Variance in food cue reactivity predicts energy intake among women with binge eating disorder: a pilot study exploring a novel approach to analyzing EMA data

Food cue reactivity, which is experienced as food-related cognitions and cravings, influences eating-related outcomes in laboratory settings, but studies exploring this relationship in an ecologically valid manner are lacking. In this proof-of-concept pilot study, we explored the relationship between dynamic patterns of food-related cognitions and cravings and energy intake among participants undergoing treatment for binge-eating disorder (BED) using two statistical approaches. Fourteen women attending a binge-eating disorder (BED) support group completed a 7-day EMA protocol using items from the General Food Cravings Questionnaire-State (G-FCQ-S) and daily 24-h dietary recalls. Linear mixed-effects models employed two analytic approaches: a traditional method examining within-person deviations in daily mean responses adjusted for between-person differences, and a novel variance-based approach modeling person-specific within-day variability in responses as a predictor of daily energy intake. Variance in responses to three out of five G-FCQ-S questions was predictive of energy intake using our novel approach. Specifically, greater daily variance predicted higher energy intake for "an intense desire to eat" (estimated kcal change = 295.33, p < 0.001), "craving as a physiological state" (estimated kcal change = 304.37, p = 0.002), and "obsessive preoccupation with food" (estimated kcal change = 278.23, p = 0.005). This pilot study demonstrated the feasibility and potential utility of a variance-based analytic approach for EMA data. Modeling within-day variance in food cue reactivity was predictive of fluctuations in energy intake among women undergoing treatment for BED. This methodological approach may inform future EMA research investigating dynamic patterns of food cue reactivity in individuals with BED. Level V, evidence obtained from an observational study.