Processed results of an EffTox dose-update analysis, yielded by
efftox_process.
The model is invoked on data and RStan performs sampling from the posterior
distribution.
efftox_process then conducts the required processing to ultimately
arrive at a dose recommendation.
A list with elements identified above.
The object is essentially a list with slots:
dose_indices, a vector of integers representing the dose-levels
under consideration.
recommended_dose, an integer representing the dose recommended
for the next patient or cohort. TODO: what happens when stop recommended?
prob_eff, the posterior mean probabilities of efficacy at doses
1:n; a vector of numbers between 0 and 1.
prob_eff, the posterior mean probabilities of toxicity at doses
1:n; a vector of numbers between 0 and 1.
prob_acc_eff, the posterior mean probabilities that efficacy at
the doses is acceptable, i.e. that it exceeds the minimum acceptable efficacy
threshold; a vector of numbers between 0 and 1.
prob_acc_eff, the posterior mean probabilities that toxicity at
the doses is acceptable, i.e. that it is less than the maximum toxicity
threshold; a vector of numbers between 0 and 1.
utility, the utilities of doses 1:n, calculated by plugging the
posterior mean probabilities of efficacy and toxicity into the utility
formula, as advocated by Thall & Cook. Contrast to post_utility;
a vector of numbers.
post_utility, the posterior mean utilities of doses 1:n,
calculated from the posterior distributions of the utilities. This is in
contrast to utility, which uses plug-in posterior means of efficacy
and toxicity, as advocated by Thall & Cook; a vector of numbers.
acceptable, a vector of logical values to indicate whether
doses 1:n are acceptable, according to the rules for acceptable efficacy &
toxicity, and rules on not skipping untested doses.
fit, An object of class stanfit, containing the posterior samples.