Validation of CRASH Model in Prediction of 14-Day Mortality and 6-Month Unfavourable Outcome of Pediatric Traumatic Brain Injury

Value of Corticosteroid Randomisation after Significant Head Injury (CRASH) prognostic model has not been assessed in children with traumatic brain injury (TBI). This study is designed to examine the value of CRASH model in prediction of 14-day mortality and 6-month unfavourable outcome of pediatric TBI. In a cross-sectional study, 738 children with TBI brought to the emergency ward of four hospitals were studied. For assessing the predictive value of the CRASH model discrimination power and calibration of CRASH basic model and CRASH CT model were examined. The areas under the curve (AUC) of CRASH basic and CRASH CT models in prediction of 14-day mortality were 0.89 and 0.91, respectively. AUCs of the CRASH basic and CRASH CT models in predicting unfavourable outcome were 0.93 and 0.94. The value of two models in prediction of 14-day mortality (p=0.20), and 6-month unfavourable outcome (p=0.22) were equal. Both models had proper calibrations in predicting 14-day mortality and 6-month unfavourable outcome. As calculations of the basic model are easier than those of the CT model and it does not necessitate CT scanning, the CRASH basic model is suggested in the field of pediatrics.