Rural–urban disparities in diabetes quality of care with accountable care organization participation

PURPOSE: To evaluate rural-urban disparities over time in the association of ACO participation and diabetes-related quality measures among health clinics. METHODS: We used data from the Wisconsin Collaborative for Healthcare Quality all-patient all-payer electronic health records data system between 2011 and 2018, for patients 18-75 years. Difference-in-differences regression models estimated the association between ACO participation and eight diabetes quality measures among populations in rural and urban areas, separately. Triple-difference models were also estimated to assess urban-rural disparities. FINDINGS: Considering the two measures used in ACO performance evaluation, patients in ACO clinics were less likely to receive tobacco cessation advice relative to those in non-ACO clinics (rural: marginal effect estimate (MEE) = -0.025, p = 0.033; urban: MEE = -0.231, p < 0.001). The triple difference across rurality was not statistically significant (MEE = -0.007 p = 0.56). For the remaining six ACO-non-incentivized measures, rural patients at ACO clinics performed better relative to their non-ACO counterparts on kidney function monitored, and diabetes all-or-none optimal testing and control. CONCLUSIONS: ACO participation appeared to be more favorable for rural versus urban patients with diabetes. ACOs have potential to contribute to reducing existing rural-urban disparities in diabetes process measures.