Related Experiment Video
Updated: Aug 6, 2026
Concept Development and Use of an Automated Food Intake and Eating Behavior Assessment Method
Published on: February 19, 2021
Methodology development for analyzing eating behavior, diet quality, and food environment in occupational settings
Valeria Cecchini1, Negar Zand Miralvand2, Vasileios Papapanagiotou2
1Renal Research Novum, Division of Renal Medicine, Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm, Sweden.
Background:
Mobile-based dietary assessment tools have the potential to capture real-time eating behaviors and food environment exposure, but their feasibility and usability in workplace contexts remain underexplored.
Objective:
To evaluate the feasibility and usability of the Mobile Sense App among employees and students at the Karolinska Institute.
Methods:
This study employed the Mobile Sense App, which was developed to record dietary intake and food purchasing patterns through photographs and geospatial data. The study combined data from the pre-pilot and pilot studies (n=56). Participants, who were students and employees at a public institution, completed a baseline questionnaire and used the application (app) for two weeks to photograph all meals, snacks, and drinks consumed during working hours on campus. Adherence was evaluated based on the number of photographs submitted over a six-day monitoring period. Usability was assessed using the System Usability Scale (SUS). Image data were manually annotated, ultra-processed foods (UPFs) were identified according to the NOVA classification system, and inter-rater reliability was evaluated using Cohen's κ. Food purchasing 'hotspots' were examined through geospatial analyses.
Results:
Participants submitted n=1178 photographs. Cohen's κ indicated strong agreement between independent researcher annotation (κ=1.00 for meals/snacks; κ=0.91 for drinks). The mean SUS score was 75.2 ± 12.2. Store-bought meals showed a higher prevalence of UPFs (64%) than home-made meals (13%) (χ2 (1)=178.63, p<.001). Global Positioning System (GPS) data indicated food purchases clustered around campus dining areas. Open-ended feedback emphasized the need for reminders, clarity in user interface (UI) instructions, and flexibility in annotation options as desired improvements.
Conclusions:
The proposed approach is a feasible and scalable methodology for real-world dietary monitoring, supporting further evaluation for larger-scale implementation.
