To date, there has been little or no systematic research on the quality of medical care delivered to patients with AIDS. For Pneumocystis carinii pneumonia (PCP), we have found that hospital death rates in California very as much as threefold within institutions across time periods and more than twofold across institutions. How much of this variation can be attributed to quality of care differences is unknown. We propose to examine medical records of 300 patients who presented with PCP as the initial AIDS diagnosis to determine how the structure and process of care affect 30-day mortality. Three study institutions were selected to represent a diverse sample of acute care facilities used by AIDS patients in Southern California. The object of study is the """"""""episode of care"""""""" for individual patients, including the pre-hospital interval of care and extending 30 days following hospital admission. Study variables are categorized as outcomes of care (death within 30 days and serve respiratory failure), technical process indicators, interpersonal process indicators, structural features of care, and prognostic factors assessed on admission. To identify process or structural features of medical care that predict important outcomes among patients which similar prognoses, we will develop a multipredictor model of expected death rates. Both cumulative incidence and person-time measures for the outcomes will be examined by stratified analyses and by multivariate modeling. We will also examine aspects of care that are important in the longer term (discharge disposition) and that reflect interpersonal art of care surrounding information-giving and clinical decisionmaking. This project will contribute to clinical research and patient care by identifying which features of care make the greatest difference in patient outcomes, for the most common presenting manifestation of AIDS.
