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.

Cleaning the dataset extracted at T2 resulted in a 2.8% rejection. 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. 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. Bias was estimated using Bland-Altman plots. Original and cleaned T2 values were correlated with the Nubel calculated values. These limits were applied to the MyFitnessPal dataset extracted at T2 to remove extremely high and likely erroneous values. 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. Nutrient intake values were calculated either manually, using the food composition database Nubel, or automatically, using the database coupled to MyFitnessPal. In this study, we validate the database of MyFitnessPal versus the Belgian food composition database, Nubel.Īfter 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. The reliability of such platforms depends on the quality of the associated food database. Digital food registration via online platforms that are coupled to large food databases obviates the need for manual processing of dietary data.

Myfitnesspal nutrition facts