An approach for sample size determination of average bioequivalence based on interval estimation

Area of Science:

• Pharmacokinetics
• Biostatistics
• Drug Development

Background:

• Average bioequivalence (ABE) is crucial for drug approval, typically assessed using crossover designs.
• Long drug half-lives necessitate alternative study designs, such as parallel designs, for bioequivalence assessment.
• Current methods for parallel bioequivalence studies may lack robust sample size calculation strategies.

Purpose of the Study:

• To develop and validate a method for assessing parallel average bioequivalence (ABE).
• To establish a sample size calculation approach for parallel bioequivalence studies based on asymptotic power.
• To provide a practical method for determining bioequivalence in parallel drug studies.

Main Methods:

• Utilized two-sided interval estimation methods (Satterthwaite's, Cochran-Cox's, Howe's approximations) for parallel ABE assessment.
• Derived the asymptotic joint distribution of confidence interval limits as bivariate normal.
• Developed a sample size calculation method based on asymptotic power and the target ABE interval (-0.223, 0.223).

Main Results:

• The proposed method allows for sample size calculation to ensure confidence intervals fall within the bioequivalence range.
• Simulation studies confirmed that the method achieves sufficient empirical power for parallel bioequivalence assessment.
• A real-world example demonstrated the practical application of the proposed method.

Conclusions:

• The developed method provides a statistically sound approach for parallel bioequivalence studies.
• Accurate sample size determination is crucial for reliable bioequivalence assessment in parallel designs.
• This methodology enhances the evaluation of drug product equivalence when crossover designs are not feasible.