Use of Claims Data to Screen for Functional Limitations Among Medicare Beneficiaries.

Abstract

The only functional impairment data available for all Medicare beneficiaries identify those who qualified for Medicare before age 65 years due to work-related disability. This administrative measure omits impairments arising after age 64 years or affecting nonwork-related activities. Surveys may offer more accurate assessments, but survey data on functional limitations are not routinely available. Claims data may provide a practical, scalable alternative. To describe the development of a survey-based index of functional limitations and a new claims-based model for predicting limitations. Data from respondents to the 2024 Medicare fee-for-service (FFS) Consumer Assessment of Healthcare Providers and Systems (CAHPS) survey were used to create the FFS CAHPS functional limitations index (FCFLI), a measure of the extent to which functional limitations affect overall health and the benchmark for the study's claims-based measure. The FCFLI scores of respondents were linked to Medicare FFS claims from April 2023 to March 2024 to develop a claims-based predictive model (FCFLI-claims). The final model was applied to all FFS beneficiaries enrolled in Medicare Parts A and B during the same period. The FCFLI was derived from beneficiaries' self-rated health and ability to perform 6 basic activities of daily living and 1 instrumental activities of daily living (collectively, I/ADLs). Nine sets of claims-based indicators of functional impairment were used to predict FCFLI scores. Data were collected from 67 596 respondents (53.4% female; mean [SD] age, 75 [8] years). In the FCFLI model, all I/ADL indicators were negatively associated with self-rated health, with the largest effect estimates for difficulty walking (-13.9 points; 95% CI, -14.6 to -13.2 points) and running errands (-11.7 points; 95% CI, -12.7 to -10.8 points). In the FCFLI-claims model, 14 claims indicators had relative importance values that were 5% or greater of the maximum value. The model reliably identified beneficiaries with functional limitations (80.4% positive predictive value). Applied to the 2024 Medicare FFS population, approximately 12% were identified as likely to have functional limitations. Nearly 63% (approximately 2 450 000) of those identified by the algorithm were not identified by Centers for Medicare & Medicaid Services administrative measure of disability, which focuses on limitations that develop before age 65 years. In this cross-sectional study, FCFLI-claims identified beneficiaries likely to have functional limitations who could be prioritized for a survey assessment to verify the algorithm's results. Integrating this information with entitlement data could enhance Centers for Medicare & Medicaid Services' monitoring of functional status, especially among age-eligible beneficiaries.