summary-PseudoSimulations-method: Summarize the simulations, relative to a given truth

In Roche/crmPack: Object-Oriented Implementation of CRM Designs

Summarize the simulations, relative to a given truth

Description

Summarize the simulations, relative to a given truth

Usage

## S4 method for signature 'PseudoSimulations'
summary(object, truth, targetEndOfTrial = 0.3, targetDuringTrial = 0.35, ...)

Arguments

object	the PseudoSimulations object we want to summarize
truth	a function which takes as input a dose (vector) and returns the true probability (vector) for toxicity
targetEndOfTrial	the target probability of DLE wanted to achieve at the end of a trial
targetDuringTrial	the target probability of DLE wanted to achieve during a trial
...	Additional arguments can be supplied here for truth

Value

an object of class PseudoSimulationsSummary

Examples

emptydata <- Data(doseGrid = seq(25, 300, 25))

# The design incorporate DLE responses and DLE samples.
# Specify the model of 'ModelTox' class eg 'LogisticIndepBeta' class model.
my_model <- LogisticIndepBeta(
  binDLE = c(1.05, 1.8),
  DLEweights = c(3, 3),
  DLEdose = c(25, 300),
  data = emptydata
)

# The escalation rule.
td_next_best <- NextBestTD(
  prob_target_drt = 0.35,
  prob_target_eot = 0.3
)

# Cohort size is 3 subjects.
my_size <- CohortSizeConst(size = 3)

# Allow increase of 200%.
my_increments <- IncrementsRelative(intervals = 0, increments = 2)

# Stopp when the maximum sample size of 36 patients has been reached or the next
# dose is NA.
my_stopping <- StoppingMinPatients(nPatients = 36) | StoppingMissingDose()

# Specify the design. (For details please refer to the 'TDDesign' example.)
my_design <- TDDesign(
  model = my_model,
  nextBest = td_next_best,
  stopping = my_stopping,
  increments = my_increments,
  cohort_size = my_size,
  data = emptydata,
  startingDose = 25
)

# Specify the truth of the DLE responses.
my_truth <- probFunction(my_model, phi1 = -53.66584, phi2 = 10.50499)

# For illustration purpose 50 burn-ins to generate 200 samples are used.
my_options <- McmcOptions(burnin = 50, step = 2, samples = 200)

# Refer to design-method 'simulate TDDesign' examples for details.
# For illustration purpose only 1 simulation is produced.
my_sim <- simulate(
  object = my_design,
  args = NULL,
  truth = my_truth,
  nsim = 1,
  seed = 819,
  parallel = FALSE,
  mcmcOptions = my_options
)

# Produce a summary of the simulations.
summary(
  my_sim,
  truth = my_truth
)

# Example where DLE samples are involved.

# Specify the next best rule.
td_next_best <- NextBestTDsamples(
  prob_target_drt = 0.35,
  prob_target_eot = 0.3,
  derive = function(samples) {
    as.numeric(quantile(samples, probs = 0.3))
  }
)

# The design.
my_design <- TDsamplesDesign(
  model = my_model,
  nextBest = td_next_best,
  stopping = my_stopping,
  increments = my_increments,
  cohort_size = my_size,
  data = emptydata,
  startingDose = 25
)

# For illustration purpose 50 burn-ins to generate 200 samples are used.
my_options <- McmcOptions(burnin = 50, step = 2, samples = 200)

# For illustration purpose 2 trials are simulated.
my_sim <- simulate(
  object = my_design,
  args = NULL,
  truth = my_truth,
  nsim = 2,
  seed = 819,
  mcmcOptions = my_options,
  parallel = FALSE
)

# Produce a summary of the simulations.
summary(
  my_sim,
  truth = my_truth
)

Roche/crmPack documentation built on April 30, 2024, 3:19 p.m.