Cognitive Impairment Risk in Middle-Aged and Older Adults with Rheumatoid Arthritis: Development and Validation of a China Health and Retirement Longitudinal Study–Based Prediction Model
TL;DR
A nomogram based on 5 easily obtained variables was developed to help identify RA patients at high risk of cognitive impairment and showed promise in supporting targeted screening and early management in community settings.
Abstract
Background/Aims: Cognitive impairment is frequent in rheumatoid arthritis (RA) and worsens function and quality of life; yet, simple tools to identify high-risk community-dwelling patients are scarce in China. This study aimed to develop and test a practical tool to predict cognitive impairment risk in adults with RA. Materials and Methods: A cross-sectional analysis was performed using the China Health and Retirement Longitudinal Study (CHARLS) database for participants aged ≥45 years with self-reported physician-diagnosed RA and complete cognitive data. Cognitive impairment was defined as a total cognitive score <11 based on episodic memory and mental status tests. Factors included age, education, lifestyle, disease history, mood, and blood tests. The sample was randomly split into training and validation groups. Least absolute shrinkage and selection operator (LASSO) and logistic regression were used to build the prediction model and nomogram and checked its accuracy and usefulness. Results: Among 1280 community-dwelling adults with RA in the 2015 CHARLS wave, 41.6% met criteria for cognitive impairment. The participants were randomly divided into a training group and a validation group. The LASSO regression identified 5 factors: age, marital status, education, sleep duration, and social activity, which were used to develop the nomogram. Older age, not being married, lower education, and short or long sleep were linked to higher risks, while participating in social activities was linked to lower risks. The proposed model showed good separation between people with and without cognitive impairment, with area under the curve values above 0.75 in both groups. Conclusion: A nomogram based on 5 easily obtained variables was developed to help identify RA patients at high risk of cognitive impairment and showed promise in supporting targeted screening and early management in community settings. Cite this article as: He X, An G, Zhang S, et al. Cognitive impairment risk in middle-aged and older adults with rheumatoid arthritis: Development and validation of a China Health and Retirement Longitudinal Study– based prediction model. ArchRheumatol. Published online August 3, 2026. doi: 10.5152/ArchRheumatol.2026.25290.