Integrating Food Preference Profiling, Behavior Change Strategies, and Machine Learning for Cardiovascular Disease...

Area of Science:

• Digital Health Interventions
• Personalized Nutrition
• Cardiovascular Disease Risk Prediction

Background:

• Personalized dietary advice must balance health risks with individual food preferences.
• Tailoring recommendations to personal tastes enhances adherence and effectiveness.

Purpose of the Study:

• To develop a digital health intervention (DHI) providing personalized nutrition recommendations.
• To create a system based on individual food preference profiles (FPP) and cardiovascular disease (CVD) risk.

Main Methods:

• Utilized UK Biobank data from 61,229 participants to develop a DHI pipeline.
• Created a CVD prediction model using food preference profiles and machine learning algorithms.
• Incorporated the Behavior Change Wheel to design intervention features and behavior change techniques.

Main Results:

• Identified 14 key food items for effective food preference profiling.
• The food preference profile set achieved prediction accuracy comparable to traditional methods.
• Selected four DHI features: food/portion info, recipes, recommendation system, and community platforms.

Conclusions:

• Demonstrated proof of principle for a DHI pipeline integrating personal needs and preferences.
• The DHI has the potential to improve dietary choices and reduce CVD risk.
• Making healthy eating accessible and sustainable can significantly impact public health.