The objective of our study was to evaluate the applicability of the Acute Physiology and Chronic Health Evaluation (Apache II) scoring system in an Israeli general intensive care unit (GICU), and to monitor the outcome of care using the Apache II severity score. We compared over a 3-year period the distribution of Apache II score, and of hospital mortality in our GICU (790 consecutive admissions) to that of the original American study, using a prediction formula based on 5,815 admissions in the United States. Our admissions were classified as post-elective surgery, post-emergency surgery or nonsurgical. Patients' ages were 53 +/- 21 years. The mean APache II score was 13.3 +/- 7.8, and was significantly lower in the 612 survivors (11.1 +/- 6.7) than in the 178 nonsurvivors (19.2 +/- 8.4). The in-hospital mortality rate was higher, 22.5%, compared to the predicted 18.3% (odds ratio 1.14, 95% CI 1.01, 1.27). The majority of excess mortality occurred in the nonsurgical group (odds ratio 1.27, 95% CI 1.09, 1.48). We found the Apache II score useful for determining the severity of illness in our GICU. The Apache equation accurately predicts the mortality of elective and emergency surgery admissions, but underestimates the mortality of nonsurgical admissions.