II Oncology Trials

titesim_ss

R Documentation

Expanded version of the dfcrm::titesim function to incorporate design elements into the time-to-event continual reassessment method that are especially useful for seamless simulations.

Description

This is the simulator function for a TITE-CRM trial. It is meant to be called in the context of twostage_simulator() rather than by the user directly. Required input includes both the design elements as well as the true, dose-toxicity curve that is generally unknown in the real world. The simulator runs a certain number of simulated trials, creating data according to the dose-toxicity curve and making assignments according to the titecrm model. Various operating characteristics are reported.

Usage

titesim_ss(
  PI,
  prior,
  target,
  n,
  x0,
  nsim = 1,
  restrict = TRUE,
  obswin = 1,
  tgrp = obswin,
  rate = 1,
  accrual = "fixed",
  surv = "uniform",
  surv_rate = obswin/10,
  scheme = "polynomial",
  scheme_args = list(scheme_power = 1),
  count = TRUE,
  method = "bayes",
  model = "empiric",
  intcpt = 3,
  scale = sqrt(1.34),
  seed = 1009,
  conf.level = 0.5,
  no.exceed = Inf,
  cohort.size = 1,
  first.cohort.only = T,
  n.at.MTD = Inf,
  followup_b4_esc = obswin,
  earliest_stop = 6
)

Arguments

`PI`	Numeric vector with entries between 0 and 1 of true toxicity probabilities; assumed to match the order of the dose labels.
`prior`	Numeric vector with entries between 0 and 1 of anticipated toxicity probabilities, assumed to match the order of the dose labels. More commonly and accurately called the skeleton.
`target`	Scalar value between 0 and 1 giving the targeted rate of DLT.
`n`	A positive integer indicating the maximum number of patients to enroll.
`x0`	A positive integer indicating the starting dose level.
`nsim`	A positive integer indicating the number of simulated trials to conduct.
`restrict`	A logical value. If TRUE, indicates that safety constraints should be enacted. There are four: (1) No skipping doses in escalation; (2) No escalation before followup of followup_b4_esc on at least one patient at current or larger dose; (3) No assignment to dose with estimated DLT rate beyond no.exceed + target; (4) Stopping trial altogether if at any point after patient earliest_stop the estimated DLT rate at all dose levels exceeds no.exceed + target.
`obswin`	A positive number indicating the number of units of time over which DLTs are defined.
`tgrp`	This argument is an artefact of the dfcrm package and should be left blank.
`rate`	A positive number indicating the number of patients expected per unit obswin.
`accrual`	A string, either "fixed" or "poisson", denoting the patient accrual process.
`surv`	A string, either "uniform" or "exponential", denoting the time-to-DLT distribution.
`surv_rate`	A positive number. If surv is "exponential", this is the rate for the time-to-DLT distribution.
`scheme`	A string indicating the weighting scheme for patients who are free of DLT but have not completed followup. Can be either "polynomial", "logistic", or "adaptive."
`scheme_args`	A named list with elements "scheme power" (if scheme = "polynomial"), "scheme_int" and "scheme_slope" (if scheme = "logistic"), or no elements (if scheme = "adaptive").
`count`	A logical value; if TRUE, the progress of the simulations will be plotted.
`method`	A string indicating the method for fitting model. The original titesim function allows "mle" or "bayes"; here, only "bayes" is allowed.
`model`	A string indicating the type of model. The original titesim function allows "empiric" (sometimes known as the power model) or "logistic".
`intcpt`	A numeric value giving the intercept parameter when using the "logistic" model.
`scale`	A numeric value giving the prior standard deviation on the parameter beta.
`seed`	A positive integer random seed.
`conf.level`	A number between 0 and 1 indicating the confidence limits to report.
`no.exceed`	A positive number indicating by how much the toxicity rate can exceed the DLT threshold before the dose is unacceptable. To be more clear, no dose will be assigned with estimated pr(DLT) > target + no.exceed, and the trial will stop altogether if this holds for dose level 1.
`cohort.size`	A positive integer indicating the size of cohorts between successive queries of the CRM model.
`first.cohort.only`	A logical value indicating if cohort.size applies only to the first cohort.
`n.at.MTD`	A positive integer which is the number of patients that should be at the estimated MTD before the trial stops. If n.at.MTD patients have been assigned to a single dose, and this dose is the current MTD that would be assigned to the next patient, then the trial stops.
`followup_b4_esc`	A positive number indicating how much follow-up time at least one patient has to have on a given dose before the model escalates to the next dose.
`earliest_stop`	A positive integer indicating how many patients must be observed before the trial can stop.

Value

The function returns a list, with named components last_sim and all_sim. Access the last_sim to see the individual details of the single last simulation, or access all_sim to see summaries across all simulations performed.