baskwrap

The baskwrap package supplies a unified wrapper to several basket trial packages (basksim and baskexact) using a unified syntax.

Installation

You can install the development version of baskwrap from GitHub with:

# install.packages("pak")
pak::pak("LukasDSauer/baskwrap")

Example

The baskwrap package provides a simple interface to switch between two methods for calculating basket trial characteristics, numerical integration (“exact”) and Monte Carlo simulation (“simulated”).

library(baskwrap)
# INPUT PARAMETERS
n <- 20
p1 <- c(0.2, 0.5, 0.5)
lambda <- 0.95
epsilon <- 2
tau <- 0.5
design <- setup_fujikawa_x(k = 3, p0 = 0.2, backend = "exact")
# DETAILS USING EXACT BACKEND
get_details(design = design, n = n, p1 = p1, lambda = lambda,
            epsilon = epsilon, tau = tau)
#> $Rejection_Probabilities
#> [1] 0.1656753 0.9623016 0.9623016
#> 
#> $FWER
#> [1] 0.1656753
#> 
#> $EWP
#> [1] 0.9983541
#> 
#> $Mean
#> [1] 0.2358052 0.4958199 0.4958199
#> 
#> $MSE
#> [1] 0.009524536 0.009835315 0.009835315
#> 
#> $Lower_CL
#> numeric(0)
#> 
#> $Upper_CL
#> numeric(0)
#> 
#> $ECD
#> [1] 2.758928
#> 
#> $p0
#> [1] 0.2
#> 
#> $p1
#> [1] 0.2 0.5 0.5
#> 
#> $backend
#> [1] "exact"
# DETAILS USING MC BACKEND
get_details(design = set_backend(design, "sim"),
            n = n, p1 = p1, lambda = lambda,
            epsilon = epsilon, tau = tau)
#> $Rejection_Probabilities
#> [1] 0.174 0.961 0.948
#> 
#> $FWER
#> [1] 0.174
#> 
#> $EWP
#> [1] 0.999
#> 
#> $Mean
#> [1] 0.2367023 0.4891821 0.4880540
#> 
#> $MSE
#> [1] 0.009438693 0.009745407 0.009909880
#> 
#> $Lower_CL
#> [1] 0.09531401 0.33277549 0.33140031
#> 
#> $Upper_CL
#> [1] 0.3917317 0.6456339 0.6448407
#> 
#> $ECD
#> [1] 2.735
#> 
#> $Rejection_Probabilities_SE
#> [1] 0.011988494 0.006122009 0.007021111
#> 
#> $FWER_SE
#> [1] 0.01198849
#> 
#> $EWP_SE
#> [1] 0.0009994999
#> 
#> $ECD_SE
#> [1] 0.01445628
#> 
#> $p0
#> [1] 0.2
#> 
#> $p1
#> [1] 0.2 0.5 0.5
#> 
#> $backend
#> [1] "sim"