OBJECTIVE - To measure the quality of diabetic care as indicated by HbA(1c) testing frequency and HbA(1c) values and to demonstrate improvement in care after an appropriate quality improvement intervention. RESEARCH DESIGN AND METHODS- The quality improvement project used computerized claims and laboratory data relating to HbA(1c) testing among the private practices of nine physicians caring for diabetic Medicare patients. Nine indicators evaluated three main areas: HbA(1c) testing frequency HbA(1c) values, and frequency of office visits. A quality improvement intervention consisting of a physician component and a patient component was implemented. RESULTS - There were 835 patients and 4,367 visits studied. After the intervention, statistically significant improvements in HbA(1c) testing frequency and values were noted. Rates of seized opportunities for testing HbA(1c) improved from 17.7 to 33.9% (P < 0.0001), The percentage of patients with a current HbA(1c) value improved from 31.3 to 47.6% (P < 0.0001), The median HbA(1c) values fell from 8.5 to 7.8% (P < 0.006). Patients achieving good or fair control (HbA(1c) less than or equal to 8%) improved from 43.8 to 56.9% (P = 0.007). The median time between physician visits fell from 70 days to 60 days (P < 0.0001). CONCLUSIONS - The study revealed that HbA(1c) testing was underused but that after a quality improvement initiative, a significant increase in testing use could be achieved, The quality improvement initiative also resulted in significant improvements in glycemic control. The techniques and interventions used in this study could be used to intervene in larger populations and practice settings to improve medical care for diabetic patients.