Bioequivalence Design With Sampling Distribution Segments.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacokinetics

Background:

• Power analyses are crucial in bioequivalence studies to determine necessary sample sizes for detecting clinically significant differences.
• Traditional methods often rely on time-intensive computer simulations to approximate sampling distributions for complex test statistics.
• Accurate power calculations ensure efficient and ethical clinical trial conduct.

Purpose of the Study:

• To develop a novel, efficient simulation-based method for approximating power curves in bioequivalence testing.
• To enable unbiased sample size recommendations by exploring relevant segments of sampling distributions.
• To demonstrate the method's applicability to two-group bioequivalence tests with unequal variances and its potential in broader clinical designs.

Main Methods:

• A novel simulation-based approach is proposed to efficiently approximate power curves by selectively exploring sampling distributions.
• The method avoids estimating the entire sampling distribution, reducing computational time.
• Implementation is demonstrated using the 'dent' package in R for two-group bioequivalence tests with unequal variances.

Main Results:

• The proposed method provides a computationally efficient way to approximate power curves for bioequivalence tests.
• Despite not estimating the full sampling distribution, the approach yields unbiased sample size recommendations.
• The method is illustrated for two-group bioequivalence with unequal variances, showing practical utility.

Conclusions:

• A novel, efficient simulation technique accelerates power curve approximation in bioequivalence studies.
• This method facilitates accurate and unbiased sample size determination for clinical trial planning.
• The developed approach and R package ('dent') offer broader applicability in clinical study design.