Prospective validation of an out-of-hospital decision rule to identify severely injured children involved in motor vehicle crashes

Study objectives: We prospectively validate a previously published out-of-hospital clinical decision rule (based on Glasgow Coma Scale [GCS] score, passenger space intrusion, and restraint use) to identify severely injured children involved in motor vehicle crashes (MVCs) with a population-based sample. Methods: We selected 20 Los Angeles County fire stations (paramedic squad and engine company) according to pediatric MVC call volume during the previous year and geographic diversity, and 21 hospitals clustered around the fire stations. All stations (60 shifts) were trained in data collection and study background. Immediately after a pediatric MVC call, each engine or paramedic squad completed a data form (14 variables, including vital signs, GCS score, and vehicular and crash characteristics) for children evaluated at the scene of a crash during a 24-month period. Outcomes were assessed by retrospective medical record review. We performed surveillance at all hospitals to identify nontransported children presenting within 7 days of the crash. Results: The primary outcome was an Injury Severity Scale (ISS) score of 16 or greater, with a secondary outcome of the need for specialized trauma care (emergency intubation, operative intervention, or death in the first 24 hours). We calculated the sensitivity and specificity of the validation sample, including binominal confidence intervals (CIs). One hundred seventy-five children were prospectively enrolled in the study; 34 children were excluded because of lack of hospital participation (n=10) and missing hospital records (n=24). There were 4 (3%) children with ISS score of 16 or greater, 6 (4%) children requiring specialized trauma care, and 7 (5%) with 1 of the 2 outcomes. The sensitivity and specificity of the rule for each outcome was 100% (95% CI 40% to 100%) and 73% (95% CI 65% to 80%) for severe injury and 67% (95% CI 22% to 96%) and 73% (95% CI 64% to 80%) for specialized trauma care. All previously published rules would have missed 1 child requiring emergency surgery. Conclusion: Although there were a small number of severely injured children in this sample, the decision rule identified all of them with less combined under- and overtriage than any previously published pediatric triage rule. The decision rule was less sensitive for identifying children requiring specialized trauma care. Larger validation studies are needed to adequately assess the utility of this rule before implementation.