Bioequivalence and Bioavailability Forum

dear all!

I have a questions regarding testing for equivalence and acceptance criteria for lot releases. In these kind of studies several lots (test) and one reference lot are investigated in-vitro each with replicated measurements (typically 3 replicates per lot are available).

For an overall statement (i.e. combined lots) the R package mratios may be used to model the lot and potential day and/or operator effects assuming normal distribution and independent replicates and use corresponding contrasts to get the ratio and corresponding CIs for an overall statement. However, I do not really understand the comparison of individual lots with the reference lot. In statistics, CIs are used get an idea regarding the population average. By comparing individual lots with the reference lot the statistical unit reflects only the assay variability.

I would be happy if you can help me in answering the following questions:

1. Is anyone aware of acceptance criteria for these kind of evaluations (I think the scaled average approach may be also applicable as the CV of the assay is known and a parallel group design is adequate).
   
2. What is the interpretation (i.e. statement regarding the population) when equivalence is shown (for a given acceptance criteria) for a specific test lot compared to the reference lot.

best regards

martin