ewoc (version 0.3.0)

ewoc_d1ph: Escalation With Overdose Control

Description

Finding the next dose for a phase I clinical trial based on Escalation with Overdose Control (EWOC) design considering parametrization for time to event response and single agent.

Usage

ewoc_d1ph(
  formula,
  theta,
  alpha,
  tau,
  type = c("continuous", "discrete"),
  rho_prior,
  mtd_prior,
  shape_prior = NULL,
  min_dose,
  max_dose,
  first_dose = NULL,
  last_dose = NULL,
  dose_set = NULL,
  max_increment = NULL,
  no_skip_dose = TRUE,
  distribution = c("exponential", "weibull"),
  rounding = c("down", "nearest"),
  n_adapt = 5000,
  burn_in = 1000,
  n_mcmc = 1000,
  n_thin = 1,
  n_chains = 1
)

Arguments

formula

an object of class Formula: a symbolic description of the model to be fitted with only one regressor term corresponding to the dose for the right side and a matrix as a response containing time and status for the left side.

theta

a numerical value defining the proportion of expected patients to experience a medically unacceptable, dose-limiting toxicity (DLT) if administered the MTD.

alpha

a numerical value defining the probability that the dose selected by EWOC is higher than the MTD.

tau

a numerical value defining the period of time for a possible toxicity be observed.

type

a character describing the type of the Maximum Tolerable Dose (MTD) variable. It can be 'discrete' or 'continuous'.

rho_prior

a matrix 1x2 of hyperparameters for the Beta prior distribution associated with the parameter rho.

mtd_prior

a matrix 1x2 of hyperparameters for the Beta prior distribution associated with the parameter MTD.

shape_prior

a matrix 1x2 of hyperparameters for the Gamma prior distribution associated with the shape parameter r for the Weibull distribution. It is only necessary if distribution = 'weibull'.

min_dose

a numerical value defining the lower bound of the support of the MTD.

max_dose

a numerical value defining the upper bound of the support of the MTD.

first_dose

a numerical value for the first allowable dose in the trial. It is only necessary if type = 'continuous'.

last_dose

a numerical value for the last allowable dose in the trial. It is only necessary if type = 'continuous'.

dose_set

a numerical vector of allowable doses in the trial. It is only necessary if type = 'discrete'.

max_increment

a numerical value indicating the maximum increment from the current dose to the next dose. It is only applied if type = 'continuous'.

no_skip_dose

a logical value indicating if it is allowed to skip doses. It is only necessary if type = 'discrete'. The default is TRUE.

distribution

a character establishing the distribution for the time of events. It can be defined as 'exponential' or 'weibull'.

rounding

a character indicating how to round a continuous dose to the one of elements of the dose set. It can be 'nearest' or 'down'. It is only necessary if type = 'discrete'.

n_adapt

the number of iterations for adaptation. See adapt for details.

burn_in

the number of iterations before to start monitoring.

n_mcmc

the number of iterations to monitor.

n_thin

thinning interval for monitors.

n_chains

the number of parallel chains for the model.

Value

next_dose the next recommend dose.

mtd the posterior MTD distribution.

rho the posterior rho_0 distribution.

sample a list of the MCMC chains distribution.

trial a list of the trial conditions.

References

Tighiouart M, Liu Y, Rogatko A. Escalation with overdose control using time to toxicity for cancer phase I clinical trials. PloS one. 2014 Mar 24;9(3):e93070.

Examples

# NOT RUN {
time <- 9
status <- 0
dose <- 20

test <- ewoc_d1ph(cbind(time, status) ~ dose, type = 'discrete',
                 theta = 0.33, alpha = 0.25, tau = 10,
                 min_dose = 20, max_dose = 100,
                 dose_set = seq(20, 100, 20),
                 rho_prior = matrix(1, ncol = 2, nrow = 1),
                 mtd_prior = matrix(1, ncol = 2, nrow = 1),
                 distribution = 'exponential',
                 rounding = 'nearest')
summary(test)
plot(test)

# }