节点文献
基于因子正交旋转的非独立多终点校正法
An easy-to-use multiplicity adjustment based on factor orthogonal rotation for correlated multiple endpoints in clinical trials
【Author】 Ling Wang,Chanjuan Li,Zhiwei Jiang,Yongji Wang,Jielai Xia Department of Health Statistics,Faculty of Preventative Medicine, The Fourth Military Medical University,Shannxi Xi’an,710032
【机构】 第四军医大学卫生统计学教研室；
【摘要】 目的探讨在新药临床试验中非独立多终点设计对疗效评价的影响并提出一种新的校正方法。方法通过对现有方法的改进提出一种基于因子正交旋转的、适合于非独立多终点的FR校正法,并通过Monte Carlo模拟实验比较新方法与现有方法在控制总Ⅰ类错误率和检验效能方面的表现。结果对于非独立的连续型多终点和两分类多终点,FR法在多终点间不同的相关程度下均很好的控制了试验总Ⅰ类错误率并具有足够的检验效能,较之现有的校正方法具有精确和实用的双重优势。结论 FR法被初步证实是一种专门校正非独立多终点的有效方法,有良好的应用前景。
【Abstract】 Objective To explore the efficacy evaluations of clinical trials with multiple correlated endpoints and propose a new multiplicity adjustment.Methods Especially for the adjustment of correlated endpoints,FR adjustment is modified from Sidak adjustment and based on factor orthogonal rotation.The FR adjustment was compared with other popular procedures for correlated continuous or binary endpoints by Monte Carlo simulations.Results Simulation results showed that the FR adjustment was an effective method for multiple correlated endpoints,by which the Family-Wise type I Error Rate (FWER) were controlled well at the level of 0.05 in the range[0,1) of correlation coefficient and the power was adequate.Conclusions Because of the excellent performance with respect to FWER and power,the objective adjustment to all endpoints with equal weight,the distribution free of endpoint and the concise expression,FR method is a good alternative of present popular adjustments for multiple correlated endpoints.
【Key words】 Multiple correlated endpoints； Family-wise type I error rate； Power； Monte Carlo simulation；
- 【会议录名称】 2011年中国卫生统计学年会会议论文集
- 【会议名称】2011年中国卫生统计学年会
- 【会议时间】2011-07-27
- 【会议地点】中国陕西西安
- 【分类号】R96
- 【主办单位】中国卫生信息学会统计理论与方法专业委员会、中华预防医学会卫生统计专业委员会