Current Status and Risk Factors of Mild Cognitive Impairment among Older Adults in Wuhan: A Machine Learning-Based Analysis

Background This study seeks to understand the current situation of mild cognitive impairment (MCI) in Wuhan, explore the related risk factors, and identify high-risk groups. Methods Multistage cluster sampling was adopted to recruit a total of 2,190 adults aged 60 years and above from 30 residential (village) committees in 13 administrative districts. They were screened for MCI using the Community Screening Instrument for Dementia (CSI-D).The Least Absolute Shrinkage and Selection Operator (LASSO) was employed to screen predictive variables. A variety of machine learning classification models were integrated to analyze and identify the optimal model. Multiple evaluation indicators were utilized to compare the performance of models, including the area under the receiver operating characteristic curve (AUC) and decision curve analysis (DCA). Binary logistic regression was adopted to determine the effect values of risk factors, and SHapley Additive ex Planations (SHAP) was applied to interpret machine learning models. Results 35.3% participants were diagnosed with MCI. The Final model included age, occupation, sleep disorders and literacy level. Based on the AUC and DCA in the validation group, the XGBoost model demonstrated excellent performance, the most crucial features in this model and their effect values as follows: sleep disorders (OR: 1.44, 95%CI: 1.31–1.82), aged above 75 (OR: 1.52, 95%CI: 1.12–2.08), High school and above (OR: 0.39, 95%CI: 0.28–0.56), Middle school (OR: 0.48, 95%CI: 0.35–0.65), Farmers were included but without significant effect. Conclusion The current status of MCI among older adults in Wuhan is not optimistic. Therefore, early screening and intervention should be carried out for farmers and individuals with advanced age, low literacy,and sleep disorders.