Trial Probability of Success for Testing 3-Way PK/PD Similarity With Multiple Endpoints

Pharmacokinetics and pharmacodynamics (PK/PD) similarity trials typically involve multiple coprimary endpoints and a 3-way treatment comparison. The purpose of these trials is to demonstrate the similarity between a biosimilar candidate and two versions of the originator drug. The sample size for these trials is often based on point estimates of the expected treatment difference and/or variability, derived from historical reference data, without considering the uncertainty associated with these estimates. This uncertainty, especially when there are multiple comparisons, can lead to an unreliable estimate of study power. In this paper, we address the power and application of the assurance method in PK/PD similarity studies to account for the uncertainty surrounding treatment differences and/or variability in multiple coprimary endpoints when considering sample size. We introduce an assurance method that can handle multiple comparisons and propose a strategy to elicit joint prior distributions of parameters based on the availability of historical data. These methods are implemented in an R shiny app using the Monte Carlo method. Additionally, we provide a real data example to illustrate the practical application of these methods. Our findings demonstrate that the proposed methods significantly enhance our understanding of study power. Therefore, we recommend incorporating assurance methods as a complement to conditional power in sample size considerations.