(P28-012-23) Testing the Validity of the PortionSize™ Smartphone App for Estimating Dietary Intake in Free-Living Conditions: A Pilot Study

Objectives: The PortionSize™ smartphone app (PS) provides users with real-time feedback on food selection and intake, which may assist in modifying dietary intake in order to adhere to a specific diet. This free-living pilot study tested the validity of dietary intake estimates from the PS app compared to intake determined by trained raters who used the validated Remote Food Photography Method (RFPM).

Methods: Fourteen adults (71% female; mean ± SD, age: 26 ± 11 y) used the PS app to record their food intake of all eating occasions across 3 consecutive days in free-living conditions. Participants measured the portion size of foods and drinks by using embedded customizable portion templates in the PS app. Then the app automatically estimated dietary intake using the templates and a built-in nutrient database (Food and Nutrient Database for Dietary Studies). Two trained researchers simultaneously quantified food in the pictures captured by the app using the RFPM, which served as the criterion measure and was used to evaluate the validity of the PS app. Two one-sided T-tests (± 25% equivalence bounds) were performed, and Bland-Altman analysis determined if error variance differed over levels of intake.

Results: Across 3 days, participants recorded 147 meals, excluding 6 outlier meal entries from both methods, resulting in a total of 141 meals. Estimated food intake by the PS app (mean ± SD, 1105 ± 418 g/day) was similar to the estimated intake by RFPM (1086 ± 396 g/day) and the means were equivalent (p < 0.001). PS app and RFPM estimates of energy intake (1561 ± 447 vs. 1350 ± 389 kcal/day); however, were not equivalent (p=0.08), resulting in an overestimation of intake using the PS app (211 ± 318 kcal/day) compared to RFPM. Bland–Altman analysis indicated that the measurement error of the PS app in estimating food and energy intake was consistent over different levels of food (R2=0.04, p=0.50) and energy (R2=0.06, p=0.39) intake.

Conclusions: Based on this free-living pilot study, the PS app produced equivalent estimates of food intake compared to the RFPM without requiring human rating, and its measurement error was similar across levels of food intake. PS app estimations are currently being improved by a dynamic process of app refinement and larger validation studies comparing estimated energy intake to weighed meals.

Funding Sources: NIDDK, LA CaTS, LIFT Funds