Approved Research

Risk prediction using machine learning and decision-making models for cardiovascular disease

Lay summary

Cardiovascular disease (CVD) is the leading cause of death and brings a high disease burden. The current guidelines have recommended adopting risk prediction models to help clinicians to make better choices in treatment decisions for patients. Machine learning algorithms have advantages in identifying complicated relationships between exposures and outcomes, which could improve model accuracy. We aim to (1) explore the associations between exposures and CVD using machine learning algorithms and mendelian randomization methods;(2) develop and validate the risk prediction models for CVD using machine learning based on various variables (e.g., lifestyles, imaging data, genetic information); (3) evaluate the impact and efficiency of risk-based preventive strategies for CVD using simulation models. Our research will last 36 months and the findings will provide evidence to apply precise risk assessment tools and take risk-based personalized measures, which will guide decision-making in disease prevention and management for CVD and the reduce disease burden.