Association between hospital-physician vertical integration and medication adherence rates

To test the association between vertical integration of primary care providers (PCPs) and adherence rates for anti-diabetics, renin angiotensin system antagonists (RASA), and statins.

Medicare Part B outpatient fee-for-service claims and Medicare Part D event data from 2014 to 2017.

We estimated difference-in-differences regressions, comparing changes in adherence among patients with PCPs who converted from independent to integrated to changes among patients whose PCPs remained independent or integrated during the study period. To test for heterogenous impacts by patient demographics, we estimated triple difference regressions that included additional interaction terms by comorbidity rates, age group, and race/ethnicity.

We extracted Medicare claims for adults with continuous enrollment in Parts B and D during the study period.

The proportion of patients who had a vertically integrated PCP increased by approximately 23% over the study period. Changes in adherence did not differ significantly between patients based on whether their PCP became integrated (Statins: 0.18, 95% CI -0.13, 0.49; RASA: -0.13, 95% CI -0.46, 0.19; Anti-Diabetics: -0.20, 95% CI -0.78, 0.38). Among patients with PCPs who became integrated, there were significant decreases in adherence for patients who were Black, Asian, Hispanic, or Native American, above 80 years old, and had greater comorbidities for all three classes.

While there were no average changes in adherence following vertical integration of PCPs, health equity worsened, with significant declines in adherence for Black, Asian, Hispanic, and Native American patients, patients over 80 years old, and patients with greater comorbidities. These findings suggest that integration may reduce clinicians' incentives to compete based on the quality of care delivered. Given the price increases associated with integration, integration may be a net welfare loss.