The development of effective methods for predicting the survival or mo
rtality of patients is a major focus of trauma research. We address th
is problem by applying neural network modelling, using as inputs estab
lished descriptors of physiologic and anatomic injury severity, We use
two different ''backpropagation'' techniques with a variety of traini
ng strategies and examine the importance of training and test set comp
osition, performance evaluation criteria, and interpretation of networ
k outputs, The results of our study indicate that neural networks show
significant promise for trauma patient outcome evaluation.