Sample size determination for testing nonequality under a three-treatment two-period incomplete block crossover...

Area of Science:

• Clinical Trials
• Biostatistics
• Experimental Design

Background:

• Crossover trials are lengthy for comparing multiple treatments.
• Incomplete block designs offer a potential solution to reduce trial duration.
• Efficient sample size calculation is crucial for study feasibility.

Purpose of the Study:

• To develop a sample size calculation procedure for incomplete block crossover trials.
• To compare the efficiency of incomplete block crossover designs with parallel group designs.
• To provide a practical method for sample size determination in specific trial scenarios.

Main Methods:

• Development of a novel sample size calculation procedure.
• Utilizing Monte Carlo simulations to evaluate the procedure's accuracy.
• Application to a crossover trial comparing formoterol doses and placebo.

Main Results:

• The developed sample size procedure is accurate across various situations.
• Incomplete block crossover designs offer substantial sample size reductions versus parallel designs.
• Demonstrated practical application of the procedure in a real-world trial.

Conclusions:

• The incomplete block crossover design is an efficient alternative for multi-treatment comparisons.
• The proposed sample size calculation procedure facilitates the use of this design.
• This methodology can lead to more resource-efficient clinical trials.