Development and Validation of a Prediction Model for Mild Cognitive Impairment Among Middle-aged Chinese Adults

Background: Mild Cognitive Impairment (MCI) represents a transitional state between normal aging and dementia, with 10-15% of cases progressing to dementia annually. With China bearing approximately 25% of the global dementia burden, early identification of MCI in middle-aged adults is crucial for implementing timely interventions. However, limited research has focused on developing prediction models for MCI risk assessment in primary healthcare settings. Methods: Using data from the China Health and Retirement Longitudinal Study (2011-2020), we analyzed 5,289 middle-aged adults (45-60 years) with normal cognitive function at baseline. Three penalized regression models (Lasso, Elastic-net, and Ridge) were employed to identify predictors of MCI, with the best-performing model used to construct a nomogram. Model performance was assessed through discrimination and calibration metrics with internal validation. Results: Among 33 potential predictors, seven variables were identified as key predictors through three penalized regression: Education, Blind or Half Blind, Body Pains, Play Cards, Go to Sports Club, Surf the Internet, Health Status. The Elastic-net model showed the best performance with an AUC of 0.703 (95% CI: 0.669-0.738) in the validation set. A nomogram was subsequently developed for clinical application. Conclusion: This prediction model provides a practical tool for early identification of cognitive impairment risk in middle-aged Chinese adults, with potential applications in primary healthcare settings through its user-friendly nomogram and identification of modifiable risk factors.