(PS1-77) Association Between User Preference and Accuracy for Estimating Food Intake with Myfitnesspal in Controlled and Free-living Conditions

In controlled laboratory conditions, participants (N = 42, 19-70 y, 56% F, body mass index, BMI, 19-46 kg/m2) used the MyFitnessPal app on their smart phone to estimate energy intake during a simulated meal. A separate set of participants in free-living conditions (N = 43, 24-62 y, 54% F, BMI 19-47 kg/m2) used MyFitnessPal to estimate energy intake over a three-day period. Across three days, participants were given pre-weighed food in a cooler where they used MyFitnessPal to estimate their energy intake after each meal (the cooler was returned the next day to weight actual intake).  

Participants’ accuracy of estimated energy intake was assessed by calculating the percent error between their reported energy intake and the estimate obtained from covertly weighing their food provision and waste. Lower error values indicated higher accuracy. User preference was assessed using a survey that compared features of MyFitnessPal with another app that participants had used for a similar period. For this analysis, preference was quantified as the percentage of times participant’s preferred MyFitnessPal (vs. The other app) across 11 preference-related questions. Linear mixed models, adjusting for sex and age, were used to test the relationship between MyFitnessPal preference and percent error of energy intake estimates.  

There was no significant relationship between user preference for MyFitnessPal and error rates for energy intake estimates in laboratory conditions (slope = 0.05 ± 0.20; p = 0.80). Greater preference for MyFitnessPal was associated with less error in participant’s energy intake estimates in free-living conditions (slope = -0.2 ± 0.09; p = 0.03). The results suggest that individuals who have greater preference for their food tracking app may have higher accuracy when measuring their energy intake, in “real world” conditions.