fit_single_exp_covar {BPrinStratTTE}

Fit single model to data from a two-arm trial with an exponentially distributed time-to-event endpoint and one predictor of the intercurrent event

Description

Fit single model to data from a two-arm trial with an exponentially distributed time-to-event endpoint and one predictor of the intercurrent event

Usage

fit_single_exp_covar(data, params, summarize_fit = TRUE)

Arguments

data

Data frame of a structure as generated by sim_dat_one_trial_exp_covar().

params

List, containing model parameters:

• tg Positive integer value, number of intervals to calculate restricted mean survival time using the trapezoidal rule.

• p Positive integer value, number of predictors of the intercurrent event of interest (i.e. the event that determines. the principal stratum membership).

• prior_delta px2 matrix of positive numerical values, containing normal priors (mean and standard deviation) of the model parameter delta.

• prior_0N Numeric vector of length 2, containing parameters (alpha, beta) of the gamma prior on lambda_0N.

• prior_1N Numeric vector of length 2, containing parameters (alpha, beta) of the gamma prior on lambda_1N.

• prior_0T Numeric vector of length 2, containing parameters (alpha, beta) of the gamma prior on lambda_0T.

• prior_1T Numeric vector of length 2, containing parameters (alpha, beta) of the gamma prior on lambda_1T.

• t_grid Numeric vector of length tg, containing time points defining the time grid (in months) to calculate restricted mean survival time using the trapezoidal rule.

• chains Positive integer value, specifying the number of Markov chains.

• n_iter Positive integer value, specifying the number of iterations for each chain (including warmup).

• warmup Positive integer value, specifying the number of warmup (aka burnin) iterations per chain.

• cores Positive integer value, specifying the number of cores to use when executing the chains in parallel.

• open_progress Logical value, indicating whether the progress of the chains will be redirected to a file that is automatically opened for inspection.

• show_messages Logical value, indicating whether to print the summary of informational messages.

summarize_fit

Logical, if TRUE (default), the output is restricted to a summary of results on key parameters over all chains, if FALSE, the complete stanfit object is returned.

Details

The data supplied as params are used either as priors (prior_delta, prior_0N, prior_1N, prior_1T), to inform the model setup (tg, p, t_grid), or as parameters to rstan::sampling() which is invoked internally (chains, n_iter, warmup, cores, open_progress, show_messages).

Value

tibble() containing a summary of results on key parameters, or a stanfit object (S4 class), depending on summarize_fit.

See Also

fit_single_exp_nocovar() and rstan::sampling()

Examples

d_params_covar <- list(
  n = 1000,        
  nt = 500,       
  prob_X1 = 0.4, 
  prob_ice_X1 = 0.5, 
  prob_ice_X0 = 0.2,
  fu_max = 48*7,       
  T0T_rate = 0.2,     
  T0N_rate = 0.2,     
  T1T_rate = 0.15,     
  T1N_rate = 0.1
 )
dat_single_trial <- sim_dat_one_trial_exp_covar(
  n = d_params_covar[["n"]], 
  nt = d_params_covar[["nt"]],
  prob_X1 = d_params_covar[["prob_X1"]],
  prob_ice_X1 = d_params_covar[["prob_ice_X1"]],
  prob_ice_X0 = d_params_covar[["prob_ice_X0"]],
  fu_max = d_params_covar[["fu_max"]],  
  T0T_rate = d_params_covar[["T0T_rate"]],
  T0N_rate = d_params_covar[["T0N_rate"]],
  T1T_rate = d_params_covar[["T1T_rate"]],
  T1N_rate = d_params_covar[["T1N_rate"]] 
)
m_params_covar <- list(
  tg = 48,
  p = 2, 
  prior_delta = matrix(
    c(0, 5, 0, 5),
    nrow = 2, byrow = TRUE),
  prior_0N = c(1.5, 5),
  prior_1N = c(1.5, 5),
  prior_0T = c(1.5, 5),
  prior_1T = c(1.5, 5),
  t_grid =  seq(7, 7 * 48, 7) / 30,
  chains = 2,
  n_iter = 3000,
  warmup = 1500,
  cores = 2,
  open_progress = FALSE,
  show_messages = FALSE   
)

fit_single <- fit_single_exp_covar(
  data = dat_single_trial,
  params = m_params_covar,
  summarize_fit = FALSE
)
print(fit_single)