Spillover Effects of the Comprehensive Care for Joint Replacement Program Among Non‐Medicare Patients

Research Objective Through the Comprehensive Care for Joint Replacement (CJR) Model, mandatory bundled payments for lower extremity joint replacement (LEJR) have produced 3‐4% episode cost savings with stable quality. Though CJR targets Medicare fee‐for‐service patients, it may prompt broad care delivery changes that “spill over” to other, non‐Medicare fee‐for‐service patients. Prior work demonstrating spillovers in CJR only evaluated utilization outcomes among a subset of hospitals, thereby compromising the program’s randomized design and failing to evaluate program‐wide utilization and spending spillovers. Therefore, we used more comprehensive data and an approach that preserved CJR’s randomized design to examine the impact of program participation on spillovers for commercially insured and Medicare Advantage patients. Study Design We used 2011‐2017 claims data from the Health Care Cost Institute (full claims of 25% of all insured US individuals younger than 65) to construct LEJR episodes for commercially insured and Medicare Advantage patients, and data from Medicare to identify markets (CJR markets) mandated to participate in CJR from April 2016 to December 2017. Leveraging the market‐level randomized design of CJR, we used hierarchical multivariable generalized linear models—adjusted for patient and market factors and incorporating time fixed effects and hospital and market random effects—to evaluate changes in LEJR episode spending, discharge to institutional postacute care, and quality outcomes (LEJR complication rate, readmissions, emergency department visits) in CJR versus non‐CJR markets. Additionally, we conducted stratified analyses by evaluating the association between CJR participation and outcomes for Medicare Advantage patients and commercially insured patients separately. To evaluate the robustness of our findings, we conducted sensitivity analyses that incorporated low volume CJR hospitals, hospital factors, and LEJR episodes with low index hospitalization spending. Statistical tests were 2‐tailed and significant at α=0.05. Population Studied 418,016 Medicare Advantage or commercially insured patients receiving LEJR in 75 CJR or 121 non‐CJR markets. Principal Findings There were small differences in patient characteristics, including prior health care utilization, between CJR and non‐CJR markets. In adjusted analysis, patients who underwent joint replacement in CJR and non‐CJR markets did not differ with respect to episode spending (difference of ‐$157, 95% CI ‐$1,043 to $728, P  = .73) or probability of discharge to institutional postacute care (difference of ‐1.1%, 95% CI ‐3.2% to 1.0 %, P  = .31). Patients in CJR and non‐CJR markets also did not differ with respect to changes in quality outcomes. Stratified and sensitivity analyses yielded similar findings. Conclusions There was a lack of evidence supporting spillovers from CJR to patients insured through private health plans. Implications for Policy or Practice Unlike voluntary programs, for which spillovers have been demonstrated, mandatory reforms may fail to create strong enough incentives for participants to enact broad care redesign and generate spillovers. Additionally, alternatives to the market‐level randomization used in CJR may be needed to address smaller sample sizes and increase the ability to detect meaningful spillovers. Both issues pose notable implications as policy makers consider whether and how to implement mandatory versus voluntary reforms in the future. Finally, this analysis also underscores the need to go beyond subgroup analyses and conduct program‐wide spillover evaluations, which generate the most policy‐ and practice‐relevant results. Primary Funding Source Leonard Davis Institute of Health Economics.