Deimplementation of inappropriate feeding practices in early care and education: a Hybrid Type 3 cluster-randomized trial.
BACKGROUND: The science of deimplementation-reducing harmful or ineffective practices-has focused almost exclusively on clinical prescribing, with no studies conducted in community or educational settings. Early care and education (ECE) settings offer a strategic venue for shaping eating behaviors, with children consuming up to 500 meals annually in these environments. However, ECE educators routinely use feeding practices that undermine self-regulation, including pressuring children to eat, rushing mealtimes, and offering food as reward. These practices contribute to food aversions, diminished self-regulation, and obesity risk. METHODS: We will conduct a Hybrid Type 3 cluster-randomized trial evaluating a co-designed deimplementation strategy package (WISE Words) across 88 ECE sites in Arkansas and Louisiana. Sites will be randomized 1:1 to WISE Words or usual practice, with usual practice sites receiving the intervention after two years (waitlist design). WISE Words includes six strategies: dynamic training using improvisation methods, peer learning collaboratives with goal setting, external facilitation, audit and feedback, environmental reminders, and tailored educational materials. The primary outcome is de-adoption of inappropriate feeding practices measured via direct mealtime observation (Table Talk). Secondary outcomes include adoption of evidence-based practices, acceptability, appropriateness, and sustainability at 12- and 24-months post-intervention. Child outcomes include Body Mass Index, skin carotenoid levels (Veggie Meter) willingness to try new foods (observed) and food neophobia (teacher and caregiver report). An explanatory sequential mixed methods design will test mechanisms of change derived from the Implementation Trust Building Theory of Change examining whether trust mediates strategy effects on outcomes. DISCUSSION: This trial extends deimplementation science into community settings by targeting culturally embedded behavioral practices rather than clinical prescribing behaviors. Results will inform approaches to shifting entrenched practices in ECE and similar settings while testing trust as a deimplementation mechanism. Sustainability assessments will address a notable gap, as few studies have examined whether deimplementation effects persist. TRIAL REGISTRATION: NCT07101321, July 20, 2025.