version 0.1.0

DESCRIPTION

@@ -0,0 +1,20 @@
+Package: frequentistSSDBinary
+Type: Package
+Title: Screened Selection Design with Binary Endpoints
+Version: 0.1.0
+Authors@R: c(person("Chia-Wei", "Hsu", role = c("aut", "cre"),
+                     email = "Chia-Wei.Hsu@stjude.org"),
+             person("Zongheng", "Cai", role = "aut"),
+             person("Haitao", "Pan", role = "aut"))
+Maintainer: Chia-Wei Hsu <Chia-Wei.Hsu@stjude.org>
+Description: A study based on the screened selection design (SSD) is an exploratory phase II randomized trial with two or more arms but without concurrent control. The primary aim of the SSD trial is to pick a desirable treatment arm (e.g., in terms of the response rate) to recommend to the subsequent randomized phase IIb (with the concurrent control) or phase III. The proposed designs can “partially” control or provide the empirical type I error/false positive rate by an optimal algorithm (implemented by the optimal_2arm_binary() or optimal_3arm_binary() function) for each arm. All the design needed components (sample size, operating characteristics) are supported.
+License: GPL-2
+Encoding: UTF-8
+Depends: mvtnorm, clinfun, ph2mult
+NeedsCompilation: no
+Packaged: 2024-06-26 03:27:31 UTC; chsu1
+Author: Chia-Wei Hsu [aut, cre],
+  Zongheng Cai [aut],
+  Haitao Pan [aut]
+Repository: CRAN
+Date/Publication: 2024-06-26 13:00:06 UTC

MD5

@@ -0,0 +1,17 @@
+c62f28da72e6ca111fff9c53216bc9eb *DESCRIPTION
+a18a470f0c0ec4fd42b9eefb5e4e2cca *NAMESPACE
+a5414add1ba51479192d4de122f57834 *R/SSD.2arms_notext.R
+1f4c3279393114f5f0f48c20c0ce918b *R/SSD.3arms_notext.R
+96228454fea3882ab153aa9995091f77 *R/get_oc_2arm_binary.R
+a4e694f635733a48a1c01387cbb049dd *R/get_oc_3arm_binary.R
+9a9d7a808d6554e18fc32d470bcaf1b5 *R/initial_sample.R
+804bd3eaf33ed182eff4b263e4c3c40f *R/optimal_2arm_binary.R
+8bb81ece38421ab84e9d813b1271f8a2 *R/optimal_3arm_binary.R
+f227a41a17f6c215266d123dfac9c629 *R/sample_size_2arm_binary.R
+57032dde461604fd063228fff758b9a8 *R/sample_size_3arm_binary.R
+79864831d321ae2896f42ae3e1c74bf5 *man/get_oc_2arm_binary.Rd
+6492f3caf82c33147640d97a9bb86e52 *man/get_oc_3arm_binary.Rd
+6a39b906b60a25bd7ef272fd3e1ec6a4 *man/optimal_2arm_binary.Rd
+d5057f51a5047944aff79bab3c0ffe84 *man/optimal_3arm_binary.Rd
+1899f915681d2ad5ab0766cc67c77497 *man/sample_size_2arm_binary.Rd
+2d136fff5c8143cd8f5b07cf601efcf4 *man/sample_size_3arm_binary.Rd

NAMESPACE

@@ -0,0 +1,10 @@
+importFrom("stats", "qnorm", "rbinom", "rmultinom", "runif")
+importFrom("mvtnorm", "qmvnorm", "pmvnorm")
+importFrom("clinfun", "ph2simon")
+importFrom("ph2mult", "binom.power")
+export(get_oc_2arm_binary)
+export(get_oc_3arm_binary)
+export(optimal_2arm_binary)
+export(optimal_3arm_binary)
+export(sample_size_2arm_binary)
+export(sample_size_3arm_binary)

