Advertisement

P019 Validation of a Smartphone-Based App for Assessing Energy Intake: A Pilot Study

      Background

      Smartphone-based methods are innovative and promising tools to accurately quantify energy intake (EI). The PortionSizeTM app measures dietary intake using embedded templates and provides real-time feedback. PortionSize also sends food images to a server where they can be analyzed using the Remote Food Photography Method (RFPM). The RFPM accurately measures EI; however, it requires human raters to analyze food images; therefore, the RFPM does not provide dietary feedback in real-time.

      Objective

      To compare the validity of EI estimates from PortionSize and RFPM to weighed meals (WM).

      Study Design, Setting, Participants

      Fifteen adults used the PortionSize app during covertly-weighed simulated meals in a laboratory setting. Trained personnel also quantified EI from the images captured from the PortionSize using the RFPM.

      Measurable Outcome/ Analysis

      Demographics and body mass index (BMI). Dependent t-tests were performed to investigate whether EI measured with templates in PortionSize or using the RFPM differed to EI from WM.

      Results

      Mean (± SD) age and BMI of participants were 28 (± 12) years and 24.1 ± 6.6 kg/m2, respectively, and 73.3% were female. PortionSize estimated EI was 743 ± 328 kcal, EI RFPM-estimated intake was 660 ± 196 kcal, and weighed intake was 659 ± 191 kcal. The mean differences of EI estimation between PortionSize and WM (84 ± 288 kcal) and between RFPM and WM (1 ± 32 kcal) were not significant (P > .05). The mean difference in EI estimation (83 ±284 kcal) between PortionSize and RFPM was not significant (P > .05).

      Conclusions

      RFPM produced very close estimates of EI to WM but requires a human rater and is not easily scalable. These findings are important as PortionSize provides real-time feedback on dietary intake, which may assist with timely behavior change. Improvements to make PortionSize more scalable are ongoing. This work warrants future well-powered trials.

      Funding

      NIH; Pennington Biomedical (PBRC)/Louisiana Nutrition Obesity Research Center; Louisiana Clinical and Translational Science Center

      SUPPLEMENTARY DATA

      Supplementary data related to this article can be found at https://doi.org/10.1016/j.jneb.2022.04.059.

      Appendix. SUPPLEMENTARY DATA