Login / Signup

Using Wearable Cameras to Assess Foods and Beverages Omitted in 24 Hour Dietary Recalls and a Text Entry Food Record App.

Virginia ChanAlyse J DaviesLyndal Wellard-ColeSilvia LuHoi NgLok TsoiAnjali TisciaLouise SignalAnna Maria RanganLuke GemmingMargaret Allman-Farinelli
Published in: Nutrients (2021)
Technology-enhanced methods of dietary assessment may still face common limitations of self-report. This study aimed to assess foods and beverages omitted when both a 24 h recall and a smartphone app were used to assess dietary intake compared with camera images. For three consecutive days, young adults (18-30 years) wore an Autographer camera that took point-of-view images every 30 seconds. Over the same period, participants reported their diet in the app and completed daily 24 h recalls. Camera images were reviewed for food and beverages, then matched to the items reported in the 24 h recall and app. ANOVA (with post hoc analysis using Tukey Honest Significant Difference) and paired t-test were conducted. Discretionary snacks were frequently omitted by both methods (p < 0.001). Water was omitted more frequently in the app than in the camera images (p < 0.001) and 24 h recall (p < 0.001). Dairy and alternatives (p = 0.001), sugar-based products (p = 0.007), savoury sauces and condiments (p < 0.001), fats and oils (p < 0.001) and alcohol (p = 0.002) were more frequently omitted in the app than in the 24 h recall. The use of traditional self-report methods of assessing diet remains problematic even with the addition of technology and finding new objective methods that are not intrusive and are of low burden to participants remains a challenge.
Keyphrases
  • convolutional neural network
  • deep learning
  • young adults
  • optical coherence tomography
  • high speed
  • physical activity
  • weight loss
  • blood pressure
  • human health
  • high resolution