Uses and abuses of statistical models for evaluating trauma care.

OBJECTIVE: A statistical model is a powerful tool for evaluating trauma care. With accident and emergency clinicians becoming more aware of the importance of monitoring their department's performance, there is a danger that such models may be inappropriately applied to new sets of trauma cases, resulting in misleading evaluations. The primary objective is to present a method for assessing whether an existing survival probability model can be correctly applied to a different data set.
METHODS: Correct and incorrect utilization of statistical models for evaluating the mortality of trauma patients are presented. Circumstances that may lead to the modification of an existing model, and the use of statistical models for evaluating different aspects of trauma care, are suggested.