Development and validation of a Cognitive Impairment Risk Prediction Model for Elderly Patients with Multimorbidity

Background: Cognitive impairment is a prevalent issue among the elderly population. Multimorbidity has been pinpointed as a clinical risk factor that is both potential and easy to recognize. However, the association between multimorbidity and cognitive decline among older adults in China remains underexplored.Based on this,this study aims to develop a risk prediction model for cognitive impairment in elderly individuals aged 60 and above with multimorbidity. Methods: This investigation used information from the 2020 China Health and Retirement Longitudinal Study (CHARLS), including a total of 5,977 elderly patients with multimorbidity. The Least Absolute Shrinkage and Selection Operator (LASSO) regression method was used to select feature variables. To address the issue of imbalanced cognitive impairment data distribution, the Synthetic Minority Oversampling Technique (SMOTE) algorithm was applied for data balancing. A cognitive impairment risk prediction model was constructed, and its performance was evaluated using the area under the Receiver Operating Characteristic (ROC) curve (AUC), accuracy, sensitivity, specificity, calibration curves, and decision curve analysis (DCA). Results: Among the 5,977 elderly patients with multimorbidity, 3,047 (50.98%) had normal cognition, while 2,930 (49.02%) were diagnosed with cognitive impairment. Logistic regression analysis identified 11 influencing factors for cognitive impairment in elderly patients with multimorbidity, including age, gender, education level, marital status, type of residence, pension insurance, health insurance, social participation, basic activities of daily living (BADL), instrumental activities of daily living (IADL), and depression. Based on these 11 variables, a cognitive impairment risk prediction model was developed. In the training dataset, the model achieved an AUC of 0.809 (95% CI: 0.796–0.822), while in the validation dataset, the AUC was 0.819 (95% CI: 0.800–0.839). The accuracy was 0.742 and 0.749, sensitivity was 0.775 and 0.720, and specificity was 0.711 and 0.779, respectively, demonstrating a strong consistency between predicted and actual values. Conclusion: The cognitive impairment risk prediction model developed in this study exhibited good predictive performance, providing scientific evidence for community healthcare professionals in the early assessment and identification of cognitive impairment risk in elderly patients with multimorbidity.