To assess the relation between HIV/AIDS hospital caseload and mortality in Canada.
Descriptive, population-based study.
All hospitals in Canada that admitted any patients with HIV or AIDS between Mar. 31, 1987, and Apr. 1, 1994.
All patients with a diagnostic code on their hospital discharge abstract for HIV infection, AIDS, or with positive serological or viral culture findings for HIV (International Classification of Diseases, 9th revision, 042, 043, 044 or 795.8).
Over the study period, 38,075 admissions attributed to HIV/AIDS (33,380 of men and 4695 of women) were recorded in 513 Canadian hospitals. Of these hospitals, 230 (45%) had fewer than 1 admission per year of patients with HIV/AIDS; 200 (39%) had between 1 and 9; 68 (13%) had between 10 and 99; and 15 (3%) had 100 or more. HIV/AIDS-related admissions ending in death were independently associated with the patient being admitted to lower-volume hospitals, being hospitalized for longer periods of time, and being older, male and at a more advanced stage of disease. During the study period, hospitals with 100 or more admissions per year reported 36% lower mortality among patients with HIV/AIDS than those that had fewer than 1 admission per year.
There is an inverse relation between hospital caseload and in-hospital mortality among patients with HIV/AIDS in Canada. We attribute this association at least in part to the propensity of high-volume hospitals to deal more effectively with seriously ill patients with HIV/AIDS.