Two hundred one consecutive adult admissions to a 3-bed Special Care Unit in a 25-bed military hospital were scored using APACHE II (Acute Physiology and Chronic Health Evaluation). Outcome measures included APACHE II scores and mortality predictions; active intensive care interventions; transfers for specialized care; and mortality. Although higher scores generally reflected the need for intensive care intervention, admissions with nondiagnostic chest pain had scores that did not accurately predict their course. This finding could be explained by bias in the original APACHE case mix and by the need for further subclassification of cardiovascular disease. APACHE II scoring can be applied in small intensive care settings. Scoring criteria and logistic regression equations may need to be customized accordingly.