AGSDest
(Package: AGSDest) :
Estimation in adaptive group sequential trials
The package allows to compute repeated confidence intervals as well as confidence intervals based on the stage-wise ordering in groug sequential designs (GSD; see Jennison and Turnbull, 1989; Tsiatis, Rosner, Mehta, 1984) and adaptive groug sequential designs (Mehta, Bauer, Posch, Brannath, 2007; Brannath, Mehta, Posch, 2008). For adaptive group sequential designs the confidence intervals are based on the conditional rejection probability principle of Mueller and Schaefer (2001). This principle allows us to perform data dependent changes to the sample size, the spending function, and the number and spacing of interim looks while preserving the overall type I error rate. Currently the procedures do not support the use of futility boundaries as well as more than one adaptive interim analysis. Furthermore, the package is currently restricted to the computation of lower one-sided confidence intervals.