Factors Associated with Reduced Medicare Spending in the Accountable Care Organization (ACO) Investment Model

RESEARCH OBJECTIVE: The ACO Investment Model (AIM) was designed to encourage providers in rural or underserved areas to participate in the Medicare Shared Savings Program (SSP) by providing up‐front financial support for care transformation. In the second performance year, there were 41 ACOs that continued participation in AIM, with 423,434 beneficiaries assigned them. Our research objective was to determine whether the significant reductions in total Medicare spending we found in the first performance year were sustained in the second year, and whether specific ACO or market characteristics were associated with spending in both performance years. STUDY DESIGN: We utilized a difference‐in‐differences evaluation design to estimate differences in total spending between Medicare fee‐for‐service (FFS) beneficiaries attributed to AIM ACOs and comparison beneficiaries, before versus after the start of AIM. The comparison group included beneficiaries residing in each AIM ACO’s market who were eligible but not attributed to an ACO. Regression and weighting were used to adjust for demographic characteristics and health status. We stratified first and second performance year results based on ACO and market characteristics, including whether an ACO used a management company; partnered with a hospital; had fewer than 6,500 enrolled beneficiaries; was in the top quartile of rurality among AIM markets; had a discontinuous market; or had total Medicare spending at baseline that was above the median among all AIM markets. POPULATION STUDIED: Medicare FFS beneficiaries attributed to ACOs in AIM, which targeted ACO formation in rural/underserved areas in 2016 and 2017. PRINCIPAL FINDINGS: Across all 41 AIM ACOs, we estimated a significant reduction in total spending of $36.94 per beneficiary per month (PBPM) in the second performance year—a reduction of 3.5% and similar to findings from the first performance year. There were no significant differences in AIM impacts based on any of our six stratifications, but results across the two performance years suggest that greater reductions were achieved by ACOs working with a management company. CONCLUSIONS: ACOs participating in AIM sustained reductions in total Medicare spending across the first two performance years and across diverse ACO and market characteristics. IMPLICATIONS FOR POLICY OR PRACTICE: Results from the second performance year strengthen the evidence that ACOs can help reduce Medicare spending in rural or underserved areas. Among AIM markets, variations in market‐level baseline spending, market dispersion, and high rurality were not barriers to achieving spending reductions. ACOs locating in rural or underserved areas may benefit through partnership with a management company. PRIMARY FUNDING SOURCE: Centers for Medicare and Medicaid Services.