Title: A proper statistical framework for range-based comparisons of quality attributes
Authors: Vera Hofer - University of Graz (Austria)
Walter Goessler - University of Graz (Austria)
Gerhard Goessler - University of Graz (Austria) [presenting]
Abstract: When comparing industrial goods (e.g., drug products) with respect to quality attributes (QAs) one often has to deal not only with measurement error but also with a considerable amount of product-related variability. When product-related variability is present, it is not sufficient to just compare mean values - instead, the whole range of possible realizations must be considered. To this end, special statistical tests are needed. Looking, for example, at the respective information published by agencies like the EMA or the FDA, one can see that, unfortunately, such tests together with the proper statistical framework, are at best still under way. Therefore, we give a thorough discussion of the question of when two products can be considered equivalent with respect to a certain property when product-related variability is present. Based on these considerations, we attempt to create a suitable statistical framework for such a comparison which is centered around the novel concept of k-equivalence of two drug products with respect to a certain QA.