fit_cce_vs() fits two complementary estimators for a binary treatment
comparison: a Cox-model-based g-formula standardization, an IPTW-weighted
Kaplan-Meier curve (iptw_km), and an IPTW-weighted Cox standardization
(iptw_cox). The function returns tidy curves, effects, diagnostics, and
machine-readable metadata.
Usage
fit_cce_vs(
data,
arm = "arm",
time = "time",
event = "event",
covariates,
subgroup = NULL,
tau = NULL,
landmark_times = NULL,
n_grid = 100L,
bootstrap = 0L,
seed = 1L,
weight_cap = 50,
warning_max_weight = 10,
fail_max_weight = 50,
warning_smd = 0.1,
fail_smd = 0.2
)Arguments
- data
Analysis-ready data frame.
- arm, time, event
Column names identifying treatment assignment, follow-up time, and event indicator.
- covariates
Character vector of baseline adjustment covariates.
- subgroup
Optional subgroup column name. When supplied, the result includes overall and subgroup-specific summaries.
- tau
RMST truncation horizon. Defaults to the 90th percentile of observed follow-up.
- landmark_times
Survival-difference time points.
- n_grid
Number of time points used to summarize curves.
- bootstrap
Number of bootstrap resamples used to derive intervals.
- seed
Random seed used for bootstrap resampling.
- weight_cap
Hard upper cap applied to stabilized IPTW weights.
- warning_max_weight, fail_max_weight
Diagnostic thresholds for the largest stabilized weight.
- warning_smd, fail_smd
Diagnostic thresholds for absolute SMD.