Diagnose Drift Between Observed and Natural-Course TVCs — diagnose_tvcqgcomp

Compare observed and simulated natural-course trajectories over time for time-varying covariates. Categorical variables are summarized through time-specific level proportions, while continuous variables are summarized through time-specific means, standard deviations, and selected quantiles.

Usage

diagnose_tvcqgcomp_drift(observed_data = NULL, natural = NULL, run_obj = NULL,
  config = NULL, variables = NULL, include = c("both", "categorical",
  "continuous"), continuous_probs = c(0.1, 0.5, 0.9))

Arguments

observed_data: Observed longitudinal data used as the calibration reference.
natural: Simulated natural-course data to compare against the observed data.
run_obj: Optional object returned by tvcQGComp_survival(). When supplied, config, observed_data, and natural are filled from the run object when available.
config: Optional configuration object. If omitted, run_obj$config is used.
variables: Optional character vector of variables to check. Defaults to the configured time-varying covariates.
include: Whether to return checks for "categorical", "continuous", or "both" types.
continuous_probs: Quantiles to compare over time for continuous variables.

Value

A list with the inferred variable types and separate detailed and summary tables for categorical and continuous drift over time.