Early renewal of accountable care organization agreements to reset benchmarks in the Medicare Shared Savings Program: a preregistered retrospective cohort study
Abstract Background Medicare Shared Savings Program (MSSP) benchmarks are reset at each new agreement from an accountable care organization’s (ACO’s) recent spending, adjusted toward its region. Ending an agreement early reset the benchmark sooner and was a BASIC-track ACO’s only way to reach the ENHANCED track’s 75% share of savings without leaving the program or waiting out its agreement. No peer-reviewed study has examined why. This study asked which ACOs renewed early, weighing a reset motive (a higher benchmark) against a sharing-rate motive, and whether spending then changed. Methods This preregistered retrospective cohort study included the 334 MSSP ACOs whose agreements began July 1, 2019 or January 1, 2020, using public Medicare files (2016–2024). The plan and code were registered before outcome data were read. The reset benchmark was rebuilt from the regulation and validated on 101 other ACOs. 29 ACOs renewing early in January 2022 were compared with 277 stayers on the regional pull (primary), the own-spending component and BASIC risk level, by Bayesian bootstrap, permutation tests and logistic regression. Spending growth in 2022–2024 was compared by difference-in-differences. Results The rebuilt benchmark tracked the published one (ρ = 0.94). Early renewers’ regional pull was 3.4 percentage points higher than stayers’ (95% credible interval [CrI], 1.7 to 5.4; P <.001), but they also beat their benchmarks by more (own-spending component, − 2.1 points; 95% CrI, − 3.8 to − 0.5), and the total expected gain did not differ (− 0.5 points; 95% CrI, − 2.0 to 0.8). Post hoc, early renewers’ published 2022 benchmarks averaged 1.3% below 2021. Among 206 BASIC-track ACOs, a higher regional pull and BASIC level were associated with early renewal; 26 of all 27 BASIC-track early renewers moved to ENHANCED. Spending growth did not differ detectably (− 0.4%; 95% CrI, − 4.0% to 2.5%); a preregistered pre-trend condition failed. Conclusions Early renewers were already beating their benchmarks; for almost all, renewal was the route to a 75% sharing rate, and the reset did not on average raise benchmarks. This fits the sharing-rate motive better, though the design cannot separate the motives; the rules changed in 2024. Trial registration: OSF Registries, doi:10.17605/OSF.IO/C53W9. Registered 26 October 2026.
Authors
- Samuel Thomas Harrold (ORCID: https://orcid.org/0009-0008-4596-6921)
Publication Details
- Journal
- Research Square
- Published
- 2026-11-16
- DOI
- https://doi.org/10.21203/rs.3.rs-11565926/v1
- Primary Topic
- Healthcare Policy and Management
- Type
- preprint