Statistical validation of a severity of illness measure.

This paper provides statistical detail on the predictive power of a new severity of illness scale--APACHE (Acute Physiology and Chronic Health Evaluation)--on 833 consecutive medical admissions to an intensive care unit (ICU). In a multivariate logistic regression analysis of routine physiologic and other data obtained within 24 hours of ICU admission, severity of illness and age were significantly (p less than .0001) related to survival. Using the estimated equation to forecast death rates for independent data, APACHE allowed accurate estimates of death rates for groups of patients whose mortality at hospital discharge varied from 3 to 80 per cent. The Acute Physiology Score of APACHE is also strongly and significantly associated with outcome within a number of specific cardiovascular, neurologic, respiratory, and gastrointestinal diagnoses. After multi-institutional validation studies, APACHE could prove useful in a wide range of studies involving acutely ill patients.