(PS13-A10) Portionsize: An App-based Method to Improve Adherence to Myplate Dietary Guidelines in Real Time

Current United States Department of Agriculture (USDA) dietary recommendations are promoted with MyPlate, which provides people with age- and sex-specific daily energy intake targets and a recommended number of servings of specific food groups (e.g., 2 cup-equivalents of fruits, 3 cup-equivalents of vegetables, 6 ounce-equivalents of protein foods, etc.). Behavior change, including improving dietary intake, is enhanced when people receive objective information about the target behavior when, or shortly after, it occurs. Improving dietary intake is challenging due to previous dietary assessment methods: 1) failing to provide feedback about intake of specific MyPlate food groups, and 2) providing limited real time feedback. The PortionSize app aims to provide people with information about the energy and food group content of their: 1) food selection (before eating occurs), 2) food intake (after eating occurs), and 3) cumulative food intake throughout the day in relation to their MyPlate dietary intake goals. Thus, the PortionSize app was designed to help individuals modify their intake when food is selected or to modify subsequent food intake to better adhere to their MyPlate goals. The app can be deployed remotely at no cost on iOS devices and can be coded for Android devices; hence, it may help alleviate inequities in health promotion. Herein, the preliminary validity of PortionSize is reported to evaluate the ability of the app to guide dietary change.

Participants (N=42, 18-70 years old, body mass index or BMI 18.5-50 kg/m²) used PortionSize to estimate food intake during a laboratory-based test meal. Two One-Sided t-tests, with 21% equivalence bounds, were used to assess equivalence between weighed intake (the criterion measure) and intake estimated with PortionSize. The outcome variables were assessed at the meal level and included energy (kcal) content and servings of MyPlate food groups (i.e., fruits, vegetables, grains, dairy, and protein).

The sample was 54.8% female. Results are presented as Mean (SD). Mean body mass index and age were 28.2 (6.9) kg/m² and 36.3 (16.6) years, respectively. Compared to weighed intake, PortionSize underestimated intake by 151 (489) kcal or 13.3% and the values were not equivalent (p=0.13). PortionSize underestimated the five MyPlate food groups and no values were equivalent to weighed values (p-values > 0.18). However, the magnitude of food group errors was relatively small for fruits (-0.10 servings), vegetables (-0.22 servings), grains (-0.35 servings), and dairy (-0.32 servings). Protein foods had error of 0.45 servings.

At the meal level, PortionSize failed to produce energy and food group values that were equivalent to weighed foods. The magnitude of error, however, was modest for an app that automatically estimates food selection and intake. The estimates can be useful when facilitating dietary adherence, but the clinical utility of the app would be enhanced if: 1) its accuracy was improved, and 2) food intake was evaluated over more than one meal. The app is being updated to improve accuracy and usability, and the app is undergoing further testing.