Simulate results from a single arm Bayesian Phase 2 Trial

Trial is a single arm Bayesian phase 2 trial with sequential stopping boundaries as outlined in Thall, Peter F., and Richard Simon. "Practical Bayesian guidelines for phase IIB clinical trials." Biometrics (1994): 337-349.

ph2_single_bayes_seq_sim(
  omega_s_mu,
  omega_s_width,
  omega_s_w_conf_level = 0.9,
  omega_e_c,
  delta_0,
  n_min,
  n_max,
  sim_n = 1,
  pr_low = 0.05,
  pr_high = 0.95,
  mu_e = omega_s_mu + delta_0,
  verbose = FALSE,
  quiet = FALSE
)

Arguments

omega_s_mu	Mean of prior distribution on omega_s
omega_s_width	Width of the probability interval (set by w_conf_level=) running from the (1 + w_conf_level) / 2 to the (1 - w_conf_level) / 2 percentiles
omega_s_w_conf_level	Confidence level of width of probabilty specified
omega_e_c	Concentration parameter for prior distribution on omega_e
delta_0	Targeted improvement of treatment for treatment E over S
n_min	Minimum number of patients that may be enrolled
n_max	Maximum number of patients that may be enrolled
sim_n	Number of simulated trials to create
pr_low	Lower probability limit for concluding treatment not promising
pr_high	Upper probablity limit for concluding treatment promising
mu_e	True rate of success with experimental treatment in omega_s_width
verbose	When TRUE, additional information is returned as an attribute
quiet	Run with no notes, progress bars, etc.

Examples

# simulate trial results
sim_results <- ph2_single_bayes_seq_sim(
  # setting priors for standard treatment
  omega_s_mu = 0.2, omega_s_width = 0.20, omega_s_w_conf_level = 0.90,
  # setting priors for experimental treatment
  omega_e_c = 2, delta_0 = 0.15,
  # other trial parameters
  n_min = 10, n_max = 65,
  pr_low = 0.05, pr_high = 0.95,
  # true effect of experimental tx
  mu_e = 0.35,
  # number of simulations
  sim_n = 1000
)

# tabulate summary
library(gtsummary)
sim_results %>%
  dplyr::select(-sim_id) %>%
  tbl_summary(
    label = list(result ~ "Trial Result",
                 n_enrolled ~ "No. Enrolled in Trial")
  ) %>%
  add_stat_label() %>%
  as_kable()
#> 
#> 
#> |**Characteristic**    |**Statistic** |**N = 1000** |
#> |:---------------------|:-------------|:------------|
#> |Trial Result          |              |             |
#> |futile                |n (%)         |158 (16%)    |
#> |inconclusive          |n (%)         |759 (76%)    |
#> |success               |n (%)         |83 (8.3%)    |
#> |No. Enrolled in Trial |median (IQR)  |65 (65, 65)  |