|
Abstract:
|
One advantage to use the Two-One-Sided-Tests Procedure according to Schuirmann for the assessment of the similarity of two laboratories in an analytical methods transfer is that irrelevant differences are accepted to be negligible. Since the irrelevant difference (or equivalence range) has to be defined before the transfer testing starts, it is always difficult to determine. This talk presents an approach to derive an equivalence range by taking testing specification limits, production variability, and acceptance probabilities for later (e.g., release testing procedures into consideration). Thus, statisticians can support a study director to define an equivalence range in order to control the future risk of rejecting a good quality batch.
|
