Tabulate rank-based desirability scores for a BOIN12 trial
Usage
boin12_rds(
sample_sizes,
phi_t,
phi_e,
u1 = 100,
u2,
u3,
u4 = 0,
c_t = 0.95,
c_e = 0.9,
prior_alpha = 1,
prior_beta = 1
)Arguments
- sample_sizes
integer vector, cohort sample sizes to consider
- phi_t
Probability of toxicity threshold
- phi_e
Probability of efficacy threshold
- u1
utility of efficacy without toxicity, 100 by default
- u2
utility of no efficacy and no toxicity, between u1 and u4
- u3
utility of efficacy and toxicity, between u1 and u4
- u4
utility of toxicity without efficacy , 0 by default
- c_t
certainty required to flag excess toxicity, 0.95 by default
- c_e
certainty required to flag deficient efficacy, 0.9 by default
- prior_alpha
first shape param for prior on beta prior, 1 by default
- prior_beta
second shape param for prior on beta prior, 1 by default
Value
data.frame with columns Patients, Toxicity, Efficacy containing the numbers of patients, patients with toxicitiy, and patients with efficacy; Admissble, containing the character labels Admissble or Not Admissible; RDS, containing a character label of the numerical desirability score or the character string "E", where a combination is eliminated; and RDS_x, containing the desirability scores as numbers, with NA where a combination should be eliminated.
References
Lin, R., Zhou, Y., Yan, F., Li, D., & Yuan, Y. (2020). BOIN12: Bayesian optimal interval phase I/II trial design for utility-based dose finding in immunotherapy and targeted therapies. JCO Precision Oncology, 4, 1393-1402.
Examples
# Table 3 in Lin et al.
x <- boin12_rds(
sample_sizes = c(0, 3, 6, 9),
phi_t = 0.35,
phi_e = 0.25,
u1 = 100,
u2 = 40,
u3 = 60,
u4 = 0,
c_t = 0.95,
c_e = 0.9,
prior_alpha = 1,
prior_beta = 1
)