comparison_sim: Simulate model-based and rule-based trials, and compare...
In dominicmagirr/dosecombo: Combination Dose Escalation

Simulate model-based and rule-based trials, and compare results.

comparison_sim(truth, doses, starting_dose, route_through_doses, max_n_cohort,
  cohort_size, n_sims, seed, max_cohort_size, max_n, max_p_target,
  add_random = FALSE, frac_target = 0.75, over_limit_end = NA,
  n_iter = 2500)

`truth`	A vector of 'true' p(DLT)'s, corresponding to the rows of 'doses'.
`doses`	A list containing doses on the original and log scale.
`starting_dose`	The row number of 'doses' corresponding to the starting dose.
`route_through_doses`	The row numbers of 'doses' corresponding to the doses used in the '3+3' design.
`max_n_cohort`	The maximum number of cohorts.
`cohort_size`	The number of patients per cohort.
`n_sims`	The number of simulations.
`seed`	The random number seed.
`max_cohort_size`	Maximum size of cohort. If this is reached, and the max_n or max_p_target is reached then the trial can stop.
`max_n`	Maximum total sample size. If this is reached and the max_cohort_size is reached for recommended dose then the trial can stop.
`max_p_target`	Maximum probability of targeted toxicity. If this is reached and the max_cohort_size is reached for recommended dose then the trial can stop.
`add_random`	Should some randomness be added to choice of next dose? Default is FALSE.
`frac_target`	If randomness is added, how close to the 'best' dose is acceptable. Default is 75 percent of P(target toxicity).
`over_limit_end`	The safety limit to be applied at the end of the trial
`n_iter`	Number of iterations in MCMC.

A list of 4 items: a plot of the true p(DLTs), a graphical comparison of the 2 methods, a summary of model-based simulations, and a summary of rule-based simulations.

dominicmagirr/dosecombo documentation built on May 6, 2019, 4:32 p.m.