Objective To clearly state the identification of key correlates of depression and construction of a cross-sectional association-based nomogram for individualized current risk assessment, and clarify the role of each predictor via SHapley Additive exPlanations (SHAP) analysis. Methods This cross-sectional study included 3,701 participants with Activities of Daily Living dysfunction from the China Health and Retirement Longitudinal Survey Wave 3. Data were split into training (70%, n=2590) and testing (30%, n=1111) sets. Least Absolute Shrinkage and Selection Operator regression screened predictors from 79 variables, multivariate logistic regression built the nomogram, and model performance was validated using Receiver Operating Characteristic curves, Area Under the Curve (AUC), calibration plots, and Decision Curve Analysis. SHAP analysis interpreted predictor contributions. Results Ten key predictors were identified: age, pain, disability, fall history, right grip strength, waist circumference, self-rated health, sleep duration, social activity level, and memory problems. The nomogram showed acceptable discriminatory ability (AUC=0.757 in training set, 0.751 in testing set), good calibration, and clinical utility. Pain and disability were top risk factors, while right grip strength and self-rated health were protective. Conclusion The validated nomogram integrates multidimensional predictors to enable individualized depression risk assessment in this population, supporting early screening and targeted interventions to improve mental health outcomes.
Peng et al. (Wed,) studied this question.