OBJECTIVE: To evaluate the impact of a continuous quality improvement (CQI) intervention on glycemic control of patients with diabetes mellitus attending a primary care clinic. METHODS: A CQI process designed by the Minnesota Department of Health to improve diabetes care was implemented at a volunteer clinic, with another similar clinic not using the CQI process assessed for comparison. Adults with diabetes were identified at both clinics using diagnostic and pharmacy databases. Glycosylated hemoglobin (HbA1c) values (reference range, 4.3%-6.1%) and out-patient utilization and charges were compared for all patients with diabetes at each clinic for the 12 months before and 18 months after initiation of the CQI intervention. RESULTS: The mean HbA1c value at the intervention clinic fell from 8.9% at baseline to 8.4% at 12 months and to 7.9% at 18 months. The mean HbA1c value at the comparison clinic was 8.9% at baseline, 8.9% at 12 months, and 8.8% at 18 months (difference between clinics, t = 4.13, P < .001). Differences after the intervention in the proportion of patients at the comparison clinic (n = 121) vs the intervention clinic (n = 122) with HbA1c values of 8% or less (40% vs 51%), between 8% and 10% (33% vs 37%), and 10% or greater (27% vs 12%) were unlikely due to chance (chi 2 = 9.7, 2 df, P = .008). The intervention was not associated with increased utilization of outpatient visits or outpatient charges. CONCLUSIONS: Involvement of nurses, physicians, and managers in a CQI process can improve patients' glycemic control in some health maintenance organization primary care settings, without increasing utilization or charges. Health maintenance organizations should consider CQI as one possible method to improve diabetes outcomes.