Abstract
BACKGROUND: Tracking of dietary intake is key to enhancing weight loss. Mobile apps may be useful for tracking food intake and can provide feedback about calories and nutritional value. Recent technological developments have enabled image recognition to identify foods and track food intake. OBJECTIVE: We aimed to determine the effectiveness of using photography as a feature of a smartphone weight loss app to track food intake in adults who were overweight or obese. METHODS: We analyzed data from individuals (age, 18-65 years; body mass index≥25 kg/m(2); ≥4 days of logged food intake; and ≥2 weigh-ins) who used a mobile-based weight loss app. In a retrospective study, we compared those who used the photo feature (n=9871) and those who did not use the feature (n=113,916). Linear regression analyses were used to assess use of the photo feature in relation to percent weight loss. RESULTS: Weight loss was greater in the group using the photo feature (Δ=0.14%; 95% CI 0.06-0.22; P<.001). The photo feature group used the weight loss app for a longer duration (+3.5 days; 95% CI 2.61-4.37; P<.001) and logged their food intake on more days (+6.1 days; 95% CI 5.40-6.77; P<.001) than the nonusers. Mediation analysis showed that the weight loss effect was absent when controlling for either duration or number of logged days in the program. CONCLUSIONS: This study was the first to examine the effect of a food photo feature to track food intake on weight loss in a free-living setting. Use of photo recognition was associated with greater weight loss, which was mediated by the duration of app use and number of logged days in the program.