Two-stage designs for cross-over bioequivalence trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacokinetics

Background:

• Two-stage designs are increasingly relevant for bioequivalence (BE) studies, as noted in recent literature and regulatory guidance.
• Existing research on sequential designs in BE is limited, and adaptive approaches have primarily been used in superiority/non-inferiority trials.
• A direct comparison of the performance characteristics of different two-stage designs for BE studies is lacking, leaving the optimal choice unclear.

Purpose of the Study:

• To discuss and compare classical group sequential designs with three types of adaptive designs for bioequivalence studies.
• To evaluate the performance characteristics, including power and sample size, of these different two-stage designs.
• To provide guidance on selecting the most appropriate design for bioequivalence trials.

Main Methods:

• Comparative analysis of 'classical' group sequential designs and three adaptive designs with mid-course sample size recalculation.
• Comprehensive simulation study to assess design performance.
• Illustration of methods using a real-world bioequivalence study example.

Main Results:

• Group sequential designs can achieve power comparable to adaptive designs in bioequivalence studies.
• Adaptive designs with mid-course sample size recalculation require a higher average sample size compared to identified group sequential designs.
• The study identified specific group sequential designs that are efficient for bioequivalence applications.

Conclusions:

• Group sequential designs represent a viable and potentially more sample-size-efficient alternative to adaptive designs for bioequivalence studies.
• The findings suggest that careful selection of group sequential designs can match the performance of adaptive designs with reduced sample size requirements.
• This research aids in optimizing clinical trial design for bioequivalence assessments.