One hundred geriatric patients who suffered injury severe enough to necessitate hospitalization were compared retrospectively to a random group of 100 younger patients. The elderly suffered different types of injury and died six times as often as their younger peers, even when controlled for injury severity. The PRE method was employed to examine outcome in both groups and was found to be strongly predictive of death in young patients. Age stratification aided significantly in predicting mortality in elderly patients. Regression analysis was employed to examine the data set to determine the relative importance of several variables in the prediction of ultimate mortality. By incorporating all the data from the entire data set, curves describing the contribution of age and shock to mortality corrected for all factors is possible. Increasing age after 65 increases mortality and this effect is dramatically increased by the presence of shock. This information may be useful in counselling the injured elderly and their families.