Importance: The incentive structure of accountable care organizations (ACOs) may lead to participating physician groups selecting fewer vulnerable patients.
Objective: To test for changes in the percentage of racial minority patients and patients with low socioeconomic status cared for by physician groups after joining the ACO.
Design, setting, and participants: This retrospective cohort consisted of a 15% random sample of Medicare fee-for-service beneficiaries attributed to physician groups from 2010 to 2016. Medicare Shared Savings Program (MSSP) participation was determined using ACO files. Analyses were conducted between January 1, 2019, and February 25, 2020.
Exposures: Using linear probability models, we conducted difference-in-differences analyses based on the year a physician group joined an ACO to estimate changes in vulnerable patients within ACO-participating groups compared with nonparticipating groups.
Main outcomes and measures: Whether the patient was black, was dually enrolled in Medicare and Medicaid, and poverty and unemployment rates of the patient's zip code.
Results: In a cohort of 76 717 physician groups caring for 7 307 130 patients, 16.1% of groups caring for 27.8% of patients participated in an MSSP ACO. Using 2010 characteristics, patients attributed to ACOs from 2012 to 2016, compared with those who were not, were less likely to be black (8.0% [n = 81 698] vs 9.3% [n = 270 924]) or dually enrolled in Medicare and Medicaid (12.8% [n = 130 957] vs 18.2% [n = 528 685]), and lived in zip codes with lower poverty rates (13.8% vs 15.5%); unemployment rates were similar (8.0% vs 8.5%). In the difference-in-differences analysis, there was no statistically significant change associated with ACO participation in the proportions of vulnerable patients attributed to ACO-participating groups compared with nonparticipating groups. After joining an ACO, ACO-participating groups had 0.0 percentage points change (95% CI, -0.1 to 0.1 percentage points; P = .59) for black patients, -0.1 percentage points (95% CI, -0.2 to 0.1 percentage points; P = .32) for patients dually enrolled in Medicare and Medicaid, 0.2 percentage points (95% CI, -3.5 to 4.0 percentage points; P = .91) in poverty rates, and -0.4 percentage points (95% CI, -2.0 to 1.2 percentage points; P = .62) in unemployment rates.
Conclusions and relevance: In this cohort study, there were no changes in the proportions of vulnerable patients cared for by ACO-participating physician groups after joining an ACO compared with changes among nonparticipating groups.