Use of a Mobile Food Record, Enhanced by a Novel Artificial Intelligence-Based Informatics Framework to Expand Capture of Dietary Intake

Human dietary habits remain challenging to measure. Even the "gold-standard" interviewer-administered 24 hour dietary recall has established underreporting issues especially in high-risk populations, such as patients with obesity. Performance of precision nutrition research requires state of the art tools to capture individual dietary patterns. Given the widespread availability of smart phones, mobile-phone based food records present a prime opportunity to capture "in the field" dietary intake. Mobile food records, however, are not yet widely used in nutrition research. Several reasons exist. First, the mobile food record needs to be validated relative to the "gold standard" of dietary intake, the 24-hour dietary recall. Second, many available mobile food records are commercially based, with caveats about data availability and meeting research-quality data security practices (HIPAA compliance).

In response, this study includes using a freely available, research-focused, HIPAA compliant, mobile food record (mCC: my Circadian Clock app) actively used in research studies (including our own ongoing work) to test the hypothesis that a mobile food record can expand capture of dietary intake. The proposed aims include the following: Using a randomized, cross-over design, evaluation of augmenting interviewer-administered 24-hour dietary recalls by a mobile food record (Augmented Recall) results in less energy intake underreporting than a standard (un-augmented) interviewer administered 24-hour dietary recall (Standard Recall).