Abstract 70: Improving the Prehospital Trauma Triage Guidelines for Physiologic Derangement: Can We Do Better?

Abstract

Introduction: The American College of Surgeons (ACS) recommends immediate transport to a trauma center for all injured persons with field physiologic derangement (ACS Step 1 trauma triage criteria). However, it remains unclear whether these criteria could be refined to identify high-risk trauma patients with improved predictive value. Hypothesis: We assessed the hypothesis that an easily applied, prehospital, clinical decision rule could be developed to better identify high-risk injured persons meeting ACS Step 1 criteria. Methods: We conducted a prehospital prospective cohort study of injured adults > 15 years where field resuscitation was attempted and ACS physiologic criteria were present: SBP <<26>90, RR <10 or >29 breaths/minute, GCS <<26>12, or field intubation. Population-based data were collected from December 1, 2005 through November 30, 2006 by 268 EMS agencies transporting to 287 acute care hospitals (trauma and non-trauma centers) in 11 sites across the U.S. and Canada. Ten prehospital variables were assessed: SBP, RR, GCS, pulse, pulse oximetry, shock index (pulse/SBP), mechanism of injury, penetrating injury, age, and gender. High risk injured persons included: death (before or after admission) or hospital LOS >2 days. 60% of the sample was randomly selected for rule derivation and analyzed using classification and regression tree analysis. The remaining 40% were used for rule validation. The final rule was based on a targeted sensitivity of >90%. Results: Of 4,983 eligible patients, 4,326 injured persons had complete outcome information and were included in the analysis. The sample included 2,495 (58%) persons with death or LOS >2 days, as well as 1,061 (25%) patients that did not require admission. The final rule included the following variables (in order): field intubation, GCS < 8, shock index > 1.4, mechanism of injury, and age >=70 years. Rule validation demonstrated the following accuracy measures: sensitivity 94.9% (95% CI 93.5–96.3%), specificity 14.1% (95% CI 11.6 –16.6%). Conclusions: While we were able to generate a highly sensitive decision rule, the rule appears too complex for field application and is non-specific. Further refining ACS physiologic triage criteria to increase predictive value remains an ongoing challenge.