Redesigning Medicaid Frailty Algorithms:Improved Identification of Medically Frail Adults Under Community Engagement

Abstract Between 2018 and 2024, fourteen US states sought or obtained Section 1115 waivers to condition Medicaid expansion coverage on community engagement requirements, with medically frail exemptions determined from claims-based administrative data; the One Big Beautiful Bill Act of 2025 (Public Law 119-21) codified these requirements nationally, with state implementation due by January 2027. Using American Community Survey Public Use Microdata Sample data from 75,043 Medicaid-enrolled adults aged 19–64 across 17 states, we simulated frailty identification under each state’s existing algorithm via a three-channel Monte Carlo microsimulation incorporating algorithm design, claims visibility, and documentation burden. Existing algorithms identified a mean of 31.4% of adults with functional disability as medically frail (range: 14.3% Florida, Arizona to 45.4% New York). An evidence-based redesigned algorithm incorporating expanded diagnostic criteria, health information exchange integration, ex parte determination, and elimination of physician certification requirements increased mean identification to 45.6% (+14.3 percentage points), with gains across all 17 states. In multi-dimensional equity evaluation, the redesigned algorithm narrowed the American Indian/Alaska Native–White sensitivity gap by 46% (from 11.6 to 6.3 percentage points) and the Black-White gap by 10% (from 12.8 to 11.5 percentage points); within-race rural-urban sensitivity differences of 3–5 percentage points persisted under both algorithms. Adoption of the redesigned algorithm would identify an estimated 3.8 million additional medically frail adults and avert approximately 253,000 coverage losses under full implementation. These findings support minimum algorithmic design standards for state frailty determination systems.