The Kunming CalFit study: modeling dietary behavioral patterns using smartphone data.

Area of Science:

• Digital Health
• Behavioral Science
• Nutritional Epidemiology

Background:

• Smartphone sensing offers a practical approach for objective monitoring of health-related behaviors.
• Understanding the interplay between diet, physical activity, and psychosocial stress is crucial for effective health interventions.

Purpose of the Study:

• To investigate the feasibility of using a smartphone application (CalFit) to collect detailed dietary and contextual data in young adults.
• To explore the relationship between dietary intake, physical activity, psychosocial stress, and environmental factors.
• To develop a modeling framework for inferring behavioral patterns for future health interventions.

Main Methods:

• Developed and utilized the CalFit Android application for recording meals (videos) and ecological momentary assessments.
• Collected triaxial accelerometry data for energy expenditure and GPS data for time-location patterns.
• Processed GPS data using the Foodscoremap web service to characterize food environments.
• Trained dietitians analyzed meal videos for nutrient intake.

Main Results:

• Successfully collected comprehensive data on dietary intake, physical activity, stress levels, and environmental exposures from 12 participants.
• Demonstrated the integration of multiple data streams (diet, activity, GPS, psychological state) for a holistic view of behavior.
• Established a foundation for a modeling framework to dynamically infer behavioral patterns.

Conclusions:

• Smartphone-based sensing is a viable tool for capturing complex behavioral data relevant to health.
• Integrating diverse data sources provides valuable insights into factors influencing dietary choices and overall health.
• The developed framework supports the potential for personalized, data-driven health interventions.