(PO25-20-22) Evaluation of the Comparative Validity of a Smartphone Weight-Loss Trial App With a 24-Hour Diet Recall

The ePoster content is available to NUTRITION 2022 LIVE ONLINE attendees who are logged in.<br><br>If you are registered, log in and then come back here to use this button.

Presenting Author(s)

Faiza Kalam, PhD. MS.
– Postdoctoral Research Fellow, Northwestern University, Orland Park, Illinois, United States

Co-Author(s)

SA

Syed Ali
– Benedictine University
AP

Angela F. Pfammatter
– Northwestern University
BS

Bonnie J. Spring, PhD
– Professor, Northwestern University, Chicago, Illinois, United States
AT

Ayah Takrouri
– Northwestern University
AL

Annie W. Lin, PhD, MS, RD
– Assistant Professor, Benedictine University, Lisle, Illinois, United States

Disclosure(s):

Faiza Kalam, PhD. MS.: No relevant financial relationship(s) with ineligible companies to disclose.

Annie W. Lin, PhD, MS, RD: No relevant financial relationship(s) with ineligible companies to disclose.

Objectives: Despite the presence of several weight loss apps on the market, the validity of the diet data collected from these apps has rarely been tested. A granular analysis of app validity at the food-level, rather than overall intake, is needed to closely examine factors that contribute to the variability of diet data between these apps and standard diet assessments. This study investigated the comparative validity of diet data from a weight loss app at the food level.

Methods: A weight loss app was used for the Sequential Multiple Assignment Randomization Trial (SMART) to assist participants with tracking their food intake throughout the study. A trained dietitian conducted three 24-hour diet recalls at baseline using the Nutrition Data System for Research (NDSR) protocol. Food items reported by participants were categorized into 9 major NDSR food groups (beverage, dairy, fats, fruits, vegetables, grains, protein, sweets, miscellaneous) and 2 other categories (mixed, restaurant). Intraclass correlation coefficients (ICC) determined agreement of diet data (i.e., calories and macronutrients) for all food items. ICC ranges were defined as: ≥0.90, excellent; 0.75 to < 0.90, good; 0.5 to < 0.75, moderate; and < 0.5, poor agreement. Bland Altman analyses determined the estimated mean bias and standard deviation of differences between the app and recall. A subgroup analysis was conducted to evaluate agreement of energy content by food group.

Results: Agreement between the SMART app and recall ranged from moderate to good for all diet data (ICCs = 0.71 to 0.83). Bland Altman plots also confirmed the ICC results; there were little to modest differences between the SMART app and 24-hour recall for energy (-3.0 ± 94.7 kcal), carbohydrates (-0.2 ± 12.2 g), protein (-0.1 ± 5.5 g), and fat (-0.2 ± 5.1 g). When analyzing energy content by food group, agreement between the two diet assessments was excellent for restaurant and sweet food items (ICCs = 0.93 to 0.94) and good for beverages, dairy, fruits, and miscellaneous items (ICCs = 0.76 - 0.87). There was moderate agreement for vegetables, fat, grains, proteins, and mixed dishes (ICCs = 0.51 to 0.68).

Conclusions: There was moderate to good agreement between the SMART app and recall for all diet data. Results suggest that the variability may stem from the type of food reported in the app.

Funding Sources: NIH grants R01DK108678 and T32CA193193