A menu-driven facility for sample-size calculation in multiarm, multistage randomized controlled trials with time-to-event outcomes: Update

Barthel, Royston, and Parmar (2009, Stata Journal 9: 505-523) presented a menu-driven Stata program for calculating sample size and other design characteristics for a class of multiarm, multistage trials with time-to-event outcomes. In this article, we present several new features for the package. First, we allow hazard ratios greater than 1 to be targeted under the alternative hypothesis to make the design more widely applicable to other outcomes and disease areas. Second, we introduce new subroutines that use simulation to more accurately estimate the correlation between treatment effects at different stages and that calculate the familywise error rate, and we apply them to an example using the original design of the multiarm, multistage trial in prostate cancer. Finally,. we present a new design of the dialog menu for nstage that improves its usability and incorporates options for calling the new subroutines.