欢迎访问《中国临床药理学与治疗学》杂志官方网站,今天是

中国临床药理学与治疗学 ›› 2017, Vol. 22 ›› Issue (9): 1017-1021.

• 定量药理学 • 上一篇    下一篇

国际多中心临床试验中两种一致性评价方法的比较

柏建岭1,陈梦锴1,于 浩1,赵 杨1,蔡晶晶1,黄丽红1,蒋志伟2,陈 峰1   

  1. 1南京医科大学公共卫生学院,生物统计学系,南京 211166,江苏;2默沙东(研发)中国有限公司,生物统计与决策科学部,北京 100015
  • 收稿日期:2017-02-23 修回日期:2017-04-09 出版日期:2017-09-26 发布日期:2017-09-30
  • 通讯作者: 陈峰,男,教授,博士生导师,研究方向:新药临床试验中的统计理论与方法。 E-mail: dr.chenfeng@163.com
  • 作者简介:柏建岭,男,讲师,博士,研究方向:新药临床试验中的统计理论与方法。 E-mail: jbai@njmu.edu.cn
  • 基金资助:

    国家自然科学青年基金(81302512)

Two methods for assessing consistency probability in  multiregional clinical trials: a comparison study

BAI Jianling 1, CHEN Mengkai 1, YU Hao 1, ZHAO Yang 1, CAI Jingjing 1, HUANG Lihong 1, JIANG Zhiwei 2, CHEN Feng 1   

  1. 1 Department of Biostatistics, School of Public Health, Nanjing Medical University, Nanjing, 211166 Jiangsu, China; 2 Department of Biostatistics and Decision Science, MSD R&D(China) Co. Ltd, Beijing 10015, China
  • Received:2017-02-23 Revised:2017-04-09 Online:2017-09-26 Published:2017-09-30

摘要:

目的: 探讨国际多中心临床试验中的两种一致性评价方法,为设计国际多中心临床试验提供参考。方法: 通过理论推导的方法,采用R软件编程计算样本比例和一致性概率。 结果: 在其他参数设置相同的时候,方法一中J区域疗效越好时,J区域所需要的样本比例越小。一致性概率对样本量的影响比检验效能更大。J区域所占样本比例在10%~50%时,区域数小于等于3时根据方法二得到的一致性概率均高于方法一。4个区域且J区域样本比例在10.0%~22.4%时,方法二一致性概率高于方法一;而比例在22.4%~50.0%时,方法二一致性概率低于方法一。结论: 方法一的样本量主要受把握度、一致性概率和J区域疗效。方法二还受区域数的影响,在使用方法二时,区域数不能太多。

关键词: 国际多中心临床试验, 一致性概率, 样本量

Abstract:

AIM: To evaluate two methods of consistency assessment in multiregional clinical trials so as to provide a references for the design of multiregional clinical trials. METHODS: R software was applied to compute sample size and consistency probability. RESULTS: By method one, when other parameters were set the same, the required sample size of J region was smaller as the observed treatment effect of J region was better. Compared with the power, the sample size was much more affected by consistency probability. As the size proportion of J region was 10%-50%, the consistency probabilities of method two were higher than method one with two or three regions, and directly opposite with five regions. When four regions are included, the consistency probability of method two is higher than that of method one for the fraction of 10.0%-22.4%. The consistency probability of method two was smaller than that of method one for the fraction of 22.4%-50.0%. CONCLUSION: The sample size of method one is mainly influenced by power, consistency probability and the observed treatment effect for J region. The sample size of method two is also influenced by the number of regions. When method two is used, the number of regions cannot be too much.

Key words: multiregional clinical trial, consistency probability, sample size

中图分类号: