Development and validation of a risk prediction model for disability in older adults with cardiometabolic multimorbidity

Cardiometabolic multimorbidity (CMM) is a prevalent multimorbidity pattern among older adults, significantly associated with adverse outcomes, including disability. However, the risk of disability varies considerably among individuals with CMM, and reliance solely on associated factors is inadequate for individualized risk assessment. This study aimed to develop and validate a risk prediction model for disability in older Chinese adults with CMM, facilitating the early identification of high-risk individuals and stratified management. This secondary analysis utilized data from the China Health and Retirement Longitudinal Study (CHARLS). CMM was defined as the coexistence of at least two cardiometabolic diseases, including hypertension, diabetes, heart disease, and stroke. A total of 711 adults aged ≥ 65 years with CMM and no baseline disability in 2015 were included, randomly divided into a training cohort ( n = 497) and a validation cohort ( n = 214) at a 7:3 ratio. Feature selection employed least absolute shrinkage and selection operator (LASSO) regression and the Boruta algorithm. The intersection of variables identified by both methods was used in multivariable logistic regression to develop a nomogram model for disability risk. Discrimination was assessed via the area under the receiver operating characteristic curve, calibration was evaluated using calibration plots and the Hosmer-Lemeshow test, and clinical utility was examined through decision curve analysis (DCA). Among the 711 older adults with CMM, the incidence of disability at follow-up in 2018 was 40.08%. The final model retained five predictors: age, arthritis, education level, cognitive impairment, and grip strength. Age, arthritis, and cognitive impairment emerged as risk factors, while education at the high school level or above and greater grip strength were identified as protective factors. AUCs were 0.712 in the training cohort and 0.724 in the validation cohort, reflecting moderate discrimination and stability. Hosmer-Lemeshow tests yielded nonsignificant results in both cohorts ( P > 0.05), and DCA indicated a potential net clinical benefit across a broad range of threshold probabilities. The developed disability risk prediction model for older Chinese adults with CMM demonstrated moderate discrimination, acceptable calibration, and potential clinical utility, serving as an effective tool for identifying individuals at heightened risk of disability.