Energy intake estimation from counts of chews and swallows

Area of Science:

• Nutritional Science
• Biomedical Engineering
• Wearable Technology

Background:

• Current dietary assessment methods, primarily self-report, suffer from significant inaccuracies, time demands, and user burden.
• There is a critical need for objective and reliable methods to quantify energy intake.
• Wearable sensors and video analysis offer novel avenues for objective physiological measurements.

Purpose of the Study:

• To evaluate the efficacy of individually-calibrated models based on Counts of Chews and Swallows (CCS models) for estimating energy intake.
• To compare the accuracy of CCS models against established dietary assessment techniques.

Main Methods:

• Participants consumed standardized meals in a laboratory setting for model training and validation.
• Energy intake was assessed using weighed food records (gold standard), diet diaries, photographic records, and CCS models.
• Counts of chews and swallows were captured using wearable sensors and video analysis.

Main Results:

• CCS models demonstrated the lowest reporting bias and reduced error compared to diet diaries during training meal assessments.
• For the validation meal, CCS models exhibited comparable reporting errors to diary and photographic methods.
• Observed increases in error for the validation meal may be mitigated by incorporating food physical property correction factors.

Conclusions:

• Individually-calibrated CCS models present a viable and promising alternative to overcome the limitations inherent in self-reported dietary assessments.
• Further refinement of CCS models, including correction factors, could enhance their accuracy and applicability in diverse dietary contexts.