NOTE: Page 7 of this article has been revised to correct an error in the original article which stated that “..the probability of failing a lot when it actually is not equivalent, becomes 1-2*α.” The correct equation, 1-α, has replaced the errant one and a single sentence following it was added to further clarify this concept.
ABSTRACT: Pre-clinical and clinical trials conducted to establish the minimum effective dose and the maximum tolerated dose of a viral vector assume that the assigned dose values are comparable across studies. Toxicity has been associated with high dose administration of both adenovirus and adeno-associated virus-based vectors, and increased attention must be paid to assays used to measure dose. High assay variability can be mitigated by replication and the reporting of a mean value for product lot release. The establishment of a dose specification and a testing strategy must take into account the risk of errant quality control decisions. This can be accomplished by linking assay qualification information to measurement uncertainty through a statistical framework. By adopting an equivalence approach, the risk of releasing lots with unacceptably high or low dose values is minimized by reducing measurement uncertainty. This article provides a worked-through example to introduce applicable statistical concepts and the equations necessary to facilitate their implementation in the field.