rmst_diff()

Compute the RMST difference between two groups and return a tidy DataFrame.

Usage

rmst_diff(
    surv,
    tau,
    group,
    *,
    strata=None,
    conf_level=0.95,
)

A convenience wrapper around rmst_test() with estimand="difference" that returns the result as a single-row DataFrame rather than an RMSTResult object. This is useful when you want to feed the comparison directly into a summary table or pipeline.

Parameters

surv: Surv: A right-censored Surv response built with Surv.right().
tau: float: The restriction time. Should be a clinically meaningful horizon (e.g., 365 days for one-year RMST).
group: Any: Group labels, one per observation. Must have exactly two unique levels.
strata: Any | None = None: Stratification variable for stratified RMST comparison. If provided, per-group RMST estimates are computed within each stratum and pooled with inverse-variance weights.
conf_level: float = 0.95: Confidence level for the confidence interval (the default is 0.95).

Returns

DataFrame: A single-row DataFrame with columns group1, group2, rmst1, rmst2, difference, se, lower_ci, upper_ci, statistic, and p_value.

Examples

Compute the one-year RMST difference between sexes in the lung cancer dataset:

import greenwood as gw

# Load data and build a right-censored response
lung = gw.load_dataset("lung", backend="polars")
y = gw.Surv.right(lung["time"], event=(lung["status"] == 2))

# Compute the one-year RMST difference as a tidy DataFrame
gw.rmst_diff(y, tau=365, group=lung["sex"])

shape: (1, 10)

group1	group2	rmst1	rmst2	difference	se	lower_ci	upper_ci	statistic	p_value
i64	i64	f64	f64	f64	f64	f64	f64	f64	f64
1	2	241.495085	297.46541	-55.970324	14.958126	-85.287712	-26.652936	-3.741801	0.000183