Batch effects are the sources of variation in drug substances and drug products in terms of potency. Although the implication of batch effects has been widely recognized in statistical literature, the batch variability information is usually not considered in designing and analyzing clinical studies. In this article, the impact of batch variability is systematically explored for both the design and data analysis stages of equivalence and noninferiority clinical studies. Designing studies including more batches can increase the probability of success for demonstrating equivalence or noninferiority, while maintaining the control of Type I error. Ignoring the batch effect in the data analysis may cause marked underestimation of the variability which can lead to Type I error inflation. To achieve a desired precision of the treatment estimate, a formula was provided to select appropriate number of batches. The datasets from a phase I oncology study, and a phase III vaccine study are used to illustrate the importance of considering batch effects.

批次效应是药效方面药物和药品变化的根源。尽管批处理效应的含义已在统计文献中得到广泛认可，但是在设计和分析临床研究中通常不考虑批处理变异性信息。在本文中，系统地探讨了批次变异性对等效性和非劣效性临床研究的设计和数据分析阶段的影响。设计包括更多批次的研究可以增加证明同等或非劣等成功的可能性，同时保持对I型错误的控制。忽略数据分析中的批处理效果可能会导致对可变性的明显低估，这可能导致I型错误膨胀。为了达到所需的治疗估算精度，提供了一个公式来选择适当的批次数。I期肿瘤学研究和III期疫苗研究的数据集用于说明考虑批次效应的重要性。