Sequential bioequivalence trial designs with increased power and controlled type I error rates

Area of Science:

• Biostatistics
• Pharmacokinetics
• Clinical Trial Design

Background:

• Two-stage designs for crossover bioequivalence studies are recently developed.
• Existing methods are primarily described for 80% targeted study power.

Purpose of the Study:

• To identify alpha values for two-stage bioequivalence designs targeting 90% power.
• To control type I error rates around 5% for specific test/reference ratios.
• To characterize these methods regarding sample sizes and power.

Main Methods:

• Investigated two-stage design methods for two-treatment, two-sequence, two-period crossover bioequivalence studies.
• Determined alpha sets for 90% power while controlling type I error rates at 5% for geometric mean ratios of 0.90 and 0.95.
• Analyzed resulting sample sizes and power across varying variability and stage 1 sample sizes.

Main Results:

• Actual power achieved ranges from 83% to 100%, dependent on variability and stage 1 sample size.
• Type I error inflation is minimal, with slightly less inflation observed at 90% target power using decreased stage 2 alpha.
• Identified alpha sets enable achieving 90% power in bioequivalence trials.

Conclusions:

• The developed two-stage methods effectively achieve 90% power in bioequivalence trials.
• These methods control type I error rates near 5%, offering an alternative to standard designs.
• Provides valuable options for optimizing power in bioequivalence studies without statistical penalties.