R/SSD.2arms_notext.R

@@ -0,0 +1,161 @@
+SSD.2arms_notext <- function(r1, r, n1, n, p0, p1=NULL, p, nsim, diff=0.05, seed=0802) {
+
+  set.seed(seed)
+
+  if (n > 1000)
+    stop ("sample size, n cannot exceed 1000")
+
+  if (nsim <= 1) {
+    stop(" nsim less than 2! ")
+  }
+
+  if (r < r1  ) {
+    stop("r must be >= r1")
+  }
+
+  if (n <= n1) {
+    stop("condition for n > n1 must be satisfied for a 2 stage design")
+  }
+
+  n2 <- n- n1
+
+  p1<-ifelse(is.null(p1), p[1], p1)
+
+  outcome2<-No.success<-n.Subj<-matrix(999, ncol=2, nrow=nsim)
+
+  for(i in 1:nsim)	{
+
+    for (a in 1:2) {
+
+      Stage1<-rbinom(1, n1, p[a])
+
+      outcome1<-ifelse(Stage1>r1, rbinom(1, n2, p[a]), NA)
+
+      No.success[i,a]<-sum(outcome1, Stage1, na.rm=TRUE)
+
+      outcome2[i,a]<-ifelse(outcome1+Stage1>r, 1, 0)
+
+      n.Subj[i,a] <- ifelse(Stage1>r1, n, n1)
+
+    }
+
+  }
+
+
+  Outcome<-apply(outcome2, 1, sum, na.rm=T)
+
+
+  Outcome[is.na(Outcome)]<-0
+
+  Prob.neg    <-length(Outcome[Outcome==0])/nsim
+  Prob.pos    <-length(Outcome[Outcome==2])/nsim
+  Prob.negpos <-length(Outcome[Outcome==1])/nsim
+
+  Prob.ArmA<-sum(outcome2[,1], na.rm=TRUE)/nsim
+  Prob.ArmB<-sum(outcome2[,2], na.rm=TRUE)/nsim
+
+  mean.Subj<-apply(n.Subj,2,mean, na.rm=T)
+
+  ### After 1st segment
+  Prob.select.ArmA<-sum(outcome2[,1][outcome2[,2]==0 | is.na(outcome2[,2])],na.rm=TRUE)/nsim
+  Prob.select.ArmB<-sum(outcome2[,2][outcome2[,1]==0 | is.na(outcome2[,1])],na.rm=TRUE)/nsim
+  Prob.NoArm<-Prob.neg
+
+  ### Selecting an Arm when both arms are positive  (2nd Segment)
+
+  No.success.BothArms.select<-No.success[Outcome==2,]
+
+  SSD.SelectArm<-function(x, diff)
+  {
+    NoArm<-ArmA<-ArmB<-NA
+    if(x[2]/n==1&x[1]/n!=1){
+      ArmB=1
+      ArmA=0
+    }
+    if(x[2]/n!=1&x[1]/n==1){
+      ArmB=0
+      ArmA=1
+    }
+    if(x[2]/n==1&x[1]/n==1){
+      if(diff==0){
+        ArmA<-ifelse(runif(1,0,1)<0.5,0,1)
+        ArmB<-1-ArmA
+      }
+      if(diff!=0){
+        NoArm=1
+      }
+    }
+    if(x[2]/n!=1&x[1]/n!=1){
+      test_a=sqrt(n)*(x[1]/n-p0)/sqrt((x[1]/n)*(1-x[1]/n))
+      test_b=sqrt(n)*(x[2]/n-p0)/sqrt((x[2]/n)*(1-x[2]/n))
+      ArmB<-ifelse(test_b-test_a > diff,1,0)
+      ArmA<-ifelse(test_b-test_a < -diff,1,0)
+      ## for ties
+      if(diff==0 & test_b-test_a  == 0) {
+        ArmA<-ifelse(runif(1,0,1)<0.5,0,1)
+        ArmB<-1-ArmA
+      }
+
+      ## for SSD_mod with no selection for ties or if diff <= e.g.0.05
+      if(diff!=0) {
+        NoArm<-ifelse(test_b-test_a <= diff & test_a-test_b <= diff,1,0)
+      }
+    }
+    return(list(ArmA, ArmB, NoArm))
+  }
+
+
+  ### Original SSD
+
+  if( length(No.success[Outcome==2,]) > 1 ) {
+    SSD.SelectArm.2ndSeg<-matrix(unlist(
+      apply(No.success[Outcome==2,],1,SSD.SelectArm, diff=0)   ###for each Two-stage successfull row, apply the selection criteria
+    ),ncol=3,byrow=T)}   else {SSD.SelectArm.2ndSeg <- matrix(NA,ncol=3)  }
+
+  ProbArmA.2ndSeg<-sum(SSD.SelectArm.2ndSeg[,1],na.rm=TRUE)/nsim
+  ProbArmB.2ndSeg<-sum(SSD.SelectArm.2ndSeg[,2],na.rm=TRUE)/nsim
+  ProbNoArm.2ndSeg<-sum(SSD.SelectArm.2ndSeg[,3],na.rm=TRUE)/nsim
+
+  Overall.ArmA<-Prob.select.ArmA + ProbArmA.2ndSeg
+  Overall.ArmB<-Prob.select.ArmB + ProbArmB.2ndSeg
+  Overall.NoArm<-Prob.NoArm + ProbNoArm.2ndSeg
+
+
+  ### Modified SSD ####
+
+  if( length(No.success[Outcome==2,]) > 1 ) {
+    SSD.SelectArm.2ndSeg<-matrix(unlist(
+      apply(No.success[Outcome==2,],1,SSD.SelectArm, diff)
+    ),ncol=3,byrow=T)}   else {SSD.SelectArm.2ndSeg <- matrix(NA,ncol=3)  }
+
+  ProbArmA.2ndSeg.MOD<-sum(SSD.SelectArm.2ndSeg[,1],na.rm=TRUE)/nsim
+  ProbArmB.2ndSeg.MOD<-sum(SSD.SelectArm.2ndSeg[,2],na.rm=TRUE)/nsim
+  ProbNoArm.2ndSeg.MOD<-sum(SSD.SelectArm.2ndSeg[,3],na.rm=TRUE)/nsim
+
+  Overall.ArmA.MOD<-Prob.select.ArmA + ProbArmA.2ndSeg.MOD
+  Overall.ArmB.MOD<-Prob.select.ArmB + ProbArmB.2ndSeg.MOD
+  Overall.NoArm.MOD<-Prob.NoArm + ProbNoArm.2ndSeg.MOD
+
+  if(diff==0){
+    soln<-data.frame("n"=n,
+                     "SSD Arm A"=Overall.ArmA, "SSD Arm B"= Overall.ArmB, "SSD No Arm"=Overall.NoArm,
+                     "diff"=diff,
+                     "Mean N Arm A"=mean.Subj[1],"Mean N Arm B"=mean.Subj[2])
+
+
+  }
+  if(diff!=0){
+    soln<-data.frame("n"=n,
+                     "Modified SSD Arm A"=Overall.ArmA.MOD, "Modified SSD Arm B"=Overall.ArmB.MOD,
+                     "Modified SSD No Arm"=Overall.NoArm.MOD, "diff"=diff,
+                     "Mean N Arm A"=mean.Subj[1],"Mean N Arm B"=mean.Subj[2])
+
+
+  }
+
+
+
+  soln
+
+}
+