Sequential Designs for Clinical Trials

Sequential Designs for Clinical Trials
复制标题

临床试验的序贯设计

DOI:
10.1007/978-1-4614-0140-7_3
复制
发表时间:
2012
影响因子:
1.7
通讯作者:
KyungMann Kim
KyungMann Kim
中科院分区:
生物学3区
文献类型:
--
作者:
KyungMann Kim

文献摘要

参考文献

被引文献

相似文献

本章将详细描述基于重复显着性检验方法的临床试验序贯设计。这里的术语序贯设计是指允许序贯分析并包括最大样本量估计的设计。这是基于随时间变化的有效分数测试统计数据来完成的,已知其联合分布具有独立的增量结构,从而可以轻松进行设计和分析所需的数值计算。我们首先回顾基于正态数据的通常固定设计,并基于有效分数测试推广到非正态数据。然后,我们回顾正常数据的经典组序贯设计和最大样本量的估计,并讨论基于信息的组序贯设计,包括基于膨胀因子的最大信息估计。在基于信息的组序贯设计之后,我们回顾了基于信息的组序贯分析和用于中期分析的 I 类错误支出函数方法。这种方法适应了中期分析的实际情况,包括连续分析之间统计信息的不等增量以及不可预测数量的重复显着测试。我们以 III 期非小细胞肺癌 (NSCLC) 的癌症和白血病 B 组试验 8433 为例,回顾了基于信息的分组序贯设计和随机对照试验分析的实施情况。
Sequential designs for clinical trials based on repeated significance testing approach will be described in detail in this chapter. The term sequential design here refers to a design that allows sequential analysis and includes estimation of the maximum sample size. This is done based on the efficient score test statistics over time whose joint distribution is known to have an independent increment structure, allowing for easy numerical computation necessary for design and analysis. We start by reviewing the usual fixed design based on normal data and generalizing to nonnormal data based on the efficient score test. We then review classical group sequential designs for normal data and the estimation of the maximum sample size and discuss information-based group sequential designs, including estimation of the maximum information based on an inflation factor. Following information-based group sequential designs, we review information-based group sequential analysis and the type I error spending function approach to interim analysis. This approach accommodates the practical aspects of interim analyses, including unequal increments in statistical information between successive analyses and an unpredictable number of repeated significant tests. We review the implementation of an information-based group sequential design and analysis of a randomized, controlled trial using the Cancer and Leukemia Group B trial 8433 in stage III non-small cell lung cancer (NSCLC) as an example.
DOI: 10.1002/sim.4780081003
发表时间: 1989-10-01
影响因子: 2
作者:
LAN, KKG;DEMETS, DL
通讯作者: DEMETS, DL
DOI: 10.1002/sim.4780111012
发表时间: 1992
影响因子: 2
作者:
Kim,K;Demets,DL
通讯作者: Demets,DL
DOI: 10.1056/nejm199010043231403
发表时间: 1990-10-04
影响因子: 158.5
作者:
DILLMAN, RO;SEAGREN, SL;GREEN, MR
通讯作者: GREEN, MR
组序贯临床试验的研究持续时间,并根据分层调整删失生存数据。
DOI: 10.1002/sim.4780111106
发表时间: 1992
影响因子: 2
作者:
Kim,K
通讯作者: Kim,K
SEQPWR 和 SEQOPR:用于设计基于组序贯对数秩检验的最大信息试验的计算机程序。
DOI: 10.1016/0169-2607(94)01615-m
发表时间: 1995
影响因子: 6.1
作者:
Kim,K
通讯作者: Kim,K