Equivalence tests for interchangeability based on two one-sided probabilities.

Area of Science:

• Biostatistics
• Pharmacokinetics
• Drug Development

Background:

• Interchangeability of drug treatments relies on demonstrating bioequivalence.
• Current FDA guidance suggests individual bioequivalence (IBE) and population bioequivalence (PBE) methods.
• Limitations exist in IBE and PBE regarding average bioequivalence and asymmetric margins.

Purpose of the Study:

• To reexamine and extend tolerance interval approaches for drug interchangeability testing.
• To address limitations of existing methods, particularly concerning the two one-sided tests (TOST) with asymmetric margins.
• To develop a robust statistical framework for assessing interchangeability in various trial designs.

Main Methods:

• Extension of Tsong and Shen's (2007, 2008) two one-sided tolerance interval approaches.
• Application to parallel arms trials and paired/crossover data, relaxing assumptions of equal sample sizes and variances.
• Development of exact power function and Type I error rate assessment.

Main Results:

• The proposed method extends tolerance interval approaches for interchangeability testing.
• The approach accommodates parallel, paired, and crossover trial designs with unequal sample sizes and variances.
• Exact power function and Type I error rates were developed and assessed.

Conclusions:

• The refined tolerance interval method offers a more robust approach to drug interchangeability assessment.
• This methodology provides improved statistical power and accurate sample size determination for bioequivalence studies.
• The extended approach enhances the reliability of determining treatment interchangeability in pharmaceutical development.