Exact sample-size determination in testing non-inferiority under a simple crossover trial

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Power Analysis

Background:

• Odds ratios (ORs) are recommended for assessing relative treatment efficacy in non-inferiority trials.
• Accurate sample size calculation is crucial for reliable clinical trial outcomes.

Purpose of the Study:

• To develop an exact sample-size calculation procedure for testing treatment non-inferiority based on odds ratios.
• To compare the exact procedure with asymptotic methods for sample size determination.

Main Methods:

• The study extends an existing exact test for a simple crossover design.
• A novel exact sample-size calculation procedure is derived for odds ratios of patient response rates.
• The procedure considers desired statistical power and nominal type I error rates.

Main Results:

• The exact sample-size calculation procedure is presented.
• Calculated sample sizes using asymptotic tests can be substantially smaller than those from the exact test.
• Advantages and disadvantages of both exact and asymptotic sample-size calculation methods are discussed.

Conclusions:

• The developed exact procedure provides a reliable method for sample-size calculation in non-inferiority trials.
• Clinicians should be aware of potential underestimation of sample size when using asymptotic tests.
• The study illustrates the procedure with an example involving inhalation devices for asthmatics.