Type-II generalized family-wise error rate formulas with application to sample size determination

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Power Analysis

Background:

• Multiple endpoints are common in clinical trials, necessitating advanced statistical methods.
• Traditional family-wise error rate (FWER) control may not be optimal for all trial designs.
• Generalized error rates offer more flexibility in controlling false rejections.

Purpose of the Study:

• To define and implement a type-II r-generalized family-wise error rate for multiple hypothesis testing.
• To develop general power formulas for sample size computation in clinical trials with multiple endpoints.
• To provide an accessible R package for applying these statistical methods.

Main Methods:

• Definition of a type-II r-generalized family-wise error rate.
• Derivation of general power formulas for single-step and step-wise testing procedures.
• Implementation of formulas in the R package 'rPowerSampleSize'.

Main Results:

• General power formulas applicable to various multiple testing procedures.
• The 'rPowerSampleSize' R package is available on CRAN for practical use.
• Sample size calculations demonstrated for acute heart failure drug trial and pneumococcal vaccine immunogenicity study.

Conclusions:

• The developed methods and R package facilitate accurate sample size determination for complex clinical trials.
• The approach allows for control of generalized error rates, enhancing statistical rigor.
• Practical application in real-world clinical trial scenarios is demonstrated.