More than 60% of preschool children in the US eat one or more of their daily meals in a childcare setting. Although there are childcare centers continuing to provide meals to children, a growing trend is for centers to discontinue their meal program and require that parents provide sack lunches for their children. Preliminary data indicate these packed lunches are low in servings of fruits, vegetables and whole grains. With the growing issue of child overweight, consumption of meals low in high nutrient foods as recommended in the Dietary Guidelines can lead to the development of poor eating habits that can persist over time. The primary aim of this study is to revise, implement and evaluate a behaviorally-based intervention, the Lunch Box Program, to increase the servings of fruits, vegetables and whole grains parents pack in preschool child lunches. This study will be implemented in three phases. Phase 1 (months 1 through 6 of the project) includes a planning meeting with a select advisory group as well focus groups with child care personnel and focus groups with parents to adapt the Lunch Box Program (LBP) with a behavioral focus on fruits, vegetables, whole grains as well as editing and evaluation of tools to measure program-specific behavioral outcomes (e.g., parents'expectations and intentions for packing more fruits, vegetables and whole grains in the lunches). Phase 2 (months 6 through 10) includes focus testing and finalizing adaptations to the LBP as well as recruitment of families to participate in a pilot of the adapted LBP. In Phase 3 (months 11 through 19), a pilot of the adapted LBP will be conducted with the participation of 40 families at 3 childcare centers randomly assigned to intervention and 40 families at 3 centers assigned to comparison. The intervention will be conducted over a 6 week period, with pre- and post-tests consisting of observations of the sack lunches and parent questionnaires. Process data also will be collected to evaluate dose, fidelity and feasibility of the intervention for sustainable implementation in childcare settings. Results from the pilot will be used to determine Interclass Correlation Coefficient for detectable effect computations to power an R01 with a group randomized design using childcare center as the unit of randomization and child's parent as the unit of analysis.
