The SmartFeeding4Kids: Study of a New Web-based Food Parenting Intervention

Overview

This research aims to develop and study the efficacy of a web-based brief intervention, Smart Feeding4Kids, to promote healthy dietary patterns in young children (2 to 6 years old) through changes in parents' feeding practices. The intervention is grounded on self-regulation and habit-formation models and combines the use of several effective behavioral methodologies. The multidisciplinary team integrates experienced researchers on parenting interventions, child nutrition, and the development of online applications to support personalized nutritional assessment and psychological interventions. The monitoring of the use of the platform and the knowledge about the predictors of efficacy, adherence, and involvement obtained in this project will offer professionals essential information to the development of future online interventions. The project will also contribute with knowledge concerning the most effective methodologies for changing parental feeding practices and collects unique information about the eating habits and practices of parents of Portuguese children. The main hypotheses of the study: 1. parents who enrolled in behavior change and social support interventions will report significantly higher use of effective feeding practices (child's self-regulation intake practices, food availability, and accessibility practices) and significantly lower use of ineffective feeding practices (food control, restriction, and permissiveness feeding practices); 2. children whose parents enrolled in behavior change and social support interventions will have a significantly more frequent intake of vegetables and fruits intake, and significantly less frequent intake of sugar-sweetened foods and beverages; 3. positive changes in parental feeding practices will mediate children's food intake, with increased vegetables and fruits and decreased sugar-sweetened foods and beverages intake.

A Randomized Control Trial with two groups (SmartFeeding4Kids, SmartFeeding4Kids Health) x 4 times (pre, post, 3-months, 6-month follow-up) repeated measures design will be used. Invitations to participate in the program will be made through flyers and social media messages. The invitation will show a link where parents can access detailed information about the program (study objectives, procedures, random allocation to one of the three conditions, data collected during the study). As this is a Randomized Control Trial study with repeated measurements, participants' identification must be recorded. The data relating to the parent's answers to evaluation protocols and self-monitoring, and time or interaction with the application/program, are associated with a personal account created by the participants at the beginning of the process. Consent protocol includes a paragraph that explains which data will be recorded and how it will be saved; parents must consent to researchers accessing this information. Only relevant information is collected. Data will be stored in a secure server. The team chose the server Amazon Web Services because is one of the most used services for hosting a website securely. Amazon Web Services delivers a mature set of services specifically designed for the unique security, compliance, privacy, and governance requirements of large organizations. Once registered, parents are directed to the baseline assessment protocol. After the completion of the questionnaires and 24h food recalls, parents will be randomized and allocated to one of the two conditions through an automated web-based randomization program. For both conditions, parents will be invited to participate in five sessions plus two brief booster sessions. A new session is only accessible once parents have visualized the contents of the last session and performed the tasks proposed between sessions (for experimental conditions). Although the intervention is self-guided, each session is expected to be accomplished within an average of one week. At the end of the program, parents are redirected to fill the post-intervention assessment protocol; a similar procedure is performed after 3 and 6 months. Cohen's f criteria will be used, with a small effect size of 0.15, an alpha equal to 0.05, and power of 0.80 for a two-group with four repeated measures design (with a moderate correlation pattern between time measurements of 0.35). The sample size needed is approximately 130 participants (i.e., 65 parents in each of the two-arm groups), but the total sample size was adjusted to account for a dropout rate of 50%. Therefore, the sample will be collected until 130 participants in each group (N = 260) finish the program. The proposed sample size will also be adequate for multilevel modeling with level-1 repeated measures nested within level-2 individuals, by assuring a minimum of 50 individuals required for level-2. As mentioned before, the application does not allow participants to proceed to the sessions without completing all items of the baseline evaluation protocol, nor the participants have access to a new session if they do not complete the requested between-sessions records. As such, missing data will only be due to participant's dropout, during the intervention, or during the follow-up assessments completion. The researchers will use intention-to-treat principles: participants will be analyzed in the group in which they were allocated in the randomization process, independently of whether they had completed all measurement time points and/or the intervention. An evaluation of missing data mechanisms will be performed to inform which imputation strategy should be applied between multiple imputations (MI) and maximum likelihood estimation (FIML). Assuming the data are missing at random, a sensitivity analysis will be included, considering the whole samples vs. samples with complete data separately.