[Assessment of stability variation among batches, packaging, and formulations--application of matrixing and bracketing].

Assessment of shelf-life equivalence among batches, packaging or formulations of pharmaceutical products, which was based on the range of shelf-life estimates, was proposed as an alternative method to an analysis of variance (ANOVA). The power of this analysis was not significantly affected by assay error, whereas that of ANOVA decreased markedly as assay error increased. Thus, ANOVA exhibites a tendency to overlook the stability variation from stability data of a larger assay error, but this is not the case for the proposed method.