There is marked debate about whether outcomes of care, particularly mortality, vary as a function of hospital and physician experience with a disease. This issue is especially important with respect to AIDS because greater than 200,000 individuals have now been diagnosed with this disease. We analyzed discharge data for 3,126 persons with AIDS who had Pneumocystis carinii pneumonia and who were treated at one of 73 New York City hospitals in 1987. In-hospital mortality was 25%. Factors associated with higher chances of short-term death were older age, being black, not having private health insurance, and being severely ill. A logistic regression model indicated that after controlling for differences in patient and hospital characteristics, the chances of death decreased when care was given at hospitals with higher caseloads of patients with Pneumocystis carinii pneumonia. Our findings suggest that hospital experience may decrease mortality in this subset of patients with human immunodeficiency virus disease, although it is unknown whether this is due to differences in quality of care.