A Mixed Methods Approach to Identifying “Best Practices” in Outpatient Diabetes Care in VA

Background: Nearly 25% of patients in the Veterans Healthcare System have diabetes, a major cause of morbidity and mortality. This project aims to identify "best practices" that may account for inter-facility variations in performance and help to identify barriers/facilitators which can be shared nationally and inform policy related to diabetes care delivery. Methods:This observation study used a mixed-methods approach and a national VA database to identify high to low performing sites for 2008. A total of 140 facilities and 582 community based outpatient clinics (CBOC) were included. The measure was the percentage of patients with an A1c> 9% (facility range: <5 to >30) Patient populations (insulin users, those with comorbid mental illness) were analyzed separately. Semi-structured telephone interviews were conducted with selected sites. Results: Site performance varied across patient populations Low performing sites were characterized by lack of interdisciplinary team work, insufficient support staff, inadequate planned care. Also, better performing sites were characterized by shared responsibilities such as planned visits/care e.g. telehealth and group medical appointments, time for non face to face care. Middle performing sites were able to identify appropriate interventions (for example, best practices observed in other sites) but were unable to operationalize them. Conclusion: This study illustrates notable differences in site characteristics that are suitable for quality improvement activities These findings lend themselves to local, regional and national level improvement initiatives.