Development and internal validation of a frailty prediction model in adults aged ≥ 45 years with obesity: evidence from the China Health and Retirement Longitudinal Study (CHARLS)

Abstract Objective Given the increased risk of frailty among middle-aged and older adults with obesity, we aimed to develop and validate a predictive model for incident frailty among adults aged ≥ 45 years with obesity. Methods We used data from the China Health and Retirement Longitudinal Study (CHARLS). Adults aged ≥ 45 years with obesity who were non-frail in 2013 were enrolled as the baseline cohort and followed through 2018. We conducted a complete-case analysis, excluding participants with missing data. Candidate predictors were screened using least absolute shrinkage and selection operator (LASSO) regression and then entered into multivariable logistic regression to build the model. A nomogram was created to facilitate clinical use. Model performance was evaluated by calibration plots, receiver operating characteristic (ROC) curves with the area under the curve (AUC), the Hosmer–Lemeshow goodness-of-fit test, and decision curve analysis (DCA). Results A total of 497 participants completed five-year follow-up, of whom 182 developed frailty. Age, alcohol consumption, diabetes, and grip strength were independently associated with incident frailty and were incorporated into the nomogram. The model showed good discrimination (AUC = 0.744, 95% CI: 0.668–0.800) and acceptable calibration by the Hosmer–Lemeshow test, and DCA suggested meaningful net clinical benefit across a range of threshold probabilities. Conclusions We developed a simple nomogram to predict 5-year frailty risk among adults aged ≥ 45 years with obesity. The model showed moderate discrimination and acceptable calibration in internal validation, suggesting potential utility for risk stratification pending external validation.