Abstract
BACKGROUND: Emerging approaches to building more efficient and effective behavioral interventions are becoming more widely available. The current paper provides an empirical example of the use of the engineering-inspired multiphase optimization strategy (MOST) to build a remotely delivered responsive parenting intervention to prevent obesity among children of low-income mothers with and without depressive symptoms. METHODS: Participants were 107 mothers with (n = 45) and without (n = 62) depressive symptoms who had a child aged 12 to 42 months participating in the Women, Infants and Children program. Participants were randomized to one of sixteen experimental conditions using a factorial design that included a combination of the following eight remotely delivered intervention components: responsive feeding curriculum (given to all participants), parenting curriculum, portion size guidance, obesogenic risk assessment, personalized feedback on mealtime routines, feeding curriculum counseling, goal setting, mobile messaging, and social support. This design enabled efficient identification of components with low feasibility and acceptability. RESULTS: Completion rates were high (85%) and did not statistically differ by depressive symptoms. However, mothers with depressive symptoms who received obesogenic risk assessment and personalized feedback on mealtime routines components had lower completion rates than mothers without depressive symptoms. All intervention components were feasible to implement except the social support component. Regardless of experimental condition, most participants reported that the program increased their awareness of what, when, and how to feed their children. CONCLUSIONS: MOST provided an efficient way to assess the feasibility of components prior to testing them with a fully powered experiment. This framework helped identify potentially challenging combinations of remotely delivered intervention components. Consideration of how these results can inform future studies focused on the optimization phase of MOST is discussed.