viz.plot_cif()

Plot cumulative incidence functions from a fitted Aalen-Johansen estimator.

Usage

Source

viz.plot_cif(
    aj,
    *,
    conf_int=True,
    risk_table=False,
    times=None,
    title=None,
    xlab="Time",
    ylab="Cumulative incidence",
    width=400,
    height=300,
    backend="altair"
)

Renders one cumulative incidence curve per competing cause as a right-continuous step function. The chart is faceted by cause (one panel per event type). Stratified fits (produced by passing by= to AalenJohansen.fit()) draw one colored line per group within each panel. An optional shaded confidence band shows the point-wise uncertainty.

Parameters

aj: AalenJohansen

A fitted AalenJohansen object. Unstratified fits draw a single curve per panel. Stratified fits draw one colored curve per group.

conf_int: bool = True

If True (default), draw a shaded point-wise confidence band around each curve.

risk_table: bool = False

If True, stack an aligned numbers-at-risk table beneath the curves.

times: Any = None

Query times for the numbers-at-risk table (used only when risk_table=True). Defaults to six evenly spaced times from 0 to the largest observed follow-up time.

title: str | None = None

Optional overall plot title.

xlab: str = "Time"

X-axis label (default "Time").

ylab: str = "Cumulative incidence"

Y-axis label (default "Cumulative incidence").

width: int = 400

Width of each cause panel in pixels (default 400).

height: int = 300

Height of each cause panel in pixels (default 300).

backend: str = "altair"

Plotting backend. Currently only "altair" is supported.

Returns

altair.Chart

An interactive Altair chart. Each panel corresponds to one competing cause. Groups are distinguished by color.

Examples

Fit the Aalen-Johansen estimator on the bundled mgus2 dataset, where patients may progress to malignancy (PCM) or die first:

import numpy as np
import greenwood as gw

# Load data and build a competing-risks response
mg = gw.load_dataset("mgus2", backend="polars")
etime = np.where(mg["pstat"] == 1, mg["ptime"], mg["futime"])
cause = np.where(mg["pstat"] == 1, 1, 2 * mg["death"])
y = gw.Surv.multistate(etime, event=cause, states=("pcm", "death"))

# Fit the Aalen-Johansen estimator and plot cumulative incidence
aj = gw.AalenJohansen().fit(y)
gw.plot_cif(aj)

Pass by= to stratify by a covariate and compare groups across causes:

# Stratify by sex and compare groups across causes
aj_sex = gw.AalenJohansen().fit(y, by=mg["sex"])
gw.plot_cif(aj_sex, title="Cumulative incidence by sex")

Add an aligned numbers-at-risk table beneath the curves:

# Add a numbers-at-risk table beneath the curves
gw.plot_cif(aj_sex, risk_table=True)