Effects of Comprehensive Care for Joint Replacement model on racial/ethnic disparities in joint replacement care

Background: Hip and knee replacements (lower extremity joint replacement or LEJR hereafter) improve function and quality of life for people with severe arthritis. However, there have been long-standing racial/ethnic disparities in LEJR care. This study aims to assess the effects of the Comprehensive Care for Joint Replacement (CJR) model on racial/ethnic disparities in LEJR care. Implemented in April 2016, CJR holds hospitals accountable for the cost and quality of care for Medicare patients receiving LEJR during the index hospitalization and 90 days following the hospital discharge. CJR might create strong incentives to provide high-value care for black and Hispanic patients who disproportionately represent patients with low socioeconomic status (SES) and previously received poor-quality care. If so, the CJR model will decrease racial/ethnic disparities in LEJR care.

Study Design: We used 2014-2016 Medicare claims and examined the effect of CJR on LEJR care for white, black, and Hispanic patients between hospitals located in 67 metropolitan statistical areas (MSAs) affected by CJR and those in 104 comparable MSAs not affected by CJR. CJR was implemented in 67 randomly selected MSAs, providing an opportunity to estimate the causal effect of the CJR model on racial/ethnic disparities. There were no statistically significant differences among baseline risk-adjusted outcomes when comparing CJR and non-CJR MSAs, suggesting randomization was effective. Therefore, we used post-CJR data to estimate the effects of CJR on disparities, controlling for patient, hospital, and MSA characteristics unrelated to SES. We also assessed CJR effects for dual-eligibles vs non-dual-eligibles among white, black, and Hispanic patients.

Results: Consistent with prior literature, we found substantial racial disparities in most outcomes before the announcement of CJR. During the first year of CJR, black-white and Hispanic-white disparities remained the same in most outcomes. However, CJR was associated with a $3,455 decrease in expenditures for institutional post-acute care among Hispanic patients (35% decrease from the average of Hispanics in non-CJR MSAs), and an $896 decrease among white patients (14% decrease from the average of whites in non-CJR MSAs), thereby narrowing the Hispanic-white difference in post-acute care expenditures from $3,097 to $538. The decrease in institutional post-acute care expenditures among white patients was mostly driven by a reduction in expenditures among dual-eligible patients, but the decreases did not differ by dual-eligible status within Hispanic patients.

Conclusion: Hispanic-white disparities in expenditures for institutional post-acute care decreased under the CJR model, suggesting that CJR increased the value of care for Hispanic patients, eventually decreasing Hispanic-white disparities. However, black-white and Hispanic-white disparities remained the same for all other outcomes related to quality of LEJR care (readmission rates, emergency department visit rates, and prompt start of physical therapy care after the hospital discharge). Altogether, our findings suggest that despite its potential, CJR did not improve black-white and Hispanic-white disparities in quality of care during its first implementation year, but decreased Hispanic-white disparities in Medicare expenditures for institutional post-acute care.