Feasibility of Using Ecological Momentary Assessment to Quantify Feeding to Soothe Behavior Among Low-Income WIC Mothers: A Pilot Study

Background: Without effective intervention, it is predicted that 50% of children in the US will be obese by 2030. Establishing responsive feeding practices early in life may promote self-regulation, have lasting effects on eating behavior and eventually establish healthy growth and development.   A promising point of intervention is preventing the use of non-responsive feeding practices such as use of food to soothe in non-eating contexts.  In this paper we assess the feasibility of smartphones to collect and characterize maternal use of food to soothe using ecological momentary assessment (EMA) in the context of a responsive feeding intervention among a sample of low-income mothers attending Women, Infant, and Children (WIC) clinics in central Pennsylvania.

Methods: Participants included 122 WIC mothers, >18 years of age, with an infant or toddler aged 0 - 2 years.  Mothers were randomized to one of three responsive feeding curricula conditions: brochure only, in-person, home-delivered, or in-person home delivered plus text messaging.  Mothers completed feeding practice and style questionnaires at baseline and at 6 and 12 weeks post-intervention.  In addition, EMA was used to collect food to soothe behavior in response to a fussy event over a two-day period at each measurement.  Upon completion of the study, mothers participated in an exit interview to examine aspects of feasibility. 

Results: At the end of three months, no intervention effects were observed on use of food to soothe; at baseline, 40% used food to soothe at least once. Although attrition over the study was moderate (26%-40%, depending on study condition), overall enthusiasm for EMA was high.  Nearly two thirds (of the in-home plus condition) watched the video messages, and among those, 71% said they learned something new and 84% reported they would use the information.  Eighty percent reported that the smartphone was easy to use and was not annoying to answer questions.  Food to soothe measures collected with EMA were highly correlated with other validated feeding practices such as pressure, indulgence, and laissez-faire. 

Conclusions: Results from this pilot study show promise both in terms of data collection and intervention implementation with this low-income population.  Use of smartphones to obtain EMA provides an opportunity to better understand the dynamic processes of using food to soothe, and which factors are predictive of that behavior.  Future analyses would benefit from time-varying effect models to examine the effect of theorized predictors on use of food to soothe over time.