A utility-based machine learning-driven personalized lifestyle recommendation for cardiovascular disease prevention

Area of Science:

• Cardiovascular Health
• Machine Learning Applications
• Preventive Medicine

Background:

• Cardiovascular disease (CVD) is a leading global cause of mortality.
• Many CVDs are preventable through lifestyle modifications.
• Personalized recommendations are challenging due to complex risk factor interactions and individual considerations.

Purpose of the Study:

• To develop an effective personalized lifestyle recommendation algorithm for reducing CVD risk.
• To address the complexity of risk factor relationships and individual effort-benefit considerations.
• To incorporate uncertainty in disease progression into lifestyle modification recommendations.

Main Methods:

• A data-driven approach combining machine learning and a personalized exponential utility function.
• Implementation of a classification-based prediction model for CVD risk assessment.
• Utilization of Generative Adversarial Networks (GANs) to model risk factor relationships and disease progression uncertainty.
• Proposal of a novel personalized exponential utility function for evaluating modifications.

Main Results:

• The developed algorithm effectively predicts CVD risk based on various risk factors.
• GANs successfully quantified disease progression uncertainty under lifestyle changes.
• The personalized utility function identified optimal lifestyle modifications for individuals.
• Validation on an open-access CVD dataset confirmed the method's effectiveness in reducing CVD risk.

Conclusions:

• The proposed methodology offers a promising data-driven approach for personalized CVD prevention.
• Personalized lifestyle modifications can significantly reduce cardiovascular disease risk.
• The algorithm has potential for real-world application in public health initiatives.