Department of Medical Statistics, University Medical Center Göttingen, Germany.
Stat Med. 2011 Jun 15;30(13):1528-40. doi: 10.1002/sim.4202. Epub 2011 Feb 22.
In recent years adaptive seamless phase II/III designs (ASDs) allowing treatment or dose selection at an interim analysis have gained much attention because of their potential to save development costs and to shorten time-to-market of a new compound compared to conventional drug development programmes with separate trials for individual phases. In this paper, we describe an ASD with treatment selection based on early outcome data, specifically considering the situation where no final outcomes are observed at the time of the interim analysis. Bringing together combination tests for adaptive designs and the closure principle for multiple testing, control of the familywise type I error rate in the strong sense is achieved. Furthermore, a simulation model is proposed based on standardized test statistics that allows the generation of virtual trials for a variety of outcomes. We use this simulation model to investigate the actual type I error rate of the proposed testing procedure and find that the familywise type I error rate is controlled as expected. The method is often conservative, with the degree of conservatism depending on the correlation between early and late outcome, the true mean values of the early outcome in the different treatment groups and the selection rule. The investigations are motivated and illustrated by an application of the proposed design and simulation model to progressive multiple sclerosis.
近年来，自适应无缝二期/三期设计（ASD）允许在中期分析时进行治疗或剂量选择，由于与单独的各期试验相比，它具有节省开发成本和缩短新药上市时间的潜力，因此引起了广泛关注。在本文中，我们描述了一种基于早期结果数据的 ASD，特别是考虑到在中期分析时没有观察到最终结果的情况。通过将自适应设计的组合检验与多重检验的闭合原理相结合，实现了强意义上的总体Ⅰ类错误率的控制。此外，还提出了一种基于标准化检验统计量的模拟模型，该模型允许针对各种结果生成虚拟试验。我们使用此模拟模型来研究所提出的检验程序的实际Ⅰ类错误率，并发现总体Ⅰ类错误率得到了预期的控制。该方法通常比较保守，保守程度取决于早期和晚期结果之间的相关性、不同治疗组中早期结果的真实平均值以及选择规则。所提出的设计和模拟模型的应用以及对进行性多发性硬化症的研究说明了这种情况。
