Skip to contents

Model-Predicted Survival Curves Stratified by Predictor

Source: R/plot_cox.R
plot_cox_survival.Rd

Generates stratified predicted survival curves from a Cox model. For continuous predictors, values are automatically binned into quantile strata (e.g. quartiles).

Usage

plot_cox_survival(
  fit,
  data,
  feature = NULL,
  id_col = NULL,
  n_tiles = 4,
  label_endpoints = TRUE,
  overlay_km = FALSE,
  base_size = 12
)

Arguments

fit

A fitted survival::coxph model or a cbe_cox object.

data

Data frame used for fitting the model.

feature

Character name of the predictor column.

id_col

Character name of the subject ID column (default: NULL, which uses the row number).

n_tiles

Number of quantile bins for continuous predictors (default: 4).

label_endpoints

Logical; if TRUE, repels labels for stratum values at the final time point.

overlay_km

Logical (default: FALSE). If TRUE, overlays observed Kaplan-Meier step curves (dashed) for each stratum on top of the Cox-predicted curves (solid), sharing color by stratum, with a linetype legend distinguishing "Cox-predicted" from "KM observed". The event time/status used are those the model was fit with (fit$y).

base_size

Base font size (default: 12).

Value

A ggplot2 object.