An Exact Procedure for the Evaluation of Reference-Scaled Average Bioequivalence

Area of Science:

• Pharmacokinetics and Pharmaceutical Sciences
• Biostatistics
• Regulatory Science

Background:

• Reference-scaled average bioequivalence (RSABE) is recommended by regulatory agencies like the FDA and EMA for assessing bioequivalence of highly variable (HV) and narrow therapeutic index (NTI) drugs.
• The FDA suggested an approximating procedure for RSABE evaluation, prompting the need for a numerically exact alternative.
• Existing methods may have limitations in power and accuracy for specific study designs.

Purpose of the Study:

• To develop and evaluate a numerically exact algorithm, termed 'Exact', for the assessment of RSABE.
• To compare the statistical power and consumer risk of the Exact method against FDA and EMA recommended procedures using simulations.
• To derive exact, analytical formulas for confidence intervals of RSABE.

Main Methods:

• A novel algorithm, 'Exact', was developed based on the noncentral t distribution model for RSABE.
• The algorithm's parameters were derived for both crossover and parallel-group study designs.
• Simulations were conducted for HV and NTI drugs to compare the Exact method with FDA and EMA approaches, assessing power and consumer risk.

Main Results:

• The Exact method demonstrated slightly higher statistical power compared to the FDA approach.
• Consumer risks for both Exact and FDA methods were generally below nominal error risk, with exceptions in partial replicate designs under heteroscedasticity.
• The Exact method's confidence intervals are uniformly most powerful and computationally simple, though Hedges' correction is needed for RSABE estimator bias.

Conclusions:

• The Exact algorithm provides a statistically robust and computationally efficient method for RSABE assessment.
• This approach offers improved power and more accurate consumer risk evaluation for HV and NTI drugs.
• The derived analytical formulas for confidence intervals enhance the precision of bioequivalence determination.