临床试验期中分析中不同样本空间排序方法矫正效应高估的比较
Comparison of different sample space ordering method in overestimation of efficacy correction during clinical trial interim analysis
李恒 1杨永利 1贾晓灿 1杨超君 1施学忠1
作者信息
- 1. 郑州大学公共卫生学院计算机与卫生统计学教研室 郑州 450001
- 折叠
摘要
目的:比较逐段(SW)排序法、似然比(LR)排序法和极大似然估计值(MLE)排序法对临床试验中成组序贯设计期中分析效应高估的矫正效果.方法:使用SAS 9.4 生成正态分布资料模拟数据集,模拟期中分析,采用中位无偏估计值(MUE)及其95%CI下限评估3 种样本空间排序法的矫正效果.采用一项已上市的随机对照临床试验进行验证.结果:与SW排序法相比,LR排序法和MLE排序法的MUE及其 95%CI下限更接近真实值.结论:对于正态分布数据,LR和MLE排序法能更好地矫正临床试验期中分析的效应高估.
Abstract
Aim:To compare the performance of stagewise(SW)ordering,likelihood ratio(LR)ordering,and maxi-mum likelihood estimator(MLE)ordering in correcting overestimation of efficacy in different scenarios in group sequential design interim analysis.Methods:Normal distributed quantitative simulation datasets were generated using SAS 9.4.After the simulated interim analysis,the median unbiased estimator(MUE)and its lower limit of 95%confidence interval(CI)were used to assess the performance of SW ordering,LR ordering and MLE ordering.Finally,a randomized controlled trial was used as a validation.Results:Compared with SW ordering,the MUE and its lower limit of 95%CI of LR ordering and MLE ordering were closer to the true intergroup effect.Conclusion:LR ordering and MLE ordering could obtain better cor-rection efficacy than SW ordering during clinical trial interim analysis in normally distributed data.
关键词
临床试验/期中分析/样本空间排序/效应高估矫正Key words
clinical trial/interim analysis/sample space ordering/overestimation of efficacy correction引用本文复制引用
基金项目
郑州大学青年人才企业合作创新团队支持计划(32310410)
校级教育教学改革研究与实践重点项目(2022ZZUJG063)
出版年
2024
国家科技期刊平台
NSTL
万方数据