Calorie Compensation Patterns Observed in App-Based Food Diaries

Self-regulation of food intake is necessary for maintaining a healthy body weight. One of the characteristics of self-regulation is calorie compensation. Calorie compensation refers to adjusting the current meal's energy content based on the energy content of the previous meal(s). Preload test studies measure a single instance of compensation in a controlled setting. The measurement of calorie compensation in free-living conditions has largely remained unexplored. This paper proposes a methodology that leverages extensive app-based observational food diary data to measure an individual's calorie compensation profile in free-living conditions. Instead of a single compensation index followed in preload-test studies, we present the compensation profile as a distribution of days a user exhibits under-compensation, overcompensation, non-compensation, and precise compensation. We applied our methodology to the public food diary data of 1622 MyFitnessPal users. We empirically established that four weeks of food diaries were sufficient to characterize a user's compensation profile accurately. We observed that meal compensation was more likely than day compensation. Dinner compensation had a higher likelihood than lunch compensation. Precise compensation was the least likely. Users were more likely to overcompensate for missing calories than for additional calories. The consequences of poor compensatory behavior were reflected in their adherence to their daily calorie goal. Our methodology could be applied to food diaries to discover behavioral phenotypes of poor compensatory behavior toward forming an early behavioral marker for weight gain.

Accuracy of Nutrient Calculations Using the Consumer-Focused Online App MyFitnessPal: Validation Study

Background

Digital food registration via online platforms that are coupled to large food databases obviates the need for manual processing of dietary data. The reliability of such platforms depends on the quality of the associated food database.

Objective

In this study, we validate the database of MyFitnessPal versus the Belgian food composition database, Nubel.

Methods

After carefully given instructions, 50 participants used MyFitnessPal to each complete a 4-day dietary record 2 times (T1 and T2), with 1 month in between T1 and T2. Nutrient intake values were calculated either manually, using the food composition database Nubel, or automatically, using the database coupled to MyFitnessPal. First, nutrient values from T1 were used as a training set to develop an algorithm that defined upper limit values for energy intake, carbohydrates, fat, protein, fiber, sugar, cholesterol, and sodium. These limits were applied to the MyFitnessPal dataset extracted at T2 to remove extremely high and likely erroneous values. Original and cleaned T2 values were correlated with the Nubel calculated values. Bias was estimated using Bland-Altman plots. Finally, we simulated the impact of using MyFitnessPal for nutrient analysis instead of Nubel on the power of a study design that correlates nutrient intake to a chosen outcome variable.

Results

Per food portion, the following upper limits were defined: 1500 kilocalories for total energy intake, 95 grams (g) for carbohydrates, 92 g for fat, 52 g for protein, 22 g for fiber, 70 g for sugar, 600 mg for cholesterol, and 3600 mg for sodium. Cleaning the dataset extracted at T2 resulted in a 2.8% rejection. Cleaned MyFitnessPal values demonstrated strong correlations with Nubel for energy intake (r=0.96), carbohydrates (r=0.90), fat (r=0.90), protein (r=0.90), fiber (r=0.80), and sugar (r=0.79), but weak correlations for cholesterol (ρ=0.51) and sodium (ρ=0.53); all P values were ≤.001. No bias was found between both methods, except for a fixed bias for fiber and a proportional bias for cholesterol. A 5-10% power loss should be taken into account when correlating energy intake and macronutrients obtained with MyFitnessPal to an outcome variable, compared to Nubel.

Conclusions

Dietary analysis with MyFitnessPal is accurate and efficient for total energy intake, macronutrients, sugar, and fiber, but not for cholesterol and sodium.

Self-Efficacy and Nutrition-Related Goal Achievement of MyFitnessPal Users

Mobile health applications (apps) are increasing in popularity to assist in health behavior change and maintenance. Drawing from constructs outlined within the social cognitive theory by Bandura, we explored how frequently users of a popular health app achieved their health-related goals. Participants reportedly used the MyFitnessPal (MFP) app within the past year completed an online survey on how often they used MFP, their self-efficacy for healthy eating (SE-HE), and achievement of nutrition-related goals. Findings suggest that higher levels of SE-HE and greater use of MFP predicted greater goal achievement. One important finding was that greater use of MFP related to increased reports for goal achievement when SE-HE was lower but not when participants showed high level of SE-HE. The results from this study suggest that MFP may be the most helpful for those who find it more difficult to meet their nutrition-related goals.

Engagement with MyFitnessPal in eating disorders: Qualitative insights from online forums

OBJECTIVE

Using calorie-counting and fitness-tracking technologies is concerning in relation to eating disorders. While studies in this area typically assess one aspect of use (e.g., frequency), engagement with a device or application is more complex. Consequently, important relationships between the use of these technologies and the eating disorder symptomatology might remain undetected. The current study therefore used comments from online eating disorder-related forums to generate comprehensive qualitative insights into engagement with a popular calorie-counting and fitness-tracking application, MyFitnessPal.

METHOD

First, we extracted every comment mentioning MyFitnessPal made on three eating disorder-related forums between May 2015 and January 2018 (1,695 comments from 920 commenters). Then, we conducted an inductive thematic analysis using these comments to identify important aspects of engagement with MyFitnessPal.

RESULTS

The analyses resulted in three themes: Preventing misuse, describing ways in which MyFitnessPal attempts to prevent pathological use and actions taken by users to circumvent its interventions; Accuracy, outlining distrust of MyFitnessPal's accuracy and ways in which perceived inaccuracy is reduced or compensated for; and Psychosocial factors, comprising cognitive, behavioral, and social factors that influence, or are influenced by, engagement with MyFitnessPal.

DISCUSSION

The qualitative insights provide a detailed overview of how people with high levels of eating disorder symptomatology likely engage with MyFitnessPal. The insights can be used as a basis to develop valid, quantitative assessment of pathological patterns of engagement with calorie-counting and fitness-tracking technologies. The findings can also provide clinicians with insight into how their patients likely engage with, and are affected by, these devices and applications.

Computational Approaches Toward Integrating Quantified Self Sensing and Social Media

The growing amount of data collected by quantified self tools and social media hold great potential for applications in personalized medicine. Whereas the first includes health-related physiological signals, the latter provides insights into a user's behavior. However, the two sources of data have largely been studied in isolation. We analyze public data from users who have chosen to connect their MyFitnessPal and Twitter accounts. We show that a user's diet compliance success, measured via their self-logged food diaries, can be predicted using features derived from social media: linguistic, activity, and social capital. We find that users with more positive affect and a larger social network are more successful in succeeding in their dietary goals. Using a Granger causality methodology, we also show that social media can help predict daily changes in diet compliance success or failure with an accuracy of 77%, that improves over baseline techniques by 17%. We discuss the implications of our work in the design of improved health interventions for behavior change.