The relationship between medication possession index and functional prognosis in patients with rheumatoid arthritis: a national longitudinal cohort study from CHARLS

Objectives: To prospectively investigate the association between the Medication Possession Index (MPI) and the risk of incident functional dependence in middle-aged and older rheumatoid arthritis (RA) patients in China. Methods: This national longitudinal cohort study utilized data from the China Health and Retirement Longitudinal Study (CHARLS). A total of 2495 RA patients (mean age 67 · 5 ± 10 · 9 years; 36 · 5% male) without baseline functional dependence were included. MPI (range: 0–1), a composite, cross-sectional measure of pharmacotherapeutic coverage for 12 chronic conditions, was calculated at baseline. Complex survey weights were incorporated in all Cox models using survey::svycoxph(), and missing covariates were addressed via multiple imputation ( m = 10). Participants were categorized by MPI quartiles (Qa:lowest to Qd:highest). The primary outcome was incident functional dependence (assistance needed in activities of daily living/instrumental activities of daily living). Multivariate Cox proportional hazards models and restricted cubic splines (RCS) were used to assess associations and dose–response relationships. Results: Over 8530 person-years (median follow-up 3–4 years), 918 functional dependence events occurred (incidence rate: 10.8/100 person-years). Higher MPI quartiles were associated with higher functional independence survival (log-rank P = 1.3 × 10 –4 ). Compared to Qa, Qd was associated with a 29% reduced risk in the fully adjusted model (HR = 0.71, 95% CI:0.59–0.87; P <0.001), with a significant inverse linear trend ( P -trend<0.001). RCS confirmed a monotonic dose–response relationship ( P -nonlinearity = 0.18). The association was robust across multiple sensitivity analyses. Conclusion: Higher MPI, reflecting better overall treatment coverage across multiple conditions, is independently associated with a reduced risk of functional dependence in middle-aged and older RA patients, showing a monotonic dose–response relationship. MPI shows potential as a pragmatic indicator of overall healthcare engagement for identifying high-risk patients who may benefit from interventions aimed at improving treatment coverage and health management.