van der Voet Hilko, Goedhart Paul W, Schmidt Kerstin
Wageningen University & Research, Biometris, Droevendaalsesteeg 1, 6708PB Wageningen, The Netherlands.
BioMath GmbH, Friedrich-Barnewitz-Str. 8, 18119 Rostock-Warnemünde, Germany.
Food Chem Toxicol. 2017 Nov;109(Pt 1):472-485. doi: 10.1016/j.fct.2017.09.044. Epub 2017 Sep 25.
An equivalence testing method is described to assess the safety of regulated products using relevant data obtained in historical studies with assumedly safe reference products. The method is illustrated using data from a series of animal feeding studies with genetically modified and reference maize varieties. Several criteria for quantifying equivalence are discussed, and study-corrected distribution-wise equivalence is selected as being appropriate for the example case study. An equivalence test is proposed based on a high probability of declaring equivalence in a simplified situation, where there is no between-group variation, where the historical and current studies have the same residual variance, and where the current study is assumed to have a sample size as set by a regulator. The method makes use of generalized fiducial inference methods to integrate uncertainties from both the historical and the current data.
描述了一种等效性测试方法，该方法使用在对假定安全的参比产品进行的历史研究中获得的相关数据来评估受监管产品的安全性。使用来自一系列转基因玉米品种和参比玉米品种动物饲养研究的数据对该方法进行了说明。讨论了几种量化等效性的标准，并选择研究校正的分布等效性作为适用于示例案例研究的标准。基于在简化情况下声明等效性的高概率提出了一种等效性测试，在这种简化情况下，组间无差异，历史研究和当前研究具有相同的残余方差，并且假定当前研究具有监管机构设定的样本量。该方法利用广义置信推断方法来整合来自历史数据和当前数据的不确定性。
