Abstract 4126000: The Association of Cardiologist Vertical Integration With Care Quality, Patient Outcomes, and Utilization

Introduction: U.S. healthcare has experienced substantial growth in vertical integration (physicians employed by hospitals), but little is known about its effect on patient care. Aims: To assess the association between cardiologist vertical integration and patient outcomes, care quality, and utilization among patients hospitalized with incident acute myocardial infarction (AMI) or heart failure (HF). Methods: We used a sample of all Medicare fee-for-service beneficiaries hospitalized with incident AMI and a 75% sample of beneficiaries hospitalized with incident HF between 2008-2019. We identified the accountable cardiologists that cared for these patients and determined their integration status using tax identification numbers. We used difference-in-differences methods to compare outcomes for patients treated by integrated cardiologists after switching from non-integrated to integrated practice, to outcomes for patients treated by cardiologists who remained non-integrated. Patient outcomes were in-hospital mortality, 30-day mortality, and 30-day readmission. Quality measures were the proportion of patients that received 1) a guideline-recommended test to assess cardiac function and 2) a 30-day follow-up clinic visit. Utilization measures were length of stay and the proportion of AMI patients receiving percutaneous coronary intervention. Models were adjusted for time trends and patient, hospital, and cardiologist characteristics. Results: The proportion of U.S. cardiologists employed by hospitals increased from 26% in 2008 to 61% in 2018. We identified 186,052 AMI patients and 259,849 HF patients cared for cardiologists who switched to integrated practice and 168,052 AMI patients and 245,769 HF patients cared for by non-integrated cardiologists. Patients were similar in age, sex, race/ethnicity, and comorbidities. We found no significant differences in clinical outcomes (e.g., adjusted difference in 30-day mortality 0.03% [95% CI = -0.39%, 0.45%] for AMI and -0.05% [-0.37%, 0.27%] for HF patients); no differences in most quality metrics, and no differences in utilization between patients treated by integrated versus non-integrated cardiologists. Conclusions: We found minimal evidence that cardiologist employment by hospitals improves care quality or outcomes. Regulation of vertical integration should focus on other effects of integration such as higher prices paid by insurers.