Bioequivalence and Bioavailability Forum

HI balakotu,

❝ 1. Is it possible to conduct Two-stage design approach (Method C Potvin et al.) for Partial and full replicate study designs? In this type of designs how to calculate sample size for Stage-2?

Yes.

❝ 2. It is not possible to calculate Power for USA the Partial and full replicate studies in SAS Program how to calculate power for such study designs?

Correct.
Power isn't your biggest problem. You need to do a simulation yourself to make sure the overall alpha is conserved. This requires extensive programming. For actual sample size in stage 2 d_labes' package power.TOST for R will be fine. At the moment two-stage design with replcates aren't documented in the publicly available literature.

❝ 3. Whether group*treatment interaction should be statistically insignificant are not (for Europe and Canada Guidelines not mentioned whether it is required are not).

This is a nuisance effect, but in my opinion it should not be impossible to argue that there's no issue.

❝ 4. Please specify the most suitable conditions to opt Two-Stage design?

Chose 2-stage designs when you cannot reliable predict anything about T/R or the CV prior to the trial.