Get an object to fit the BOIN COMB model using the BOIN package.
Source:R/boin_comb_selector.R
get_boin_comb.RdGet an object to fit the BOIN COMB model using the BOIN package.
Usage
get_boin_comb(
parent_selector_factory = NULL,
num_doses,
target,
use_stopping_rule = TRUE,
...
)Arguments
- parent_selector_factory
optional object of type
selector_factorythat is in charge of dose selection before this class gets involved. Leave as NULL to just use CRM from the start.- num_doses
integer vector of the number of doses of treatment 1, 2
- target
We seek a dose with this probability of toxicity.
- use_stopping_rule
TODO
- ...
Extra args are passed to
next.comb.
Value
an object of type selector_factory that can fit the
BOIN COMB model to outcomes.
References
Lin, R., & Yin, G. (2017). Bayesian optimal interval design for dose finding in drug-combination trials. Statistical methods in medical research, 26(5), 2155-2167.
Examples
num_doses <- c(3, 4)
target <- 0.25
boin_fitter <- get_boin_comb(num_doses = num_doses, target = target)
x1 <- fit(boin_fitter, outcomes = "1.1NNN")
x1
#> Patient-level data:
#> # A tibble: 3 × 5
#> Cohort Patient Dose_string Dose Tox
#> <int> <int> <chr> <list> <int>
#> 1 1 1 1.1 <int [2]> 0
#> 2 1 2 1.1 <int [2]> 0
#> 3 1 3 1.1 <int [2]> 0
#>
#> Dose-level data:
#> # A tibble: 13 × 8
#> dose tox n empiric_tox_rate mean_prob_tox median_prob_tox admissible
#> <ord> <dbl> <dbl> <dbl> <dbl> <dbl> <lgl>
#> 1 NoDose 0 0 0 0 0 TRUE
#> 2 1.1 0 3 0 0.0161 0.000000215 TRUE
#> 3 1.2 0 0 NaN 0.5 0.500 TRUE
#> 4 1.3 0 0 NaN 0.5 0.500 TRUE
#> 5 1.4 0 0 NaN 0.5 0.500 TRUE
#> 6 2.1 0 0 NaN 0.5 0.500 TRUE
#> 7 2.2 0 0 NaN 0.5 0.500 TRUE
#> 8 2.3 0 0 NaN 0.5 0.500 TRUE
#> 9 2.4 0 0 NaN 0.5 0.500 TRUE
#> 10 3.1 0 0 NaN 0.5 0.500 TRUE
#> 11 3.2 0 0 NaN 0.5 0.500 TRUE
#> 12 3.3 0 0 NaN 0.5 0.500 TRUE
#> 13 3.4 0 0 NaN 0.5 0.500 TRUE
#> # ℹ 1 more variable: recommended <lgl>
#>
#> The model targets a toxicity level of 0.25.
#> The model advocates continuing at dose 1.2.
x2 <- fit(boin_fitter, outcomes = "1.1NNN 2.1TNT")
x2
#> Patient-level data:
#> # A tibble: 6 × 5
#> Cohort Patient Dose_string Dose Tox
#> <int> <int> <chr> <list> <int>
#> 1 1 1 1.1 <int [2]> 0
#> 2 1 2 1.1 <int [2]> 0
#> 3 1 3 1.1 <int [2]> 0
#> 4 2 4 2.1 <int [2]> 1
#> 5 2 5 2.1 <int [2]> 0
#> 6 2 6 2.1 <int [2]> 1
#>
#> Dose-level data:
#> # A tibble: 13 × 8
#> dose tox n empiric_tox_rate mean_prob_tox median_prob_tox admissible
#> <ord> <dbl> <dbl> <dbl> <dbl> <dbl> <lgl>
#> 1 NoDose 0 0 0 0 0 TRUE
#> 2 1.1 0 3 0 0.0161 0.000000215 TRUE
#> 3 1.2 0 0 NaN 0.5 0.500 TRUE
#> 4 1.3 0 0 NaN 0.5 0.500 TRUE
#> 5 1.4 0 0 NaN 0.5 0.500 TRUE
#> 6 2.1 2 3 0.667 0.632 0.525 TRUE
#> 7 2.2 0 0 NaN 0.632 0.525 TRUE
#> 8 2.3 0 0 NaN 0.632 0.525 TRUE
#> 9 2.4 0 0 NaN 0.632 0.525 TRUE
#> 10 3.1 0 0 NaN 0.632 0.525 TRUE
#> 11 3.2 0 0 NaN 0.632 0.525 TRUE
#> 12 3.3 0 0 NaN 0.632 0.525 TRUE
#> 13 3.4 0 0 NaN 0.632 0.525 TRUE
#> # ℹ 1 more variable: recommended <lgl>
#>
#> The model targets a toxicity level of 0.25.
#> The model advocates continuing at dose 1.1.