An overview of sequential methods and their application in clinical trials

Issues in Planning Interim Analyses

“…Formal group se quential plans may be an efficient way to gain early answers about drug efficacy in certain disease settings. Methods appro priate to a variety of study settings, as re viewed here and elsewhere (18,19), need to be carefully considered and prespecified in the study plan. Likewise, in nonsequential trials, the need to specify plans for datamonitoring and interim tests is equally im portant to maintain the integrity of the completed study.…”

Flexibly Monitoring Group Sequential Survival Trials When Testing is Based Upon a Weighted Log-Rank Statistic

“…One approach to account for deviations from a pre-planned schedule of analyses is the constrained boundaries algorithm (Burington and Emerson (2003)). The constrained boundary algorithm is a generalization of the error spending approach of Demets and Lan (1984). Specifically, the use of error spending functions constrained the proportion of type I error spent at previous analyses using a pre-defined function of the proportion of maximal information accrued at those preceding analyses.…”

“…Interim testing can be formalized using a group sequential framework to attain desired frequentist operating characteristics (Emerson et al (2007)). To control the type I error rate under repeated tests of significance multiple authors have proposed discrete sequential stopping rules (Armitage et al (1969); Pocock (1977); O’Brien and Fleming (1979)) and error spending approaches (Demets and Lan (1984); Pampallona (1995)). Most commonly used group sequential stopping rules consider continuation sets of the form C j = ( a j , b j ] ⋃ [ c j , d j ) such that −∞ ≤ a j ≤ b j ≤ c j ≤ d j ≤ ∞for j = 1, … , J analyses.…”

On the Use of Local Assessments for Monitoring Centrally Reviewed Endpoints with Missing Data in Clinical Trials

“…The timing of sequential analyses is measured by the proportion of statistical information obtained at an interim analysis relative to the maximal information that is anticipated at the final analysis of the trial [ 3 ]. Thus it is important to reliably estimate statistical information at each interim analysis in order to properly implement and potentially re-power a chosen group sequential design [ 4 , 5 ]. However, when an IRC is used to adjudicate a trial endpoint there may be a subset of individuals who do not have verified IRC measurements at the time of an interim analysis because the final assessment of their outcome has yet to be returned by the IRC.…”