FoodCoach: Fully Automated Diet Counseling

Area of Science:

• Digital Health
• Nutrition Science
• Artificial Intelligence in Healthcare

Background:

• Unhealthy dietary habits are a primary cause of non-communicable diseases (NCDs).
• Traditional diet counseling is resource-intensive and manually operated, limiting its scalability.
• Automated solutions are needed to overcome the limitations of manual diet counseling for NCD prevention.

Purpose of the Study:

• To introduce FoodCoach, a fully automated system for diet counseling using digital receipt data.
• To present the system's recommender algorithm, architecture, and evaluation through a randomized controlled trial.
• To demonstrate the potential for scalable, automated diet monitoring and personalized recommendations.

Main Methods:

• Development of a novel dietary recommendation algorithm with clinical expert input.
• Implementation of a scalable system architecture utilizing a knowledge graph for data interoperability.
• Conducting a two-arm randomized controlled trial with 61 participants to evaluate FoodCoach's feasibility and impact.

Main Results:

• The FoodCoach system demonstrated technical feasibility for scalable, automated diet counseling.
• The randomized controlled trial showed no significant improvement in participants' food purchase healthiness.
• The study provides implementations and deployment guidance for the FoodCoach system.

Conclusions:

• FoodCoach offers a scalable, automated approach to diet monitoring and recommendations, augmenting traditional counseling.
• The system can conserve clinical resources by streamlining the diet counseling process.
• Further research and development can enhance the impact of automated diet counseling on NCD reduction.