Set Up a GBSG-Based AFT Data Generating Mechanism

Creates a GBSG-based data generating mechanism that is fully compatible with simulate_from_dgm. This is the replacement for create_gbsg_dgm(): it accepts exactly the same arguments and produces the same numeric output, but returns an object of class "aft_dgm_flex" instead of "gbsg_dgm".

Usage

setup_gbsg_dgm(
  model = c("alt", "null"),
  k_treat = 1,
  k_inter = 1,
  k_z3 = 1,
  z1_quantile = 0.25,
  n_super = 5000L,
  cens_type = c("weibull", "uniform"),
  use_rand_params = FALSE,
  seed = 8316951L,
  verbose = FALSE
)

Arguments

model

Character. Either "alt" for alternative hypothesis with heterogeneous treatment effects, or "null" for uniform treatment effect. Default: "alt"

k_treat

Numeric. Treatment effect multiplier applied to the treatment coefficient from the fitted AFT model. Values > 1 strengthen the treatment effect. Default: 1

k_inter

Numeric. Interaction effect multiplier for the treatment-subgroup interaction (z1 * z3). Only used when model = "alt". Higher values create more heterogeneity between HR(H) and HR(Hc). Default: 1

k_z3

Numeric. Effect multiplier for the z3 (menopausal status) coefficient. Default: 1

z1_quantile

Numeric. Quantile threshold for z1 (estrogen receptor). Observations with ER <= quantile are coded as z1 = 1. Default: 0.25

n_super

Integer. Size of super-population for empirical HR estimation. Default: 5000

cens_type

Character. Censoring distribution type: "weibull" or "uniform". Default: "weibull"

use_rand_params

Logical. If TRUE, modifies confounder coefficients using estimates from randomized subset (meno == 0). Default: FALSE

seed

Integer. Random seed for super-population generation. Default: 8316951

verbose

Logical. Print diagnostic information. Default: FALSE

Value

An object of class c("aft_dgm_flex", "gbsg_dgm", "list") with all fields from create_gbsg_dgm() plus:

df_super

Super-population data frame with simulate_from_dgm()-compatible column names.

model_params$tau

Copy of model_params$sigma.

model_params$censoring

Sub-list with type, mu, tau for the censoring model.

Details

Internally the function calls create_gbsg_dgm() and then:

1. Adds a df_super field with column names aligned to simulate_from_dgm() conventions (lin_pred_1, lin_pred_0, lin_pred_cens_1, lin_pred_cens_0, flag_harm).

2. Adds a model_params$tau field (= model_params$sigma) and a model_params$censoring sub-list.

3. Sets class to c("aft_dgm_flex", "gbsg_dgm", "list").

The original df_super_rand field is kept so that compute_dgm_cde() and print.gbsg_dgm continue to work.

See also

create_gbsg_dgm, simulate_from_dgm, compute_dgm_cde

Examples

# \donttest{
dgm <- setup_gbsg_dgm(model = "alt", k_inter = 2, verbose = FALSE)
dgm <- compute_dgm_cde(dgm)
print(dgm)
#> GBSG-Based AFT Data Generating Mechanism (Aligned)
#> ===================================================
#> 
#> Model type: alt 
#> Super-population size: 5000 
#> 
#> Effect Modifiers:
#>   k_treat: 1 
#>   k_inter: 2 
#>   k_z3: 1 
#> 
#> Hazard Ratios (Cox-based, stacked PO):
#>   Overall (causal): 0.7331 
#>   Harm subgroup (H): 2.9651 
#>   Complement (Hc): 0.6612 
#>   Ratio HR(H)/HR(Hc): 4.4846 
#> 
#> Average Hazard Ratios (from loghr_po):
#>   AHR (overall): 0.7431 
#>   AHR_harm (H): 3.8687 
#>   AHR_no_harm (Hc): 0.5848 
#>   Ratio AHR(H)/AHR(Hc): 6.6157 
#> 
#> Subgroup definition: z1 == 1 & z3 == 1 (low ER & premenopausal) 
#> ER threshold: 8 (quantile = 0.25)
#> Subgroup size: 634 (12.7%)
#> Analysis variables: v1, v2, v3, v4, v5, v6, v7 
sim <- simulate_from_dgm(dgm, n = 400, seed = 1)
# }