Interhospital differences in cancer survival.

We examined variations in cancer survival rates among a large number of hospitals in the United States. Survival rates for breast cancer, prostate cancer, and Hodgkin's disease were calculated from patient care studies of the American College of Surgeons and were linked to data on hospital characteristics from the surveys of the American Hospital Association. When patient, disease, treatment, and institutional characteristics were examined in multivariate analyses, medical school affiliation, residency training, and community size were not related to hospitals' cancer survival experience. Patterns of care involving greater use of staging laparotomy with splenectomy for Hodgkin's diseases, lesser use of mastectomy without axillary dissection for breast cancer and, greater use of bone scanning and lesser use of hormone therapy for prostate cancer all were associated with better survival. The differences between hospitals' survival rates were large but we found that the differences were more a function of patient characteristics, disease stage, and tumor histology than of hospital affiliations, location, size, facilities, or treatment patterns. These findings provide some data upon which future public health interventions to affect cancer mortality may be planned and evaluated.