Study objective: Most injured patients taken by ambulance to hospital emergency departments do not require emergency surgery, yet most US trauma centers require a surgeon to be present on their arrival. If a clinical decision rule could be developed to accurately identify which injured patients require emergency operative intervention, then such "secondary triage" criteria could permit a trauma center to more efficiently use their surgeons' time.
Methods: We analyzed 7.5 years of data (8,289 consecutive trauma activations) in our prospectively maintained Level I trauma center registry. We used classification and regression tree analyses to generate clinical decision rules using standard out-of-hospital variables to identify emergency operative intervention (within 1 hour) by a general surgeon (for adults) or a pediatric surgeon (if < or =14 years).
Results: Emergency operative intervention occurred in 3.0% of adults and 0.35% of children. For adults, summoning a surgeon for any one of 3 criteria (penetrating mechanism, systolic blood pressure <96 mm Hg, pulse rate >104 beats/min) could reduce surgeon calls by 51.2% while failing to identify emergency operative intervention in only 0.08% (rule sensitivity 97.2% and specificity 48.6%). For children, no rule at all (ie, never automatically summoning a surgeon) would fail to identify emergency operative intervention in only 0.35% of patients, and use of a single criterion (penetrating mechanism) would reduce surgeon calls by 96.2% while failing to identify emergency operative intervention in only 0.09% (rule sensitivity 75.0% and specificity 96.5%).
Conclusion: We have derived simple decision rules for trauma centers that, if validated, could substantially reduce the need for routine surgeon presence on trauma patient arrival. These rules demonstrate low false-negative rates.