Objective: It is common practice for hospitals to use a trauma team activation criteria (TTAC) to identify patients at risk of major trauma and to activate a multidisciplinary team to receive such patients on arrival to the ED. The aims of this study are to describe the frequency of individual criteria and the ability of one currently used system to predict major trauma, and to estimate the effect of simplified criteria on the prediction.
Design and setting: A retrospective observational study of the entire cohort of adult patients who a) received trauma team activation or b) were included in the trauma registry of Royal Darwin Hospital in 2015. From the original clinical record all components of the TTAC, and corresponding outcomes, were extracted for each case. The predictive effect of each criterion, adjusted for the presence of others, was assessed by logistic regression. The poorest predictors were sequentially “dropped” to develop a number of models of which the predictive value of the resulting hypothetical TTAC was calculated.
Main outcome measures: Major trauma (MT) was defined as a death in ED, immediate operative intervention or direct admission to ICU. Overtriage was defined as activation of the trauma team without major trauma. Undertriage was defined as major trauma without trauma team activation.
Results: 794 trauma presentations were reviewed, 428 of those presentations met TTAC. Major trauma was present in 135 (32%) of those with TTAC hence overtriage was 68%. Criteria based on mechanism of injury (MOI) were responsible for over half of the overtriage and were collectively present without other activation criteria in only 10 MTs (6%). Removal of the criteria with the worst predictive value decreased overtriage to 50% before a rise in undertriage to beyond 24%.
Conclusion: A number of criteria including those based on MOI decrease the accuracy of TTAC and lead to high rates of overtriage. Airway, respiratory and neurological compromise were the best predictors of MT. Any criteria simplification should be introduced in the context of a further audit of TTAC performance, as the estimates of the separate criteria in the current TTAC are not robustto bias or to undetected correlation.