Skip to contents

Multivariable Cox Proportional Hazards Modeling Engine

Source: R/cbe_cox_multi.R
cox_multi.Rd

Fits a multivariable Cox proportional hazards model, validates the proportional hazards assumption for every term (including the multivariate global test), formats coefficients with explicit reference levels grouped by variable, and gathers model fit and convergence metrics.

Usage

cox_multi(
  data,
  formula = NULL,
  outcome = "outcome",
  features = NULL,
  conf_level = 0.95,
  ...
)

Arguments

data

A data frame containing survival data and the candidate features.

formula

Optional. A model formula with a survival::Surv() outcome, e.g. Surv(time, status) ~ age + sex. When supplied, outcome and features are ignored.

outcome

Character string naming the Surv object or outcome column in data (e.g., "outcome" or "Surv(survival_time, status)"). Ignored if formula is supplied.

features

Character vector of candidate predictor columns in data. Required if formula is not supplied.

conf_level

Numeric confidence level (default: 0.95).

...

Additional arguments passed to survival::coxph (e.g., id, ties, weights).

Value

An object of class cbe_cox_multi containing:

  • model: The fitted survival::coxph object.

  • table: Tidy coefficient table (HR, log-HR, CI, p-value) with explicit reference rows per factor, grouped by Variable.

  • glance: One-row data frame of model goodness-of-fit metrics.

  • zph: A cbe_cox_check object (includes the multivariate global test).

  • converged: Logical; whether the fit converged within the iteration limit.

  • n_iterations: Number of Newton-Raphson iterations used by the fit.

  • features: Character vector of feature names used.

  • var_labels: Named character vector mapping feature names to labels.

See also

[cox_single()], [cox_check()], [cbe_cox_table()], [plot_cox_forest_multi()]