Accountable care organizations and Medicare payments for residents with ADRD in disadvantaged neighborhoods

Abstract

Introduction: Accountable care organizations (ACOs) are well positioned to promote care coordination. However, robust evidence of ACOs' impact on Medicare payments for residents with Alzheimer's disease and related dementias (ADRD) in disadvantaged neighborhoods remains limited.

Methods: Using a 2016 to 2020 longitudinal dataset, we examined the effects of ACO enrollment on Medicare payments for people newly diagnosed with ADRD, focusing on the neighborhood Social Vulnerability Index (SVI) and its subcategories. Multivariable generalized estimating equation (GEE) models were applied.

Results: ACO enrollment was associated with significantly reduced total payments across all SVI subcategories. The highest cost savings were observed among ADRD patients living in neighborhoods with high proportions of racial and ethnic minorities. Results also showed that higher quality ACOs were associated with lower total payments.

Discussion: ACOs have a great potential to save health-care costs for beneficiaries with ADRD living in socially vulnerable neighborhoods, particularly for those residing in areas with higher proportions of racial and ethnic minority populations.

Highlights: Accountable care organizations (ACOs) reduced Medicare payments for Alzheimer's disease and related dementias across neighborhood disadvantage levels. The cost reductions varied by specific indicators of social vulnerability. Highest cost savings were found among residents living with high proportion of racial/ethnic minorities. Cost savings were the greatest among the highest quality ACOs.

Keywords: Alzheimer's disease and related dementias; accountable care organization; cost disparities; neighborhood disadvantage; social vulnerability.

FIGURE 1

Trends of unadjusted total Medicare payments of newly diagnosed beneficiaries with ADRD by SVI and ACO enrollment status. Unadjusted total Medicare payments of newly diagnosed beneficiaries with ADRD were summarized by beneficiaries’ ACO enrollment status and residence in low or high SVI neighborhoods. No ACO represents beneficiaries who were not enrolled in ACO for a full year and those who never enrolled in ACO. ACO represents ACO enrollees for a full year. Low SVI represents residents in neighborhoods with < 50th percentile of SVI, while high SVI represents those in neighborhoods with ≥ 50th percentile SVI. Unadjusted dollar amounts are available in Table S5 in supporting information. ACO, accountable care organization; ADRD, Alzheimer's disease and related dementias; SVI, Social Vulnerability Index.

FIGURE 2

Total Medicare payments of newly diagnosed patients with ADRD by SVI and ACO enrollment status in different themes of neighborhood social vulnerabilities. Adjusted total Medicare payments were estimated using a GEE model in individual SVI themes to present variations in cost patterns by ACO enrollment status and residence in neighborhood vulnerability depending on specific measures of social vulnerability. The GEE results are available in Table S7 in supporting information and estimated US dollar amounts are available in Table S8 in supporting information. ACO enrollment status, individual SVI theme, interaction term between ACO and individual SVI, race/ethnicity, sex, age, insurance coverage, primary care provider shortage area, four common coexisting comorbidities, and year were included as covariates. No ACO includes beneficiaries who were not enrolled in ACO for a full year and those never enrolled in ACO. ACO represents ACO enrollees for a full year. Low SVI represents residents in neighborhoods with < 50th percentile of SVI, while high SVI represents those in neighborhoods with ≥ 50th percentile SVI. ACO, accountable care organization; ADRD, Alzheimer's disease and related dementias; GEE, generalized estimating equation; SVI, Social Vulnerability Index.

FIGURE 3

Total Medicare payments among continuous 5‐year ACO enrollees by ACO‐specific quality. Adjusted total Medicare payments among ACO enrollees continuously enrolled for 5 years by ACO‐specific quality are presented. The MSSP ACO overall quality scores from CMS were categorized into four quantile groups, ranging from the lowest (i.e., < 25th percentile) to the highest (i.e., ≥ 75th percentile) quality. Never ACO represents beneficiaries who were never enrolled in ACO. The GEE results used to estimate adjusted total Medicare payments among continuous 5‐year ACO enrollees are available in Table S10 and estimated US dollar amounts are available in Table S11. ACO, accountable care organization; CMS, Centers for Medicare and Medicaid Services; GEE, generalized estimating equation; MSSP, Medicare Shared Saving Program.