How many times should a cluster randomized crossover trial cross over?

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Experimental Design

Background:

• Cluster randomized crossover trials involve subjects switching between control and intervention conditions.
• Determining the optimal number of crossovers is crucial for balancing efficiency, complexity, and cost.

Purpose of the Study:

• To determine the optimal number of crossovers for a fixed trial duration in cluster randomized crossover trials.
• To identify efficient designs by varying cluster numbers and considering cost and power constraints.

Main Methods:

• Utilized linear mixed models for treatment effect estimation with continuous outcomes and recruitment.
• Modeled within-cluster outcome similarity using a correlation structure that decays over time.
• Evaluated efficiency gains across different numbers of crossover periods.

Main Results:

• Optimal designs theoretically involve crossovers with each successive subject, which is impractical.
• Significant efficiency gains are observed when moving from two to four periods.
• Further increases in crossovers beyond four periods yield diminishing returns, approaching an asymptote.

Conclusions:

• While more frequent crossovers can increase efficiency in cluster randomized crossover trials, the gains are most substantial up to four periods.
• Designs with more than four periods offer limited additional benefits and may not be cost-effective or practical.
• Four-period designs represent a practical and efficient choice for many cluster randomized crossover trials.