Sample Size Calculation for “Gold-Standard” Noninferiority Trials With Fixed Margins and Negative Binomial Endpoints

Abstract

For clinical trials in patients with asthma, chronic obstructive pulmonary disease, and relapsing-remitting multiple sclerosis, regulatory guidelines state that a noninferiority clinical trial is an option. Aiming at establishing assay sensitivity, a three-arm noninferiority trial, including an experimental treatment, a reference treatment, and a placebo, the so-called gold-standard noninferiority trial, is recommended. In addition, in the clinical trials for the above diseases, the number of events per unit time is often used as a primary or key secondary endpoint. There are two issues to be addressed in determining the sample size for gold-standard noninferiority trials: the hypotheses to be tested for assay sensitivity, and overdispersion of the number of events. To address these two issues, we propose a sample size calculation method for gold-standard noninferiority trials with two fixed margins and negative binomial endpoints. In a simulation, we show that the proposed hypothesis-testing method achieves the target power, while controlling the Type I error rate at a nominal level, and that the sample size calculation method has adequate power accuracy. Moreover, we illustrate an application of our proposed method for a clinical trial of mild asthma.

Keywords:

• Assay sensitivity
• Chronic obstructive pulmonary disease
• Relapsing-remitting multiple sclerosis
• Single-margin procedure
• Two-margin procedure

Supplementary Materials

The code for the R software to implement our proposed sample size calculation method is provided as supporting information, and can be obtained from the publishers’ website.

Acknowledgments

We would like to thank the associate editor and the anonymous reviewers for helpful comments that improved the article.