Language barriers and diabetes care among urban, underserved Latinos in a teaching hospital in northern California

Language barriers are one major cause of ineffective diabetes care, thus of diabetes disparities, among US minority populations. Latinos, the largest and fastest growing minority population in the US, are disproportionately affected by diabetes. Because at least 21 million of US Latino patients have limited English proficiency (LEP), Latinos are at high risk of experiencing language barriers in the course of receiving diabetes care. Prior research by our team has indicated that LEP is an independent predictor for poor diabetes control among insured Latinos, yet only when care is provided by language discordant (LD) providers, even when these are aided by professional medical interpreters, but not when care is provided by language-concordant (LC) providers. Further, use of professional interpreters to overcome language barriers may not be sufficient to eliminate disparities in diabetes for LEP patients, thus research comparing the clinical effectiveness in communication modalities (bilingual primary care provider vs. interpreter-mediated care) is needed. The aim of our study was to describe patients' and providers' perspectives on LC and LD diabetes medical encounters. Our study site was San Francisco General Hospital General Medical Clinic, which counts with award-wining interpreter services. Study participants either provided or received diabetes care at this site. We interviewed a convenience sample of 20 patients (half of them fluent in English and half of them LEP), and 15 providers (about one third of them bilingual and two thirds of them needing interpreter services), and observed 10 medical encounters (half of them LC and half of them LD). Using interpretive phenomenology as theoretical framework, we analyzed key themes pertaining to how patient-provider language barriers influence patient adherence to medication and clinical recommendations and providers' clinical decision making in risk factor control (e.g. clinical inertia). We present preliminary results of our investigation.