INCREMENTAL BENEFIT OF INDIVIDUAL AMERICAN-COLLEGE OF SURGEONS TRAUMATRIAGE CRITERIA

Objective: To determine the incremental benefit of individual American College of Surgeons (ACS) trauma triage criteria for prediction of se vere injuries after consideration of concurrent physiologic, anatomic, mechanism, or ''other'' criteria. Methods: A prospective cross-sectio nal study of motor vehicle crash victims transported to any of the 12 hospitals in a suburban/rural county by local ambulance services was p erformed. Demographic and individual ACS criteria were collected using structured data instruments, EDs provided patient disposition within 24 hours of patient arrival. Medical records were reviewed. Major outc omes were admission, operative interventions (OR), major nonorthopedic operative interventions or death (Maj-OR), and injury severity score (ISS). To optimize sensitivity and specificity of out-of-hospital tria ge decision rules, receiver operating characteristic (ROC) curves were derived. Results: Of 1,545 patients, 13% were admitted; 6% had OR; 1% had Maj-OR; and 3% had ISSs greater than or equal to 16. For all outc omes, the most useful criteria were physiologic and anatomic, Some add itional criteria (crash speed >20 mph, greater than or equal to 30-inc h vehicle deformity, axle displacement) substantially worsened specifi city, with minimal or no improvement in sensitivity, For example, the optimal ROC curve for Maj-OR was determined by a systolic blood pressu re <90 mm Hg, Glasgow Coma Scale (GCS) score <13, respiratory rate (RR ) <10 or >29, death of a same-car occupant, penetrating injury, and/or greater than or equal to 24-inch opposite-side compartment intrusion (sensitivity, 85% specificity, 87%). An ISS greater than or equal to 1 6 was predicted by GCS score <13, RR <10 or >29, penetrating injury, 2 proximal long bone fractures, flail chest, greater than or equal to 2 4-inch opposite-side compartment intrusion, patient ejection, rollover , and/or age <5 or >55 years (sensitivity, 86%; specificity, 70%). Con clusion: Physiologic and anatomic trauma triage criteria predicted inc reased hospital resource utilization and severe injury. On the other h and, when used concurrently with physiologic, anatomic, and ''other'' criteria, some mechanism criteria worsen specificity with negligible i mprovement in sensitivity. In particular, crash speed >20 mph and grea ter than or equal to 30-inch vehicle deformity had little predictive v alue for all outcomes.