To determine the degree of compliance with clinical standards among hospitals for care of breast cancer patients and account for variations in compliance. Analysis of cancer registry data submitted to the American Cancer Society, Illinois Division, Chicago, for a concurrent prospective descriptive study of breast cancer, supplemented by other hospital data from public sources. Ninety-nine Illinois hospitals evenly distributed among rural counties, counties with small cities outside the Chicago metropolitan area, exurban counties in the Chicago metropolitan area, suburban Cook County, and urban Chicago. A total of 5766 newly diagnosed patients with histologically confirmed breast cancer in 1988, representing 84% of the estimated 6900 new cases in the state for that year. Descriptive statistics and multiple linear regression analyses of five dependent quality variables from clinical indicators related to early diagnosis, hormone receptor determination, adjuvant therapy, radiation therapy, and axillary lymph node dissection. At the hospitals studied, (1) late stage (IIb through IV) at diagnosis was associated with urban location, higher proportion of poorly insured patients, fewer breast cancer cases treated, and lower oncology charges (proportion of variance explained, R2 = .50, P less than .00001); (2) omission of hormone receptor test for stages II through IV was associated with urban location and higher proportion of poorly insured patients (R2 = .18, P less than .00003); and (3) omission of indicated radiation therapy was associated with urban location and fewer breast cancer cases (R2 = .21, P less than .00001). Omission of adjuvant therapy and omission of axillary lymph node dissection were not significantly associated with any of the hospital variables examined. The findings suggest that there is a group of urban hospitals, generally small and marginally reimbursed, where comprehensive diagnosis and treatment of breast cancer are not obtained.