Food, Health and Choices Implementation and Context: The Case for a Comprehensive Approach to Process Evaluation in School-Based Childhood Obesity Prevention Trials

Childhood obesity is an important societal problem for reasons of justice, economy, and well-being; therefore, significant resources are expended on childhood obesity prevention interventions. We need strong evaluations to ensure that we use these resources efficiently and effectively. While randomized controlled trials are focused on internal validity, process evaluations can be focused on more generative evaluation questions and can be a powerful compliment to a randomized controlled trial if they employ methodological pluralism and methodological rigor. In this dissertation, a systematic framework for a Comprehensive Approach to Process Evaluation is presented. The Comprehensive Approach to Process Evaluation was developed based on a review of 17 process evaluation and similar studies and emphasizes the examination of both implementation and contextual factors together in a process evaluation. The Comprehensive Approach to Process Evaluation is applied in this dissertation to the Food, Health & Choices childhood obesity prevention cluster randomized trial, which was implemented in 20 high-need NYC public elementary schools and included 1,358 students in 55 5th grade classes. The Comprehensive Approach to Process Evaluation was used to evaluate Food, Health & Choices in three articles. The first used hierarchical linear modeling to examine relationships among implementation factors, contextual factors, and behavioral outcomes, while accounting for the clustering of these variables at both the classroom and school levels. Important findings were that student reception of the intervention was related to outcome behaviors; class engagement was related to curriculum recall; and teacher interest was related to delivery of the wellness intervention, the only aspect of Food, Health & Choices that depended on classroom teachers for implementation. Additionally, student responses to behavior items on questionnaires were grouped as three logical factors, which not only increased statistical power but also provided insight into how students think about energy balance related behavior. The second article used mixed methods to examine the relationship between intervention delivery and reception, energy balance related behavior, and intrapersonal and external contexts by comparing classrooms with high and low buy-in based on measures of classroom context. This comparison led to the finding that there were quantitative differences in energy balance related behavior in high buy-in and low buy-in classes at baseline and posttest, but no other constructs. A qualitative comparison of student perceptions of supports and barriers to maintaining energy balance uncovered neighborhood environment as a universal barrier, preferences as a more salient barrier for students in low buy-in classes, and more awareness of the home environment as both a support and barrier among students in high buy-in classes. The third article used qualitative methods to examine the relationship between energy balance related behavior and intrapersonal and external contexts. Findings were that four patterns of behavior adoption existed among the interviewees and that corresponding intrapersonal and external contextual factors facilitated and hindered energy balance related behavior adoption in each of these groups. In all, the Comprehensive Approach to Process Evaluation proved to be a useful and rigorous compliment to the Food, Health & Choices randomized controlled trial outcome evaluation and provided a more nuanced explanation of what happened during Food, Health & Choices. A synthesis of findings across the three studies highlights both implementation and contextual factors that were important during the intervention, including: 1) Intervention reception – satisfaction and recall of Food, Health & Choices were associated with energy balance related behaviors; 2) Intervention delivery – using trained nutrition educators seemed to positively influence the completion of the intervention; 3) Intrapersonal context – students responded differently to the intervention based on prior experiences and personality; 4) External context – neighborhood environments were powerful in influencing energy balance related behaviors. This has implications for intervention sustainability and improvement as well as research, practice, and policy in childhood obesity prevention. Research should be conducted to clarify the relationship between intervention satisfaction and behavioral outcomes; practice should consider the value of trained nutrition educators versus classroom teacher professional development, the importance of parent education about exposure to fruits and vegetables in early childhood, and how interventions might be tailored for different student energy balance related behavior adoption profiles; and policies that support children in choosing healthy food outside of school should be adopted. The Comprehensive Approach to Process Evaluation should be tested in other interventions to determine if it is recommended as an approach to process evaluation across school-based childhood obesity prevention interventions