多阶段随机化胜者优先设计的计算机模拟试验分析

中国临床药理学与治疗学 ›› 2010, Vol. 15 ›› Issue (8): 888-893.

多阶段随机化胜者优先设计的计算机模拟试验分析

周敏林, 魏永越, 张汝阳, 朱晶晶, 于浩, 陈峰

南京医科大学公共卫生学院流行病与卫生统计学系,南京 210029,江苏

收稿日期:2010-07-02 修回日期:2010-07-30 出版日期:2010-08-26 发布日期:2020-09-17
通讯作者: 陈峰,男,博士,教授,博士生导师,研究方向:生物统计理论与方法。Tel: 025-86862754 E-mail: fengchen@njmu.edu.cn
作者简介:周敏林,女,硕士,研究方向:药物经济学。Tel: 13621599942 E-mail: zminlin@163.com
基金资助:
江苏省自然科学基金“临床试验中的优化设计与统计分析方法研究”资助(BK2008449)

Simulation study of multi-stage randomized-play-the-winner-rule

ZHOU Min-lin, WEI Yong-yue, ZHANG Ru-yang, ZHU Jin-jin, YU Hao, CHEN Feng

Department of Epidemiology and Biostatistics, School of Public Health, Nanjing Medical University, Nanjing 210029, Jiangsu, China

Received:2010-07-02 Revised:2010-07-30 Online:2010-08-26 Published:2020-09-17

摘要/Abstract

摘要： 目的: 本文介绍一种新的多阶段随机化胜者优先原则,以便自适应随机分组方法在临床试验中的应用。方法: 采用计算机模拟试验,探讨传统随机分组方法、改良随机化胜者优先原则及多阶段随机化胜者优先原则对检验效能及平均减少失败人数的影响。结果: 尽管改良随机化胜者优先原则和多阶段随机化胜者优先原则的检验效能低于传统方法,但相差不大。改良随机化胜者优先原则对于平均减少失败人数的调整力度最强,随着调整分组概率时间间隔的延长,多阶段随机化胜者优先原则对于平均减少失败人数的改变逐渐减弱。在不同率及率差水平下,平均减少失败人数呈现一定的趋势:率差较小时,成功率影响较大,时间间隔影响较小;率差较大时,成功率影响较小,时间间隔影响较大。结论: 多阶段随机化胜者优先原则兼顾了改良随机化胜者优先原则的优点,同时扩展了改良随机化胜者优先原则的实际可操作性。但在组间率差较大时,使用多阶段随机化胜者优先原则要选择合理的时间间隔。

关键词: 延迟效应, 随机化胜者优先原则, 瓮模型, 响应-自适应随机化设计

Abstract: AIM: To present an adaptive randomized method, multi-stage randomized-play-the-winner rule (MSRPW), which facilitates the application of adaptive randomization in clinical trial.METHODS: By using the simulation study, we compared the powers and the mean reduced failures of three different methods, those were traditional equal-probability random allocation, modified randomized-play-the-winner rule and multi-stage randomized-play-the-winner rule.RESULTS: Compared with traditional allocation method, both MRPW and MSRPW could reduced failures in clinical trials, although their powers were decreased slightly. The MRPW reduced the most failures in all scenarios, MSRPW performed as well as MRPW in the scenarios when the difference of effects of test and control groups was not large, or the time interval of adjusting allocation probability was not long, while the traditional allocation performed worst. With the time interval prolonging, MSRPW reduced fewer and fewer failures than MRPW.CONCLUSION: The MSRPW rule proposed in this paper takes both advantages of traditional randomization and MRPW rule into consideration, and more practical than MRPW in clinical trials. MSRPW can be very useful in those scenarios when the difference of effects of the test and control groups is moderate or small, while an appropriate time interval should be chosen carefully when the difference of effects is large.

Key words: Delayed response , Randomized-play-the-winner rule , Urn model , Response-adaptive randomization

中图分类号:

R311

周敏林, 魏永越, 张汝阳, 朱晶晶, 于浩, 陈峰. 多阶段随机化胜者优先设计的计算机模拟试验分析[J]. 中国临床药理学与治疗学, 2010, 15(8): 888-893.

ZHOU Min-lin, WEI Yong-yue, ZHANG Ru-yang, ZHU Jin-jin, YU Hao, CHEN Feng. Simulation study of multi-stage randomized-play-the-winner-rule[J]. Chinese Journal of Clinical Pharmacology and Therapeutics, 2010, 15(8): 888-893.

参考文献

[1] Robbins H. Some aspects of the sequential design of experiments[J]. Bull Amer Math Soc, 1952,58:527-535.
[2] Gallo P, Chuang SC, Dragalin V, et al. Adaptive design in clinical drug development-an executive summary of the PhRMA Working Group (with discussions)[J]. J Biopharm Stat, 2006,16(3):275-283.
[3] Chow SC, Chang M. Adaptive design methods in clinical trials-a review[J]. Orphanet J Rare Dis, 2008, 3:1-13.
[4] Paulo G, Yuko P. Power and sample size simulations for randomized play-the-winner rules[J]. Contemp Clin Trials, 2007, 28(4): 487-499.
[5] Hu F, Rosenberger WF. The theory of response-adaptive randomization in clinical trials[M]. Wiley, 2006:68.
[6] Zelen M. Play the winner rule and the controlled clinical trials[J]. J Am Statist Assoc, 1969, 64(325):131-146.
[7] Wei LJ, Durham S. The randomized play-the-winner rule in medical trials[J]. J Am Statist Assoc, 1978, 73:840-843.
[8] William FR. New directions in adaptive designs[J]. Stat Sci,1996, 11:137-149.
[9] 王素珍, 夏结来, 郑亮, 等. 临床试验中自适应设计的urn模型和RPW规则及其模拟[J]. 第四军医大学学报, 2007, 28 (24): 2284-2286.
[10] Morgan CC, Coad DS. A comparison of adaptive allocation rules for group-sequential binary response clinical trials[J]. Stat Med, 2007, 26 (9) :1937-1954.
[11] 刘晓燕, 陈峰, 于浩, 等. 响应-自适应随机化分组方法[J]. 临床药理学与治疗学, 2008, 13(6):684-689.
[12] 于莉莉, 薛富波, 王素珍, 等. 临床试验适应性设计的反应变量-适应性随机化方法简介[J]. 中国卫生统计, 2008, 25(5): 534-538.
[13] Yao Q, Wei LJ. Play the winner for phase II/III clinical trials[J]. Stat Med,1996, 15 (21/22):2413-2423.
[14] Fleiss JL. Statistical methods for rates and proportions[M]. 3rd Edition. New York: John Wiley, 1981:64-83.
[15] Williams MS, Ebel ED, Wager BA. Monte Carlo approaches for determining power and sample size in low-prevalence applications[J]. Prev Vet Med, 2007, 82 (1/2):157-158.