Abstract
IMPORTANCE: Physician organization (PO) affiliation with health systems and its association with health care disparities for dual-eligible Medicare and Medicaid beneficiaries, who face significant barriers to care, remains underexplored. OBJECTIVES: To estimate the association of health system affiliation with disparities in quality of care and health care utilization for dual-eligible beneficiaries relative to non-dual-eligible Medicare beneficiaries and to decompose these associations into within-PO and between-PO components. DESIGN, SETTING, AND PARTICIPANTS: This cohort study used data from 2013 to 2019 on primary care POs and beneficiaries enrolled in traditional Medicare. Statistical analysis was performed from April 2024 to March 2025. EXPOSURES: Affiliation with health systems, defined as ownership or management relationships, using Medicare Provider Enrollment, Chain, and Ownership System data and Internal Revenue Service Form 990 data. MAIN OUTCOMES AND MEASURES: Eight quality measures assessing receipt of preventive services, chronic condition management, medication adherence, and care coordination and continuity and 5 measures of health care utilization in different settings and with different physician specialists. Linear mixed-effects models were used to estimate changes in disparities before and after affiliation for primary care POs between 2013 and 2019. RESULTS: A total of 5005 primary care POs and more than 5.6 million Medicare beneficiaries (mean [SD] age, 75.5 [7.5] years; 58.4% women) were analyzed, including approximately 700 000 dual-eligible beneficiaries. Affiliation with health systems was associated with widened disparities in diabetic eye examinations (3.5 percentage point larger relative reduction for dual-eligible beneficiaries compared with non-dual-eligible beneficiaries) and follow-up visits after acute events (3.5 percentage point larger relative reduction). On the other hand, dual-eligible beneficiaries experienced relative improvements in continuity of care with primary care clinicians (1.9 percentage points) and POs (1.4 percentage points) compared with non-dual-eligible beneficiaries, as well as larger relative improvements in statin prescribing (1.8 percentage points), widening preaffiliation differences that favored dual-eligible beneficiaries. Four of these 5 measures were associated with widening within-PO and between-PO disparities. Despite moderate and large preaffiliation disparities in primary care clinician visits and specialist visits, respectively, for dual-eligible beneficiaries relative to non-dual-eligible beneficiaries, disparities in primary care clinician visits widened (21 fewer visits per 100 beneficiaries) while disparities in specialist visits did not change meaningfully. CONCLUSIONS AND RELEVANCE: In this cohort study, health system affiliation by primary care POs was associated with both positive and negative associations with disparities for dual-eligible beneficiaries and did not reduce the largest preaffiliation disparities. Health systems must strengthen their care delivery models to expand access to specialists and avoid exacerbating disparities in follow-up care. Health systems could identify factors associated with improved care at high-percentage dual POs for replication.