Comparative analysis of cardiometabolic multimorbidity predictors in China and the USA: A machine learning approach

Area of Science:

• Public Health
• Computational Biology
• Epidemiology

Background:

• Cardiometabolic multimorbidity (CMM), the co-occurrence of cardiovascular disease (CVD) and metabolic diseases (MD), poses a significant global health challenge.
• Current screening tools for CMM are insufficient and lack applicability across different nations, particularly between the USA and China.
• There is a critical need for advanced, cross-nationally applicable screening methods to identify individuals at risk for CMM.

Purpose of the Study:

• To develop and validate a bidirectional machine learning (ML) model for predicting cardiometabolic multimorbidity (CMM) risk.
• To identify distinct and shared risk factors for CMM in China and the USA.
• To enable precision screening and inform region-specific public health interventions for CMM.

Main Methods:

• Utilized data from the Chinese Health and Retirement Longitudinal Survey (CHARLS) and the US Health and Retirement Study (HRS).
• Developed and compared bidirectional ML models, including logistic regression (LR), Gaussian Naive Bayes (GNB), and extreme gradient boosting (XGBoost).
• Evaluated model performance using Area Under the Curve (AUC) and performed cross-national validation with SHAP analysis for predictor identification.

Main Results:

• Logistic regression models demonstrated strong performance in predicting cardiovascular disease (CVD) (AUC = 0.70 in both countries).
• Predictive accuracy for metabolic diseases (MD) varied, with LR (AUC = 0.71) excelling in China and GNB (AUC = 0.65) in the USA.
• Core predictors included disease counts and moderate physical activity, with unique regional predictors identified, such as grip strength in China and pulse/emotional problems in the USA.

Conclusions:

• A validated, bidirectional ML model offers a promising tool for precision CMM screening across different countries.
• Region-specific screening strategies are recommended, emphasizing grip strength and balance for China and pulse monitoring and psychological interventions for the USA.
• This approach facilitates targeted public health interventions to mitigate the burden of cardiometabolic multimorbidity.