TABLE 1.

Brief description and available input arguments to R functions subpop.sim and treatsel.sim

		Implementation
Argument	Description	Treatment selection (treatsel.sim)	Subgroup selection (subpop.sim)
n	Sample sizes for each treatment group at stage 1 (interim) and stage 2 (final) analyses	List of sample sizes; for example, 32 in each group in stage 1 and 64 in stage 2, list(stage1=32, stage2=64)	An additional list option can be used to increase (enrich) the sample size in stage 2; for example, enrich sample size to 128 in stage 2, list(stage1=32, enrich=128, stage2=64)
effect	Effect sizes for early and final outcomes	List of effect sizes for the control (first) and each treatment group; for example, for control (0) and two effects of size 0.1 and 0.2 for early and 0.2 and 0.3 for final outcome, list(early=c(0,0.1,0.2), final=c(0,0.2,0.3))	The first element of each vector is the effect size in the subgroup and the second is the effect size in the full population; for example, for an effect size of 0.4 in subgroup and 0.2 in the full population for both early and final outcomes list(early=c(0.4,0.2), final=c(0.4,0.2)). An optional argument can be included to set the effect size for the control group; for example, default is zero, control=list(early=0, final=NULL)
outcome	Outcome type for early and final outcomes	List of outcome types, options for normal (N), time‐to‐event (T), and binary (B) are currently available; for example, normal for early and final outcomes list(early="N", final="N").
nsim	Number of simulations	An integer $<1\times {10}^7$
sprev	Subgroup prevalence		The prevalence of the subgroup in the main population; $0<\mathtt{sprev}<1$. Subgroup prevalence can be either fixed or allowed to vary at each simulation; default is sprev.fixed=TRUE
corr	Correlation between outcomes	Correlation between early and final outcomes; $-1<\mathtt{corr}<1$
seed	Seed number	Seed number to ensure repeatability of simulations
select	Method for treatment selection	Seven available options (see Section 4.3.1), with default, select=0, to select all treatments. For select=4, the epsilon rule, can be used to allow more flexibility, with default epsilon=1 and for select=6, all treatments greater than a threshold can be selected, with default thresh=1	Two options are available. The default threshold selection rule (select="thresh"), for which limits must be set; for example, selim=c(-1,1). A futility selection rule (select="futility"; See Section 5.2) is also available, for which limits must be set; selim=c(0,0).
ptest	Treatments for counts of the number of rejections	A vector of valid treatment numbers for treatment specific counts of rejections; for example, for four treatments, ptest=c(1), ptest=c(1, 2) or ptest=c(1,2,3,4) are all valid options
method	Methodology used for simulations; and for subpop.sim either	Either method="invnorm" or "fisher" to select inverse normal or Fisher combination test	Select method="CT-SD", "CT-Simes", "CT-Bonferroni" or "CEF" (see Section 4.2.1)
fu	Follow‐up options	Subjects in the dropped treatment groups can followed‐up (fu=TRUE), with default fu=FALSE (see Section 4.3.1.)
weight	Stage 1 weight	Stage 1 weight be defined with this option (0 ⩽ weight ⩽1); default weight=NULL
level	Test level	Test level must be set ($0\le \mathtt{level}\le 1$); default level=0.025
file	File name for output	If unset will default to R console. For example, set file="output.txt", direct output to file in R working directory (getwd())