schemaKm

Kaplan-Meier association schema

Description

schemaKm() defines the result and artifact contract for grouped Kaplan-Meier analyses produced by associate$regression(model = “km”).

Usage

schemaKm()

Value

A schema definition list.

Result semantics

Kaplan-Meier analyses are not coefficient models. The main result table contains one omnibus row per fit, usually a log-rank comparison of the stored curves.

  • outcome: endpoint/event-of-interest indicator.

  • time: observed time or age at event or censoring.

  • predictor: grouping variable used to define Kaplan-Meier curves.

  • term: tested grouping variable.

  • term.type: omnibus.

  • effect.scale: median.time.

  • estimate, se, lower, upper: usually missing for the omnibus row; group-level medians live in group.summary.

  • statistic: log-rank chi-square statistic.

  • p.value, adj.p.value: raw and adjusted log-rank p-values.

  • n, n.events: complete observations and endpoint events used by the curves.

  • formula: displayed survival formula.

Artifacts

observed.curves is required and contains one row per curve step with: fit.id, curve.id, curve.label, curve.type, time, estimate, lower, upper, n.risk, n.event, and n.censor.

Optional risk.table contains time-indexed numbers at risk. Optional group.summary contains group-level n, n.events, median.time, lower, and upper values.