Time-stratified modeling of cognitive impairment risk in rural aging populations: Nomogram development (2011) and external validation (2013) using CHARLS.

BACKGROUND

With China's aging population, cognitive impairment has become a pressing public health issue. Rural older adults face disproportionately higher risks, yet remain underrepresented in predictive modeling studies. This study aimed to develop and externally validate a nomogram to estimate cognitive impairment risk among rural older adults in China.

METHODS

Data were obtained from 2228 rural participants aged ≥60 years in the 2011 China Health and Retirement Longitudinal Study (CHARLS), randomly assigned to training and internal validation cohorts. An additional 1854 rural participants from the 2013 CHARLS wave served as an external validation set. Feature selection was conducted using the least absolute shrinkage and selection operator (LASSO), followed by multivariable logistic regression to identify independent predictors. A nomogram was constructed, with model performance evaluated through ROC curves, calibration plots, and decision curve analysis (DCA).

RESULTS

Six predictors-age, education, alcohol consumption, systolic blood pressure, handgrip strength, and depressive symptoms-were included in the final nomogram. The model achieved AUCs of 0.849 (training), 0.852 (internal validation), and 0.806 (external validation), indicating strong discriminative ability. Calibration showed good agreement between predicted and observed outcomes. DCA demonstrated favorable clinical utility across all cohorts.

CONCLUSION

The nomogram exhibited strong predictive performance and generalizability, offering a cost-effective and practical tool for early identification of cognitive impairment in underserved rural populations in China.