Explainable machine learning‐driven predictive modelling of incident cognitive impairment in individuals with early‐stage cardiovascular–kidney–metabolic syndrome: Insights from a longitudinal CHARLS cohort study
Authors' conclusion
Our interpretable machine learning framework, built upon readily accessible clinical and metabolic parameters, provides a robust and clinically applicable tool for predicting 10-year risk of cognitive impairment among individuals with early-stage CKM syndrome.
Study context
- Population
- Adults with cardiovascular-kidney-metabolic (CKM) syndrome stages 0–2 at baseline (2011) with complete follow-up through 2020, China Health and Retirement Longitudinal Study (CHARLS)
- Participants
- 4,462 adults
- Outcome
- Incident cognitive impairment (cognitive performance >1 SD below the age-adjusted mean)
- Follow-up
- 10 years
Methodological detailsmachine learning prediction model, nationally representative cohort, Predictors screened with Boruta algorithm and LASSO regression, then multicollinearity assessment; 9 ML models trained and validated; interpretability by SHAP
- Machine learning prediction model
- Nationally representative cohort
- Predictors screened with Boruta algorithm and LASSO regression, then multicollinearity assessment; 9 ML models trained and validated; interpretability by SHAP
Main finding
As reported in the abstract
Incident cognitive impairment · 10 years of follow-up
Developed cognitive impairment
525 participants
11.8%
XGBoost model best performance: validation AUC
0.726
Observational data: associations, not effects.
Further findings
- Predictors retained (education, total cholesterol, age, FPG, uric acid, estimated glucose disposal rate, diastolic BP, relative fat mass, CKM stage, metabolic syndrome)
- 10
- SHAP top predictors: education, total cholesterol and age
Original publication
Explainable machine learning‐driven predictive modelling of incident cognitive impairment in individuals with early‐stage cardiovascular–kidney–metabolic syndrome: Insights from a longitudinal CHARLS cohort study. Diabetes Obes Metab. Published 10 December 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →