GROUP SEQUENTIAL DESIGNS FOR ONE-SIDED AND 2-SIDED HYPOTHESIS-TESTING WITH PROVISION FOR EARLY STOPPING IN FAVOR OF THE NULL HYPOTHESIS
GROUP SEQUENTIAL DESIGNS FOR ONE-SIDED AND 2-SIDED HYPOTHESIS-TESTING WITH PROVISION FOR EARLY STOPPING IN FAVOR OF THE NULL HYPOTHESIS
复制标题
DOI：
10.1016/0378-3758(94)90187-2
复制
发表时间：
1994-11-01
影响因子：
0.9
通讯作者：
TSIATIS, AA
中科院分区：
文献类型：
--
作者：
PAMPALLONA, S;TSIATIS, AA
在这篇文章中，我们考虑了一类分组序贯临床试验的边界，当观察到微小的治疗差异时，这些试验也允许提前停止。它们基于Wang和Tsiatis(Biometrics 43(1987)193-199)提出的单参数族边界，可以精确地推导出任何类型I和类型II错误概率的选择，并且可以很容易地应用于单边和双边假设检验。根据平均样本数对不同设计参数选择对应的方案进行比较。提供了相关参数的表格，以帮助设计成组序贯试验。
In this paper we consider a class of boundaries for group sequential clinical trials that allow for early stopping also when small treatment differences are observed. They are based on the one-parameter family of boundaries introduced by Wang and Tsiatis (Biometrics 43 (1987) 193-199), can be derived exactly for any choice of type I and type II error probabilities, and can be easily applied to both one- and two-sided hypothesis testing. Plans corresponding to different choices of the design parameters are compared in terms of average sample number. Tables of relevant parameters are provided to help in the design of group sequential tests.