Marginal Event Probability / Relative Hazard Diagnostic Plot

Plots a continuous predictor against a model-derived curve, with a 95 for scale = "prob", overlaid actual binary event observations.

Usage

plot_cox_marginal(
  fit,
  data,
  feature = NULL,
  status_col = "status",
  scale = c("prob", "hr", "log_hr"),
  color = "#9D2235",
  base_size = 12
)

Arguments

fit

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

data

Data frame containing survival inputs.

feature

Character name of continuous predictor column.

status_col

Character name of status/event column (0 = censored, 1 = event; default: "status"). Only used when scale = "prob".

scale

One of "prob" (default; predicted probability of the event), "hr" (relative hazard, i.e. exp() of the centered linear predictor), or "log_hr" (the centered linear predictor itself).

color

Primary color for the fitted curve and CI band (default: Temple Cherry "#9D2235").

base_size

Base font size (default: 12).

Value

A ggplot2 object.