Explore the scholarly record.

Abstract

Abstract Background. Arthritis affects one in three Chinese adults aged 45 years and older and is a leading cause of functional decline. Whether arthritis independently predicts downstream adverse events (falls, hospitalization, and out-of-pocket [OOP] medical spending) over the long term, and whether prevalent and incident arthritis carry different risks, remains incompletely characterized. Methods. We analyzed five waves (2011, 2013, 2015, 2018, 2020) of the China Health and Retirement Longitudinal Study (CHARLS). Adults aged 45 years and older free of falls or overnight hospitalization at baseline were followed. Arthritis status was carried forward across waves and treated as a time-varying exposure. Falls (Waves 2–4) and overnight hospitalization (Waves 2–5) were modeled with time-dependent Cox regression; OOP spending was modeled with pooled logistic and two-part GEE. Three nested models adjusted for age, sex, residence, education (Model 2), and additionally for BMI, depression, smoking, drinking, comorbidity count, and ADL disability (Model 3). Pre-specified analyses stratified by sex, separated prevalent versus incident arthritis, applied a one-wave-lagged exposure, and addressed attrition bias. Results. Among 10,515 fall-cohort participants (Model 3, 25,045 person-intervals, 56,978 person-years, 3,349 incident falls), arthritis was associated with a 42% higher fall risk (HR 1.42, 95% CI 1.32–1.52; P = 1.3 x 10 –23 ). In 11,507 hospitalization-cohort participants (34,399 intervals, 76,634 person-years, 4,416 hospitalizations), the HR was 1.18 (1.11–1.25; P = 5.5 x 10  − 8 ). In repeated-interval GEE on 12,532 persons, arthritis was associated with any OOP spending (OR 1.33, 95% CI 1.25–1.42; P = 4.2 x 10 –19 ; absolute rates 29.5% vs 21.4%) but not with the amount of spending among spenders. Both prevalent (HR 1.44) and incident (HR 1.23) arthritis predicted falls; associations were consistent across sex (interaction P = 0.25), were unchanged under a one-wave exposure lag (HR 1.38), and persisted after excluding early dropouts. Cognitive decline ( > = 3-point, >= 1-SD, memory, or executive) and hip fracture yielded complementary patterns: hip fracture HR 1.20 (1.05–1.37), but all four cognitive-decline definitions were null (HR approximately 1.00). Conclusions. Arthritis, whether prevalent or incident, is an independent, persistent predictor of falls, hospitalization, and increased likelihood of any OOP medical spending in Chinese middle-aged and older adults. The absence of an arthritis-cognitive-decline association argues against a strong causal pathway. Findings support integrating fall-prevention, comorbidity management, and financial-protection strategies into routine arthritis care.