Transformer-based models for predicting cardiovascular risk in Chinese adults: development and validation

The authors aimed to develop and validate transformer-based models, named China-AIHeart, for predicting 10-year cardiovascular disease (CVD) risk in Chinese adults, addressing the limitations of traditional Cox proportional hazards models. Their study demonstrated that these models significantly outperformed Cox models and established risk scores in terms of discrimination and calibration, providing a robust tool for CVD risk stratification in this population. External validation confirmed the models' effectiveness across different cohorts, highlighting their practical applicability in clinical settings.