Variations in VA and Medicare Use Among Veterans With Diabetes: Impacts on Ambulatory Care Sensitive Conditions Hospitalizations for 2008, 2009, and 2010.

ABSTRACT

INTRODUCTION: VA and Medicare use among older Veterans has been considered fragmented care, however, it may represent access to needed care. METHODS: The population studied were Veterans with diabetes, age 66 years and older, dually enrolled in VA and Medicare. DATA SOURCE/STUDY SETTING: We conducted a dynamic retrospective cohort study with 2008, 2009, and 2010 as the outcome years (Ambulatory Care Sensitive Conditions Hospitalization (ACSC-H) or not). We analyzed administrative data to identify comorbidities; ambulatory care utilization to identify variations in use before hospitalization. We linked 2007 primary care (PC) survey data to assess if organizational factors were associated with ACSC-H. MEASURES AND ANALYSIS: We identified ACSC-Hs using a validated definition. We categorized VA/Medicare use as single system; dual system supplemental specialty care use; or primary care use. Using hierarchical logistic regression models, we tested for associations between VA/Medicare use, organizational characteristics, and ACSC-H controlling for patient-level, organizational-level, and area-level characteristics. RESULTS: Our analytic population was comprised of 210,726 Medicare-eligible Veterans; more than one quarter had an ACSC-H. We found that single system users had higher odds of ACSC-H compared with dual system specialty supplemental care use (odds ratio, 1.14; 95% confidence interval, 1.09-1.20), and no significant difference between dual-system users. Veterans obtaining care at sites where PC leaders reported greater autonomy (eg, authority over personnel issues) had lower odds of ACSC-H (odds ratio, 0.74; 95% confidence interval, 0.59-0.92). DISCUSSION: Our findings suggest that earlier assumptions about VA/Medicare use should be weighed against the possibility that neither VA nor Medicare may address complex Veterans' health needs. Greater PC leader autonomy may allow for tailoring of care to match local clinical contexts.