Designing stepped wedge trials to evaluate physical activity interventions in schools: methodological considerations.

Area of Science:

• Public Health
• Epidemiology
• Biostatistics

Background:

• Stepped wedge designs offer advantages in statistical power and logistical efficiency for intervention studies.
• However, their application in evaluating school-based physical activity interventions for children has been limited.
• This study addresses the feasibility of using stepped wedge designs in this context.

Purpose of the Study:

• To explore the feasibility of stepped wedge designs for school-based physical activity interventions.
• To identify school constraints influencing design configuration.
• To balance statistical considerations with practical school limitations and assess design power.

Main Methods:

• A three-study approach integrating qualitative methods, simulation, and power analysis.
• Study 1: Qualitative exploration of school constraints to inform stepped wedge configuration.
• Study 2: Simulation to optimize design balancing school needs and statistical properties.
• Study 3: Power analysis for an open cohort design under various sample sizes.

Main Results:

• School staff found data collection feasible with 3-4 measurements/year and a one-term implementation period.
• Incomplete stepped wedge designs with five steps were considered, favoring mixed control/intervention measurements and spread across the study.
• High power (>80% to detect a 6-min physical activity difference) was achievable with >50% pupil retention in 20 schools, compared to 42 schools for a cluster randomized trial.

Conclusions:

• Stepped wedge trials represent a viable design for school-based physical activity intervention evaluations.
• Incomplete stepped wedge designs provide essential flexibility to accommodate practical school constraints.
• This design offers a statistically powerful and logistically efficient alternative to traditional cluster randomized trials.