Ⅱ/Ⅲ期无缝临床试验中合并检验法与改良法的比较

The Comparison of Combination Test Approach and Modified Test Approach in Seamless Phase Ⅱ/Ⅲ Trial

目的研究Ⅱ/Ⅲ期无缝设计中最常用的合并检验法与最新提出的改良法的统计学特性,推荐选择较优的分析方法。方法系统地回顾无缝设计的基本设计思路和统计分析方法,采用计算机模拟技术,分别研究两种统计分析方法的总I类错误和检验效能。结果合并检验法和改良法相比,两者在控制总I类错误方面无明显差异,但在检验效能方面,合并检验法始终优于改良法。结论建议优先选择合并检验法进行Ⅱ/Ⅲ期无缝临床试验的统计分析。

Objective To compare the combination test approach,the most commonly used method and the recently proposed modified test approach in the analysis of seamless phase Ⅱ / Ⅲ clinical trials. Methods The basic principle of seamless design and two statistical methods were reviewed. Simulation studies were then used to understand the power and the familywise type I error rate in different scenarios. Results There is no significant difference between combination test approach and modified test approach in terms of controlling family-wise type I error rate. How ever,combination test approach always shows higher power. Conclusion When conducting statistical analysis of a seamless phase Ⅱ / Ⅲ trial design,combination test approach has the priority.