Patterns of care for HIV-related Pneumocystis carinii pneumonia in a university medical program: a case study

Previous studies have identified large variations in patterns of in-hospital acquired immunodeficiency syndrome (AIDS) care among groups of hospitals and physicians. We evaluated the patterns of care for patients with AIDS-related Pneumocystis carinii pneumonia (PCP) care at a single university program with patients treated at an adjacent county and Veterans' Administration (VA) hospital. All medical care was provided by physicians in a single residency program, but attending physician staffs were separate.

A randomized sample of patients with human immunodeficiency virus (HIV)-related PCP from the two hospitals who received care between 1987 and 1990.

During the study years, the VA hospital provided care for approximately one fourth as many AIDS patients as the county hospital. Patients at the VA hospital had a higher bronchoscopy rate (39.7% versus 27.7%, P = .05), higher intensive-care unit (ICU) rate (11.8% versus 2.9%, P = .008), longer hospitalizations (mean length of stay of 18.9 versus 13.9 days, P = .004), but delayed initiation of anti-PCP therapy (median of day 2 versus day 1, P < .05). The odds of death were significantly different between the VA and county hospitals, even after adjusting for differences in important patient characteristics.

Patterns of in-hospital PCP care differed between the two hospitals of this medical school. Possible explanations include organizational differences related to level of attending physician HIV experience, hospital case loads of AIDS patients, or specific hospital considerations such as concerns over tuberculosis exposure.