Bioequivalence and Bioavailability Forum

Dear all,

we know that the FDA uses R internally, e.g., in all simulations which lead to its reference-scaling method for NTIDs.

THX to ElMaestro discovering another example. Though I attended the 4^th GBHI workshop last month in Bethesda, it is hidden in the backup slides.* Quote:

Q1. What is a sample size to achieve 80% power to pass bioequivalence, assuming CV as 100%, T/R true ratio as 5%?
Q2. With a fixed sample size (n=1,000), what are maximum differences in FEV₁ metrics which pass BE with 80% power?

To address questions 1–2;

• package ‘PowerTOST’ was used to estimate sample size using observed CVs (R ver 3.6.0)
  
• Sample size was calculated for reference and treatment group

Of note, Robert Lionberger (Director Office of Research and Standards, Office of Generic Drugs) told me that he visits the forum regularly. 

---

• Liang Zhao (Division of Quantitative Methods and Modeling, ORS, OGD, CDER/FDA). FEV1 Based Bioequivalence Study for Inhaled Corticosteroids.