Background The American College of Surgeons requires trauma centers to use six minimum criteria (ACS-6) for full trauma team activation: hypotension, gunshot wound to the neck or torso, Glasgow Coma Scale (GCS) score < 9, respiratory compromise, transfers receiving blood transfusion, or physician discretion. Our goal was to evaluate the effect of adding varying shock index (SI) thresholds to the ACS-6 in an adult trauma population with the hypothesis that SI would significantly improve sensitivity at the expense of an acceptable decrease in specificity. Methods We performed a secondary analysis of EMS and trauma registry data from an urban Level I trauma center. Consecutive patients > 15 years of age were included from 1993 through 2006. SI at thresholds of >= 0.8, >= 0.85, >= 0.9, and >= 1 were evaluated. Primary outcome was emergency operative (within 1 h of arrival) or procedural (cricothyrotomy or thoracotomy) intervention (EOPI); secondary outcomes were Injury Severity Score (ISS) > 15, ISS > 24, a composite of EOPI or ISS > 15, and urgent operative intervention (within 4 h). Results A total of 20,872 patients were included, 27% with an ISS > 15 and 5% who underwent EOPI. Sensitivity and specificity of the ACS-6 alone for EOPI were 86% (95% confidence interval [CI] = 84% to 88%) and 81% (95% CI = 80% to 81%), respectively. Inclusion of SI thresholds of 0.8, 0.85, 0.9, and 1 resulted in sensitivities of 95% (95% CI = 93% to 96%), 93% (CI = 91% to 94%), 92% (95% CI = 90% to 93%), and 90% (95% CI = 88% to 92%), respectively, and specificities of 52% (95% CI = 51% to 52%), 59% (95% CI = 58% to 59%), 64% (95% CI = 64% to 65%), and 72% (95% CI = 71% to 73%), respectively. Similar trends were found for each secondary outcome. Conclusion The addition of SI to the ACS-6 for trauma team activation increased sensitivity for EOPI with a larger decrease in specificity across all thresholds. Inclusion of a SI threshold of >= 0.9 closely aligns with under- and overtriage benchmarks in this trauma registry cohort using a strict definition of trauma team activation need.