Two thousand one hundred and thirty-five consecutive acute blunt trauma admissions over a 48 month period were studied in an attempt to produce an objective scale for the severity of injury. All traumatic injuries were coded using the International Classification of Disease and Accidents (H-ICDA) prior to computer analysis. An a priori probability of survival was computed for the population under study. The patient population was divided into a "Training Set" and a "Test Set." The training set was used to compute a conditional probability of survival (P sub C) for each diagnostic code, which was used to rank the injuries in order of severity. An effective probability of survival (P sub E) was then computed for each diagnostic code. The P sub E) for each code was used to predict survival rates on the test set and five random groups of the patients studied, and to predict individual survival using a decision rule that P sub E is greater than 0.5 resulted in survival.
Abstract