OBJECTIVE: To investigate the association between selecting one’s own primary care physician and adherence to treatment regimens as measured by patients’ report of self-care behaviors and objective records of adherence to recommended prevention behaviors. STUDY DESIGN: A mail survey of physicians and their patients with diabetes. PATIENTS AND METHODS: Forty physicians and 1200 of their patients with diabetes (30 per physician) from 3 health centers of Kaiser Permanente in Northern California were surveyed. The questionnaires asked about satisfaction, treatment adherence, and physician behavior. Data from the Kaiser Permanente Diabetes Registry of Northern California were used to assess whether patients had undergone recommended prevention tests and screenings. RESULTS: Patients who chose their primary care physicians reported significantly greater adherence to their treatment regimens (P < .01) than those assigned to a primary care physician, and this relationship remained significant after controlling for possible confounding factors (eg, physician gender, patient gender, length of relationship). Objective records of prevention behaviors indicated that patients who chose their physicians were significantly more likely to have had a retinal exam (P < .02) and tests for total cholesterol (P < .001), high-density lipoprotein cholesterol (P < .03), and glycosylated hemoglobin (P < .02) during the past 12 months. CONCLUSIONS: The findings suggest that the manner in which patients are linked with primary care physicians is associated with patient behavior; therefore, HMOs might increase opportunities for patients to choose their doctors, while determining those factors that affect patient choice and why choice makes a difference.