simulate-Design-method: Simulate outcomes from a CRM design

In crmPack: Object-Oriented Implementation of CRM Designs

Simulate outcomes from a CRM design

Description

Simulate outcomes from a CRM design

Usage

## S4 method for signature 'Design'
simulate(
  object,
  nsim = 1L,
  seed = NULL,
  truth,
  args = NULL,
  firstSeparate = FALSE,
  mcmcOptions = McmcOptions(),
  parallel = FALSE,
  nCores = min(parallel::detectCores(), 5),
  ...
)

Arguments

object	the Design object we want to simulate data from
nsim	the number of simulations (default: 1)
seed	see setSeed
truth	a function which takes as input a dose (vector) and returns the true probability (vector) for toxicity. Additional arguments can be supplied in args.
args	data frame with arguments for the truth function. The column names correspond to the argument names, the rows to the values of the arguments. The rows are appropriately recycled in the nsim simulations. In order to produce outcomes from the posterior predictive distribution, e.g, pass an object that contains the data observed so far, truth contains the prob function from the model in object, and args contains posterior samples from the model.
firstSeparate	enroll the first patient separately from the rest of the cohort? (not default) If yes, the cohort will be closed if a DLT occurs in this patient.
mcmcOptions	object of class McmcOptions, giving the MCMC options for each evaluation in the trial. By default, the standard options are used
parallel	should the simulation runs be parallelized across the clusters of the computer? (not default)
nCores	how many cores should be used for parallel computing? Defaults to the number of cores on the machine, maximum 5.
...	not used

Value

an object of class Simulations

Examples

# Define the dose-grid
emptydata <- Data(doseGrid = c(1, 3, 5, 10, 15, 20, 25, 40, 50, 80, 100))

# Initialize the CRM model 
model <- LogisticLogNormal(mean=c(-0.85, 1),
                           cov=
                             matrix(c(1, -0.5, -0.5, 1),
                                    nrow=2),
                           refDose=56)

# Choose the rule for selecting the next dose 
myNextBest <- NextBestNCRM(target=c(0.2, 0.35),
                           overdose=c(0.35, 1),
                           maxOverdoseProb=0.25)

# Choose the rule for the cohort-size 
mySize1 <- CohortSizeRange(intervals=c(0, 30),
                           cohortSize=c(1, 3))
mySize2 <- CohortSizeDLT(DLTintervals=c(0, 1),
                         cohortSize=c(1, 3))
mySize <- maxSize(mySize1, mySize2)

# Choose the rule for stopping
myStopping1 <- StoppingMinCohorts(nCohorts=3)
myStopping2 <- StoppingTargetProb(target=c(0.2, 0.35),
                                  prob=0.5)
myStopping3 <- StoppingMinPatients(nPatients=20)
myStopping <- (myStopping1 & myStopping2) | myStopping3

# Choose the rule for dose increments
myIncrements <- IncrementsRelative(intervals=c(0, 20),
                                   increments=c(1, 0.33))

# Initialize the design
design <- Design(model=model,
                 nextBest=myNextBest,
                 stopping=myStopping,
                 increments=myIncrements,
                 cohortSize=mySize,
                 data=emptydata,
                 startingDose=3)

## define the true function
myTruth <- function(dose)
{
  model@prob(dose, alpha0=7, alpha1=8)
}

# Run the simulation on the desired design
# We only generate 1 trial outcomes here for illustration, for the actual study 
# this should be increased of course
options <- McmcOptions(burnin=100,
                       step=1,
                       samples=2000)
time <- system.time(mySims <- simulate(design,
                                       args=NULL,
                                       truth=myTruth,
                                       nsim=1,
                                       seed=819,
                                       mcmcOptions=options,
                                       parallel=FALSE))[3]
  
  

crmPack documentation built on June 26, 2024, 5:07 p.m.