Sex differences in cognitive function trajectories and influencing factors in older adults: A machine learning study based on CHARLS and HRS.

BACKGROUND: Identify the influencing factors that respectively affect the differences in cognitive function trajectories between China and the United States, and explore the reasons for gender and transnational differences.

METHODS: Five waves of data from the China Health and Retirement Longitudinal Study (CHARLS) and the Health and Retirement Study (HRS) were utilized, and Latent Class Growth Models (LCGMs) was employed to identify cognitive trajectories. The optimal machine learning (ML) model was adopted and the Boruta algorithm was used to screen the variables. The SHAP value is used to explain the ranking of feature importance.

RESULTS: The XGBoost model performs best among the six ML models. Education was the top influencing factor across groups. Family factors mattered more in China, while U.S. medical and endowment insurance had stronger effects. Men were more affected by external social networks; women were influenced by family responsibilities and community environments.

CONCLUSION: Differences were linked to public services, social roles, and traditional culture. These results support the development of tailored interventions targeting gender-specific and culture pathways at the individual, family, society and community levels, offering actionable insights for promoting cognitive health in aging populations globally.