1997 Volume 45 Issue 9 Pages 1482-1484
An alternative method to an analysis of variance (ANOVA)for assessing stability variation among batches, packaging, or formulations is proposed. The new method assesses the shelf-life equivalence based on the range of the shelf-like estimate obtained for each product of a given batch, packaging, or formulation. The power of this analysis was not significantly affected by assay error, whereas that of ANOVA decreased markedly as assay error increased. Therefore, the proposed method does not exhibit a tendency to overlook stability variation from stability data of a larger assay error as does ANOVA. If the critical point is assumed to be 15% of the largest shelf-life estimate, 25% stability variation can be detected at a probability of more than 80% for degradation faster than 0.2%/month. When smaller stability variation is desired to be detected, the critical point must be smaller than 15%.