Sample size determination for equivalence assessment with multiple endpoints

Area of Science:

• Biostatistics
• Pharmacometrics
• Clinical Trial Design

Background:

• Equivalence assessment between reference and test treatments commonly uses two one-sided tests (TOST).
• Sample size determination for equivalence trials relies on the joint distribution of sample mean and variance.
• Multi-endpoint equivalence trials present challenges in sample size calculation, especially when endpoints are correlated.

Purpose of the Study:

• To propose an exact power function for equivalence tests with multiple, correlated endpoints.
• To adjust sample size determination for endpoint correlations in both crossover and parallel designs.
• To compare the proposed method with a naive approach ignoring correlations.

Main Methods:

• Development of an exact power function for multi-endpoint equivalence tests.
• Application of the power function to both crossover and parallel study designs.
• Correlation adjustment for sample size determination in the presence of multiple endpoints.

Main Results:

• The proposed exact power function accurately reflects power for correlated endpoints in multi-endpoint equivalence tests.
• Correlation-adjusted sample size calculations are more efficient than naive methods that ignore correlations.
• Illustrative example using a bioequivalence study with Area Under the Curve (AUC) and Maximum Concentration (Cmax) endpoints.

Conclusions:

• Accounting for endpoint correlations is crucial for accurate sample size determination in multi-endpoint equivalence trials.
• The proposed method provides a more precise and efficient approach to sample size calculation.
• This methodology enhances the design and analysis of bioequivalence studies with multiple endpoints.