(199) The Centers for Medicare & Medicaid Services' definition of social risk may be missing those in need

Background: : The Centers for Medicare and Medicaid Services (CMS) seeks to utilize risk factors (RF) for poor health outcomes in several proposals affecting Medicare Advantage (MA) Organizations, such as the Excellent Health Outcomes for All reward. Utilizing membership data from a major health plan, we noticed about half of MA membership met CMS’ definition of RF. This high prevalence begs the question of if this definition of RF is precisely able to identify those with health-related social needs (HRSNs), the targeted, more immediate requirements arising from broader concepts such as social risk, and additionally, calls for clarification of the constructs CMS desires to assess.

Objective: : This research set out to elucidate if health plan members identified with having RF as defined by CMS had unmet HRSNs, and if there were differences in HRSNs among MA membership identified with RF by CMS compared to those who weren’t.

Methods: : Data flagging CMS RF were extracted for members of a major health plan from CMS member restated monthly files as a binary variable. These were joined to member-level assessment data, including Accountable Health Communities (AHC) HRSN Screening Tool, and Protocol for Responding to & Assessing Patients’ Assets, Risks, and Experiences (PRAPARE), as well as a social risk survey (SRS) deployed in-house across one midwestern and one southern state providing information on nine HRSNs. Measures were transformed into binary variables to indicate presence of social need, and were correlated against the variable denoting CMS RF. The population studied was any of the MA members identified in the CMS member restated monthly files in 2021 or 2022 which were included in at least one of the sources of HRSN assessment.

Results: : The correlation between CMS RF and self-reported HRSN was weak to moderate across all measures, with Pearson correlation coefficients of 0.10 and 0.29 for AHC and PRAPARE, respectively, and ranging from 0.10 to 0.45 among the nine HRSNs measured in the SRS. All correlations except the “internet” HRSN in the SRS were significant at the p< 0.05 level.

Conclusions: : There is limited concordance between CMS RF and HRSNs, as self-reported on member assessments—individuals flagged with RF by CMS may not have attested to self-reported HRSNs, and vice-versa. It is critical for stakeholders across the Medicare sphere to synergize efforts to identify and support individuals in need, and this starts with ensuring concepts such as RF applied in the context of a greater social needs framework rather than in a vacuum.

Sponsorship:: Elevance Health