crm: Executing the CRM

In dfcrm: Dose-Finding by the Continual Reassessment Method

Description

crm is used to compute a dose for the next patient in a phase I trial according to the CRM.

Usage

crm(prior, target, tox, level, n = length(level), dosename = NULL, 
    include = 1:n, pid = 1:n, conf.level = 0.9, method = "bayes", 
    model = "empiric", intcpt = 3, scale = sqrt(1.34), model.detail = TRUE, 
    patient.detail = TRUE, var.est = TRUE) 

Arguments

prior	A vector of initial guesses of toxicity probabilities associated the doses.
target	The target DLT rate.
tox	A vector of patient outcomes; 1 indicates a toxicity, 0 otherwise.
level	A vector of dose levels assigned to patients. The length of level must be equal to that of tox.
n	The number of patients enrolled.
dosename	A vector containing the names of the regimens/doses used. The length of dosename must be equal to that of prior.
include	A subset of patients included in the dose calculation.
pid	Patient ID provided in the study. Its length must be equal to that of level.
conf.level	Confidence level for the probability/confidence interval of the returned dose-toxicity curve.
method	A character string to specify the method for parameter estimation. The default method "bayes" estimates the model parameter by the posterior mean. Maximum likelihood estimation is specified by "mle".
model	A character string to specify the working model used in the method. The default model is "empiric". A one-parameter logistic model is specified by "logistic".
intcpt	The intercept of the working logistic model. The default is 3. If model="empiric", this argument will be ignored.
scale	Standard deviation of the normal prior of the model parameter. Default is sqrt(1.34).
model.detail	If FALSE, the model content of an "mtd" object will not be displayed. Default is TRUE.
patient.detail	If FALSE, patient summary of an "mtd" object will not be displayed. Default is TRUE.
var.est	If TRUE, variance of the estimate of the model parameter and probability/confidence interval for the dose-toxicity curve will be computed

Details

For maximum likelihood estimation, the variance of the estimate of β (post.var) is approximated by the posterior variance of β with a dispersed normal prior.

The empiric model is specified as F(d, β) = d^{\exp(β)}. The logistic model is specified as logit (F(d,β)) = intcpt + \exp(β) \times d. For method="bayes", the prior on β is normal with mean 0. Exponentiation of β ensures an increasing dose-toxicity function.

Value

An object of class "mtd" is returned, consisting of the summary of dose assignments thus far and the recommendation of dose for the next patient.

prior	Initial guesses of toxicity rates.
target	The target probability of toxicity at the MTD.
ptox	Updated estimates of toxicity rates.
ptoxL	Lower confidence/probability limits of toxicity rates.
ptoxU	Upper confidence/probability limits of toxicity rates.
mtd	The updated estimate of the MTD.
prior.var	The variance of the normal prior.
post.var	The posterior variance of the model parameter.
estimate	Estimate of the model parameter.
method	The method of estimation.
model	The working model.
dosescaled	The scaled doses obtained via backward substitution.
tox	Patients' toxicity indications.
level	Dose levels assigned to patients.

References

O'Quigley, J. O., Pepe, M., and Fisher, L. (1990). Continual reassessment method: A practical design for phase I clinical trials in cancer. Biometrics 46:33-48.

Cheung, Y. K. (2011). Dose Finding by the Continual Reassessment Method. New York: Chapman & Hall/CRC Press.

Examples

# Create a simple data set
prior <- c(0.05, 0.10, 0.20, 0.35, 0.50, 0.70)
target <- 0.2
level <- c(3, 4, 4, 3, 3, 4, 3, 2, 2, 2)
y <- c(0, 0, 1, 0, 0, 1, 1, 0, 0, 0)
foo <- crm(prior, target, y, level)
ptox <- foo$ptox  # updated estimates of toxicity rates

dfcrm documentation built on May 1, 2019, 10:18 p.m.