Kaplan-Meier Curve

What a Kaplan-Meier curve shows, how to read it, and what information appears alongside the figure.

A Kaplan-Meier curve is a step function that estimates the probability of surviving — or remaining event-free — past each observed time point. It is the standard way to visualize time-to-event data and is the starting point for any survival analysis.

What the figure shows

The survival curve

The y-axis shows the estimated survival probability, from 1.0 (all subjects event-free) at the start down toward 0 as events accumulate. The x-axis shows time. The curve drops at each time point where an event occurs and holds flat between events.

Group curves

When a group column is present, each group is shown as a separate curve in a distinct color. The colors follow the same palette used in other Licklider group comparison figures.

Censoring markers

Observations where the event did not occur by the end of follow-up — censored observations — are shown as small tick marks on the curve at the time they left the study. Each tick sits on the curve at the survival probability estimated at that time and is colored to match its group.

What appears in the Inspector

Median survival

For each group, the time at which the estimated survival probability crosses 0.5. When fewer than half of subjects have experienced the event, the median cannot be estimated and is shown as not reached.

Log-rank test

When two or more groups are present, the log-rank p-value is shown. For three or more groups, pairwise comparisons with correction for multiple testing are also available.

Number at risk

The number of subjects still under observation at selected time points is shown in the Inspector panel. This table is an expected part of any published survival figure.

Reading the figure

A curve that drops steeply early indicates that many events occurred early in follow-up. A curve that flattens and holds steady at a non-zero probability indicates that some subjects did not experience the event within the observation period.

When two curves separate widely and do not cross, the groups differ consistently in their survival experience. When curves cross, the survival difference is not uniform over time — this has implications for the proportional hazards assumption if Cox regression is also used.

Disclosure requirements

Before a Kaplan-Meier result can be used in a claim-bearing export, Licklider requires:

  • Censoring must be acknowledged
  • At-risk information must be available
  • If a median survival time is cited, it must be estimable from the data

For more detail → see Survival Data Detection Guard.

What this page does not cover