generate_aft_dgm_flex: Generate Synthetic Survival Data using AFT Model with Flexible Subgroups

Description

Creates a data generating mechanism (DGM) for survival data using an Accelerated Failure Time (AFT) model with Weibull distribution. Supports flexible subgroup definitions and treatment-subgroup interactions.

Usage

generate_aft_dgm_flex(
  data,
  continuous_vars,
  factor_vars,
  continuous_vars_cens = NULL,
  factor_vars_cens = NULL,
  set_beta_spec = list(set_var = NULL, beta_var = NULL),
  outcome_var,
  event_var,
  treatment_var = NULL,
  subgroup_vars = NULL,
  subgroup_cuts = NULL,
  draw_treatment = FALSE,
  model = "alt",
  k_treat = 1,
  k_inter = 1,
  n_super = 5000,
  select_censoring = TRUE,
  cens_type = "weibull",
  cens_params = list(),
  seed = 8316951,
  verbose = TRUE,
  standardize = FALSE,
  spline_spec = NULL
)

Value

A named list of class aft_dgm containing:

data: Simulated trial data frame with outcome, event, and treatment columns.
model_params: Model parameters used for data generation (coefficients, dispersion, spline info if applicable).
subgroup_info: Subgroup definition and membership indicators, if a heterogeneous treatment effect was specified.
censoring_info: Censoring model parameters and observed censoring rate.
call_args: Arguments used in the call, for reproducibility.

Arguments

data

A data.frame containing the input dataset to base the simulation on

continuous_vars

Character vector of continuous variable names to be standardized and included as covariates

factor_vars

Character vector of factor/categorical variable names to be converted to dummy variables (largest value as reference)

continuous_vars_cens

Character vector of continuous variable names to be used for censoring model. If NULL, uses same as continuous_vars. Default NULL

factor_vars_cens

Character vector of factor variable names to be used for censoring model. If NULL, uses same as factor_vars. Default NULL

set_beta_spec

List with elements 'set_var' and 'beta_var' for manually setting specific beta coefficients. Default list(set_var = NULL, beta_var = NULL)

outcome_var

Character string specifying the name of the outcome/time variable

event_var

Character string specifying the name of the event/status variable (1 = event, 0 = censored)

treatment_var

Character string specifying the name of the treatment variable. If NULL, treatment will be randomly simulated with 50/50 allocation

subgroup_vars

Character vector of variable names defining the subgroup. Default is NULL (no subgroups)

subgroup_cuts

Named list of cutpoint specifications for subgroup variables. See Details section for flexible specification options

draw_treatment

Logical indicating whether to redraw treatment assignment in simulation. Default is FALSE (use original assignments)

model

Character string: "alt" for alternative model with subgroup effects, "null" for null model without subgroup effects. Default is "alt"

k_treat

Numeric treatment effect modifier. Values >1 increase treatment effect, <1 decrease it. Default is 1 (no modification)

k_inter

Numeric interaction effect modifier for treatment-subgroup interaction. Default is 1 (no modification)

n_super

Integer specifying size of super population to generate. Default is 5000

select_censoring

Logical. If TRUE (default), fits the censoring distribution to the observed censoring times in data using survreg with AIC-based selection among Weibull and log-normal models (with and without covariates). If FALSE, no model is fitted; the censoring distribution is specified entirely by cens_params. Default TRUE.

cens_type

Character string specifying censoring distribution type: "weibull" or "uniform". Controls which parametric family is considered when select_censoring = TRUE, and determines the required structure of cens_params when select_censoring = FALSE. Default "weibull".

cens_params

Named list of censoring distribution parameters. Interpretation depends on select_censoring and cens_type:

select_censoring = TRUE: Ignored; all parameters are estimated from data.

select_censoring = FALSE, cens_type = "uniform"

Must supply min and max. If either is absent, defaults to 0.5 * min(y) and 1.5 * max(y) with a message.

select_censoring = FALSE, cens_type = "weibull"

Must supply mu (log-scale location) and tau (scale). Optionally supply type ("weibull" or "lognormal"); defaults to "weibull". Censoring is treated as intercept-only (no covariate or treatment dependence): lin_pred_cens_0 = lin_pred_cens_1 = mu.

Default list().

seed

Integer random seed for reproducibility. Default is 8316951

verbose

Logical indicating whether to print diagnostic information during execution. Default is TRUE

standardize

Logical indicating whether to standardize continuous variables. Default is FALSE

spline_spec

List specifying spline configuration for treatment effect. Must include 'var' (variable name), 'knot', 'zeta', and 'log_hrs' (vector of length 3). Default NULL (no spline)

Author

Your Name

Details

Subgroup Cutpoint Specifications

The subgroup_cuts parameter accepts multiple flexible specifications:

Fixed Value

subgroup_cuts = list(er = 20)  # er <= 20

Quantile-based

subgroup_cuts = list(
  er = list(type = "quantile", value = 0.25)  # er <= 25th percentile
)

Function-based

subgroup_cuts = list(
  er = list(type = "function", fun = median)  # er <= median
)

Range

subgroup_cuts = list(
  age = list(type = "range", min = 40, max = 60)  # 40 <= age <= 60
)

Greater than

subgroup_cuts = list(
  nodes = list(type = "greater", quantile = 0.75)  # nodes > 75th percentile
)

Multiple values (for categorical)

subgroup_cuts = list(
  grade = list(type = "multiple", values = c(2, 3))  # grade in (2, 3)
)

Custom function

subgroup_cuts = list(
  er = list(
    type = "custom",
    fun = function(x) x <= quantile(x, 0.3) | x >= quantile(x, 0.9)
  )
)

Model Structure

The AFT model with Weibull distribution is specified as: $$\log(T) = \mu + \gamma' X + \sigma \epsilon$$

Where:

$T$ is the survival time
$\mu$ is the intercept
$\gamma$ contains the covariate effects
$X$ includes treatment, covariates, and treatment x subgroup interaction
$\sigma$ is the scale parameter
$\epsilon$ follows an extreme value distribution

Interaction Term

The model creates a SINGLE interaction term representing the treatment effect modification when ALL subgroup conditions are simultaneously satisfied. This is not multiple separate interactions but one combined indicator.

References

Leon, L.F., et al. (2024). Statistics in Medicine.

Kalbfleisch, J.D. and Prentice, R.L. (2002). The Statistical Analysis of Failure Time Data (2nd ed.). Wiley.

Examples

Run this code

# \donttest{
df <- survival::gbsg
dgm <- generate_aft_dgm_flex(
  data            = df,
  outcome_var     = "rfstime",
  event_var       = "status",
  treatment_var   = "hormon",
  continuous_vars = c("age", "size", "nodes", "pgr", "er"),
  factor_vars     = "meno",
  model           = "null",
  verbose         = FALSE
)
str(dgm)
# }

Run the code above in your browser using DataLab