ETHNICITY,LANGUAGE,SPECIALTY CARE, AND QUALITY OF DIABETES CARE Tung T. Nguyen, MD; Nicholas A. Daniels, MD; Ginny L. Gildengorin, PhD; Eliseo J. Pe ´rez-Stable, MD Objective: To investigate ethnicity, language, specialty care, and quality of diabetes care in one medical center. Methods: Retrospective review of computer- ized records of patients with diabetes age $50 years who were regularly cared for in general medicine, family practice, or diabetes clinics from 1997 to 2000. Measures of processes of care were tests for creatinine, cholesterol, hemoglobin A1C (HbA1C), and microalbumin; ophthalmologic care; and total visits. Interme- diate outcomes were average systolic blood pressure (SBP) ,140 mm Hg and HbA1C ,8%. Results: Among 1323 patients, test rates for creatinine, cholesterol, microalbuminuria, and HbA1C were 76.6%, 54.7%, 17.2%, 78.8%, respectively. Only 31.0% had ophthalmology visits, 57.4% had SBP ,140 mm Hg, and 62.0% had HbA1C ,8%. In multivariate analyses, African Americans, Asians, and Lati- nos received more tests and had more total visits than Whites. Intermediate outcomes were similar except that Asians were more likely (odds ratio [OR]51.78, 95% confidence interval [CI] 1.26–2.50) to have SBP ,140 mm Hg. Limited English proficient pa- tients had more total visits (7.0) than English speakers (6.5) (P5.01). Compared to patients with only primary care, patients with a diabetes specialist had more microalbuminuria (OR 3.04, 95% CI 1.87–4.95) and HbA1C (OR 1.91, 1.12–3.26) tests, while those with both types of care were more likely to have each of the five process measures but less likely to have HbA1C ,8%. Conclusions: Quality of diabetes care was suboptimal for most patients. No ethnic disparity was seen in intermediate outcomes, which may have been achieved through more tests and visits. Combined care by primary and diabetes clinicians may be optimal. (Ethn Dis. 2007;17:65–71) Key Words: Diabetes, Ethnicity, Quality of Care INTRODUCTION Compared to non-Latino Whites, African Americans and Latinos have a higher prevalence of diabetes, worse glycemic control, and higher rates of complications, 1–3 while Asian Ameri- cans have a higher prevalence of diabetes after adjusting for body mass index. 4 In a health plan setting with similar access to care, ethnic minorities with diabetes had lower risks for myocardial infarc- tions and amputations but higher risks for renal failure. 5 Diabetic complications are reduced when patients have controlled hyperten- sion, treated hypercholesterolemia, gly- cemic control, and early treatment for early retinopathy and kidney disease. 6 The American Diabetes Association (ADA) guidelines include routine tests to monitor glucose control, complica- tions, and co-morbid conditions. 6 Most studies of diabetic quality of care use technical processes of care, such as regular receipt of tests, and measure- ments of intermediate outcomes, such as glucose control. Few studies have evaluated both ethnic and language differences in quality of diabetes care. 7–10 We aimed to examine the association of language, ethnicity, and specialty care on quality of diabetes care received by older adults in general internal medicine (GIM), family prac- tice (FP), and diabetes clinics at an academic health center. We hypothe- sized that limited English proficient (LEP) and non-White patients would have lower quality of care. METHODS Setting The University of California, San Francisco (UCSF) Medical Center serves a diverse population at two main sites, each with a hospital, emergency room, urgent care, and outpatient clinics; a third site provided outpatient FP and GIM care. In 2000, a total of 55,526 visits were recorded to GIM, 30,930 to FP, and 5718 to diabetes clinics. Insurance mix for these clinics was 40% managed care, 30% Medicare, 25% Medicaid, and 5% others (self- pay or fee-for-service). Clinics were connected to a computer database and received similar administrative support. Diagnostic laboratories were within a one-block walk. Ophthalmo- logic care was available at two sites. Attending physicians and fellows pro- vided care at all practices. Nurse practi- tioners and medical residents also pro- vided supervised care in GIM. Nearly one third of clinicians in the system were non-White, and three fourths spoke a second language. 11 Approxi- mately 30% of visits required interpre- tation, but despite availability of pro- fessional interpreters, no request was made in approximately half of these visits. 11 From the Division of General Internal Medicine, Department of Medicine, Med- ical Effectiveness Research Center for Di- verse Populations, University of California, San Francisco, California. Address correspondence and reprint re- quests to Eliseo J. Pe ´rez-Stable, MD; Box 0320; 400 Parnasssus Avenue; San Francisco, CA 94143-0320; 415-502-4088; 415-476- 7964 (fax); eliseops@medicine.ucsf.edu We hypothesized that limited English proficient (LEP) and non-White patients would have lower quality of care. Ethnicity & Disease, Volume 17, Winter 2007 65