Research Objective: In 2016, the 1% of Medicare beneficiaries with end‐stage renal disease (ESRD) constituted > 7% of total Medicare spending ($35 billion). To improve the value of care for the ESRD population, the Centers for Medicare and Medicaid Services (CMS) implemented an alternative payment model (APM) for ESRD care, the ESRD Seamless Care Organization (ESCO). This model shares savings with organizations that reduce spending for their ESRD patients below a defined benchmark. This study evaluated the relationship between key organizational, provider, community characteristics, and ESCO performance.
Study Design: We constructed a novel, linked ESCO‐level data set capturing key information for all 37 Wave 2 (2017) ESCOs. After describing the organizational diversity of 2017 ESCOs, we performed bivariate comparisons of high‐ and low‐performing (eg, above versus below median) ESCOs based on three key outcomes: gross savings/losses, composite quality score, and standardized mortality ratio. Subsequently, we estimated generalized logistic regression models of ESCO performance as a function of organizational (number and type of participating facilities, and beneficiary‐to‐provider ratio), ESCO‐participating provider (credential and specialty), and community (percent Hispanic, non‐Hispanic black, percent Medicaid eligible, and median household income) characteristics.
Population Studied: The 37 ESCOs of 2017 were characterized using information from CMS reports, the National Plan and Provider Enumeration System National Provider Identification registry, and the Area Health Resource File. Combined, these ESCOs represented 40 162 Medicare beneficiaries.
Principal Findings: ESCO performance on the three outcomes was highly varied (ranges: savings/losses, −3.9% to 10.2%; quality, 76.4% to 100.0%; standardized mortality, 0.75 to 1.14), as were the ESCOs’ characteristics. Bivariate analysis showed that ESCOs with above (vs below) median savings had more aligned physicians (57.8 vs 28.6, P = .06), fewer dialysis facilities (8.66 vs 17.2, P = .07), a smaller non‐Hispanic black population (14.1% vs 21.9%, P = .06), and higher median household income ($55 532 vs $48 952, P < .001). Facilities reporting a composite quality score of 100% (vs <100%) had fewer aligned practices (22.3 vs 42.6, P = .05) and smaller non‐Hispanic black (16.3% vs 21.3%, P = .06) and Medicaid‐eligible (6.47% vs 8.9%, P = .14) populations. Low standardized mortality ratio (vs above median) was associated with higher median household income ($58 162 vs $45 810, P < .001). Logistic regression model results were broadly consistent with these findings, though small sample size prevented finding statistically significant estimates.
Conclusions: ESCOs are highly diverse with respect to organizational composition and community characteristics; this variation partially explained the observed variation in ESCO performance. Among 2017 ESCOs, larger ESCOs (eg, with more affiliated physicians) and those in higher SES communities performed better with respect to savings and mortality. Quality measure performance was greater among smaller ESCOs, perhaps because of the difficulty of managing quality measure reporting efforts among larger physician groups.
Implications for Policy or Practice: Drawing on data from the only ESRD‐specific APM effective to date, we offer the first evidence of the impact of organizational composition and social disparities on ESCO performance. This study provides crucial evidence that will inform policy deliberations about the design and implementation of APMs in specialty care, and the organizational decisions of provider groups considering whether to participate.