Hospital Characteristics Associated With Heterogeneity in Institutional Postacute Care Spending Reductions Under the Comprehensive Care for Joint Replacement Model

Abstract

Prior research concluded that institutional postacute care spending decreased under the Comprehensive Care for Joint Replacement (CJR) model. Less is known about how changes in institutional postacute care spending varied across different types of hospitals. To measure hospital-level heterogeneity in the association of the CJR model with changes in institutional postacute care spending and to identify hospital characteristics associated with this variation. Using 100% Medicare claims data, this cross-sectional study assessed institutional postacute care spending from 2016 to 2017 among US hospitals randomly selected to participate in the CJR model and control group hospitals that were eligible but not selected for the participation in the CJR model. A causal forest was used to estimate the treatment effect of the CJR model conditional on hospital characteristics. Analysis was conducted between October 2019 and October 2021. The unit of analysis was each hospital; the outcome was the average per-episode Medicare spending for institutional postacute care within 90 days after hospital discharge for hip or knee joint replacement. This study included 531 CJR participating hospitals and 658 control group hospitals from 2016 to 2017. The CJR model was associated with a $761 reduction in institutional postacute care spending (95% CI, -$1172 to -$351). The reduction in spending under the CJR model did not vary across conditional on hospital characteristics. Limited evidence was found for greater savings among hospitals with higher pre-CJR spending. However, this finding did not hold for hospitals in the highest quintile of pre-CJR spending. In this cross-sectional study of 1189 hospitals, findings did not show strong evidence for significant heterogeneity in how the CJR model was associated with reductions in institutional postacute care spending across a range of hospital characteristics. Savings were not concentrated in hospitals with specific characteristics, such as hospitals with high-volume joint replacement or hospitals serving less medically or socially complex patients. Findings suggest that the CJR model created opportunities for savings across a spectrum of different hospitals.