Context: Treatment disparities for socioeconomically disadvantaged populations have been widely reported, but few studies have sought explanations for these disparities. Objective: To compare the quality of care for patients insured by Medicare alone, Medicare plus Medicaid, or Medicare plus private insurance and investigatemediators for potential disparities. Design, Setting, and Participants: Retrospective, random chart review of 3122 African American or White Medicare patients >65 years of age hospitalized for unstable angina in 22 Alabama hospitals, 1993-1999. Main Outcome Measures: Echocardiogram within 20 minutes of presentation; evaluation by a cardiologist; appropriate anti-platelet therapy within 24 hours of admission and at discharge, heparin for high-risk patients, betablockers during hospitalization, and performance of appropriate coronary angiography. Results: 182 (5.8%) had Medicare only, 433 (13.9%) had Medicare plus Medicaid, and 2507 (80.3%) had Medicare plus private insurance. Medicaid patients were more frequently Black, female, >85 years old, had multiple co-morbidities, or were admitted to hospitals without cardiac catheterization facilities (P<.001). Fewer Medicaid patients were admitted to hospitals with cardiac catheterization capabilities. Even after adjustment for demographics and hospital characteristics, Medicaid patients were less likely to see a cardiologist (odds ratio [OR] .57, 95% confidence interval [CI] .44-.73), receive antiplatelet therapy within 24 hours of admission (OR .66, 95% CI .50-.87), or heparin (OR .71, 95% CI .53-.97). No differences were seen with regard to having an electrocardiogram within 20 minutes of admission. Beta-blockers were used least in the Medicare-only patients, with only 37.7% receiving them (P=.04). Suitable Medicaid patients received coronary angiography less often, even after adjustment for demographics, co-morbidity, and prior revascularization (OR .68, 95% CI .48-.97). However, when adjusted for hospital characteristics, this finding was no longer observed (OR .94, 95% CI .64-1.39). Conclusions: Elderly Medicaid patients appear to receive poorer quality of care. This finding is partially, but not completely, explained by characteristics of the facilities where they are hospitalized.