Analytical method transfer: Improving interpretability with ratio-based statistical approaches

Design and Statistical Analysis of Method Transfer Studies for Biotechnology Products

“…Samples of commercial production batches are recommended to be used [5,[8][9][10][11][12][13] as testing materials. Samples from at least one batch are recommended to be tested [4,[9][10][11][12].With regard to the study designs, method transfer studies should evaluate the impact of factors including analysts, days, number of runs, equipment (brand and/or model), environment and reagent suppliers on the results obtained in the sending and receiving laboratories, using the same testing materials. Wieling proposed to keep the methods, chemicals, reagents, disposables, glassware, equipment (brand and model) and data reduction as identical as possible [6].…”

“…The performance of this approach depends on the acceptance limits ([1 -γ]μ s , [1 + γ]μ s ) and prespecified β. However, no discussion is provided on how to determine β and γ.Frömke et al [11] proposed splitting one sample into two parts (one analyzed by the sending laboratory and the other analyzed by the receiving laboratory) and conducting the equivalence test on the ratio of means obtained in two laboratories, and performing a noninferiority test on the ratio of the variances obtained in two laboratories. No discussion is provided on the equivalence margin and the noninferiority margin.…”

“…Several publications emphasize that the data should be paired for each of the multiple batches to evaluate the equivalence of the results between two laboratories, for example [11,12], in which case, many batches are needed.…”

“…With regard to statistical analysis methods, although it has been commonly accepted that the significance test should be avoided for the comparative data in method transfer studies, several statistical analysis approaches, including the equivalence test, tolerance interval method and total error method [3,6,[8][9][10][11]13] have been proposed. However, these proposals did not discuss on how to establish the acceptance criteria.…”

“…While the US FDA guidance to industry [1,2] and further publications [3][4][5][6][7][8][9][10][11][12][13][14][15] discuss the general principles for the design, analysis and evaluation of method transfer studies, they do not explicitly specify the acceptance criteria by which assay transfers are considered acceptable. Although these general principles apply to the majority of assay transfer studies, depending on the type and the intended purpose of the assays, different statistical analysis methods have been proposed to suit specific needs.…”