Diet Analysis Apps: What’s Under the Hood?
Publication Date
July 2019
Location
LSU Health Medical Education Building
Document Type
Abstract
Start Date
26-7-2019 9:00 AM
End Date
26-7-2019 12:00 PM
Description
With the technological advances made in the past couple decades, it is not surprising that 81% of Americans own a smartphone. In a time of such advances, failing to utilize these pocket computers to the fullest extent in the pursuit of a healthier public would be a grave mistake. In the burgeoning mHealth field, food-tracking apps have become very popular. Whether designed to help the individual lose weight or just maintain a generally healthier diet, the market is booming. In a 2015 survey, 58.23% of smartphone users said that they had downloaded a diet app; over half of those positive responses reported continued use and high trust in the accuracy of these apps. Several studies have shown deviations between these apps and databases recognized as the standard in the diet analysis field. These deviations could have a significant impact on an individual’s health if they choose to use this nutrient data for health guidance. The interaction between the user and the app’s interface is also critical in the reliability and accuracy of the data output; features that are more prone to miss-selection of a food entry or incorrect portion sizes will lead to inaccurate and misleading data. The focus of our study is to ultimately determine the accuracy of these apps; we begin that evaulation with a look at various features that could result in inaccuracy. In this study, apps were selected from the Google Play store using their search algorithm and two other independent methods. The following criteria were used for the selection in the Google Play Store: 1) the app had to be free; 2) a customer star rating of at least 4 out of 5, and 3) have a diet evaluation function that included more nutrients than just calories. All apps/programs not found in the Google Play store also had to adhere to criteria 1 and 3. Seventeen apps and two web-based computer programs met the criteria and were included in our evaluation. Then, how each app obtained its food database—and from what sources—was investigated, as well as characteristics related to monetization and food search techniques were determined. This information was obtained from several sources: 1) information contained within the app or computer program; 2) blogs, frequently asked question (FAQ) lists, or chatboxes on app-associated websites included with the app and/or computer program; 3) email communication with the app developers. Fifty-nine percent of apps/programs utilized the USDA food composition database. A yes-or-no qualitative assessment of the app was completed for the following features: 1) monetization (account required, in-app purchase offers, premium version, self-advertising of other apps and/or physical products, advertisements), 2) search methods (search bar, common/popular foods, bar code, photo, voice, food recall prompts, custom foods, others), and 3) other (verification labeling, track daily value of non-kcal nutrients, socializability with other users, retroactively view other entries, graphing of nutrient data). A premium subscription for several apps was needed to obtain a more extensive list of nutrients. At least eight different methods of entering foods were found, but all the selected apps and web-based programs had a search bar to find foods. Only 50% allowed quick entry of common or popular foods, while two-thirds recognize bar codes and let the user enter that food’s data. Only 40% visually display a verification label that assures the user of the accuracy of an entry. Most apps allow the user to enter a food from their diet, but not in the database, which, in some instances, can be shared with other users. This data may or may not be verified. In conclusion, some features have the potential to introduce new errors compounding inaccuracies that may exist in the nutrient databases.
Recommended Citation
Rohr, Brayden, "Diet Analysis Apps: What’s Under the Hood?" (2019). Summer Research Internship Program. 6.
https://digitalscholar.lsuhsc.edu/srip/2019/hs/6
Diet Analysis Apps: What’s Under the Hood?
LSU Health Medical Education Building
With the technological advances made in the past couple decades, it is not surprising that 81% of Americans own a smartphone. In a time of such advances, failing to utilize these pocket computers to the fullest extent in the pursuit of a healthier public would be a grave mistake. In the burgeoning mHealth field, food-tracking apps have become very popular. Whether designed to help the individual lose weight or just maintain a generally healthier diet, the market is booming. In a 2015 survey, 58.23% of smartphone users said that they had downloaded a diet app; over half of those positive responses reported continued use and high trust in the accuracy of these apps. Several studies have shown deviations between these apps and databases recognized as the standard in the diet analysis field. These deviations could have a significant impact on an individual’s health if they choose to use this nutrient data for health guidance. The interaction between the user and the app’s interface is also critical in the reliability and accuracy of the data output; features that are more prone to miss-selection of a food entry or incorrect portion sizes will lead to inaccurate and misleading data. The focus of our study is to ultimately determine the accuracy of these apps; we begin that evaulation with a look at various features that could result in inaccuracy. In this study, apps were selected from the Google Play store using their search algorithm and two other independent methods. The following criteria were used for the selection in the Google Play Store: 1) the app had to be free; 2) a customer star rating of at least 4 out of 5, and 3) have a diet evaluation function that included more nutrients than just calories. All apps/programs not found in the Google Play store also had to adhere to criteria 1 and 3. Seventeen apps and two web-based computer programs met the criteria and were included in our evaluation. Then, how each app obtained its food database—and from what sources—was investigated, as well as characteristics related to monetization and food search techniques were determined. This information was obtained from several sources: 1) information contained within the app or computer program; 2) blogs, frequently asked question (FAQ) lists, or chatboxes on app-associated websites included with the app and/or computer program; 3) email communication with the app developers. Fifty-nine percent of apps/programs utilized the USDA food composition database. A yes-or-no qualitative assessment of the app was completed for the following features: 1) monetization (account required, in-app purchase offers, premium version, self-advertising of other apps and/or physical products, advertisements), 2) search methods (search bar, common/popular foods, bar code, photo, voice, food recall prompts, custom foods, others), and 3) other (verification labeling, track daily value of non-kcal nutrients, socializability with other users, retroactively view other entries, graphing of nutrient data). A premium subscription for several apps was needed to obtain a more extensive list of nutrients. At least eight different methods of entering foods were found, but all the selected apps and web-based programs had a search bar to find foods. Only 50% allowed quick entry of common or popular foods, while two-thirds recognize bar codes and let the user enter that food’s data. Only 40% visually display a verification label that assures the user of the accuracy of an entry. Most apps allow the user to enter a food from their diet, but not in the database, which, in some instances, can be shared with other users. This data may or may not be verified. In conclusion, some features have the potential to introduce new errors compounding inaccuracies that may exist in the nutrient databases.
Comments
Mentor: Lauri O. Byerley, Department of Physiology