var_surv_weibull_bootstrap: Bootstrap Variance Estimation for Weibull Survival Functions

Description

Estimates bootstrap variance for counterfactual survival functions from surv_weibull(). Supports binary and multiple treatment groups.

Usage

var_surv_weibull_bootstrap(
  data,
  surv_result,
  treatment_var,
  time_var,
  event_var,
  censoring_formula,
  censoring_control = list(maxiter = 350),
  B = 100,
  parallel = FALSE,
  mc.cores = 2,
  seed = NULL,
  boot_level = "full"
)

Value

List containing:

var_matrix: Matrix [time x group] of bootstrap variances.
se_matrix: Matrix [time x group] of bootstrap standard errors.
boot_samples: List of length B with survival matrices from each iteration.
boot_allocation: Matrix [B x group] of sample sizes for each group in each bootstrap iteration.
n_success_by_group: Matrix [time x group] of successful iterations.
n_failed_by_group: Matrix [time x group] of failed iterations.
B: Total bootstrap iterations.

Arguments

data: Data frame used in original surv_weibull() call.
surv_result: Output from surv_weibull().
treatment_var: Name of treatment variable.
time_var: Name of time variable.
event_var: Name of event indicator variable.
censoring_formula: Formula for censoring score model.
censoring_control: Control parameters for survreg(). Default list(maxiter = 350).
B: Number of bootstrap iterations. Default 100.
parallel: Logical. If TRUE, use parallel computation via mclapply. Default FALSE.
mc.cores: Number of cores for parallel computation. Default 2.
seed: Optional random seed for reproducibility. Ensures identical results across runs and between sequential and parallel execution. Default NULL.
boot_level: Bootstrap sampling level: "full" (default) or "strata". "full" resamples from entire dataset (observational studies). "strata" resamples within treatment groups preserving group sizes (RCTs).

Details

Each bootstrap iteration resamples observations with replacement and re-estimates propensity scores, weights, and survival functions using the same specifications as the original analysis. Treatment groups may fail independently if not sampled or if models fail to converge.