Kaplan-Meier Survival Estimator
Estimate survival over time with the product-limit estimator from right-censored durations.
Description
Estimate a right-censored survival curve with the Kaplan-Meier product-limit method.
Kaplan-Meier estimation summarizes time-to-event data when some subjects leave observation without an event. It is widely used for survival, reliability, retention, and other time-to-event analyses.
When to use Kaplan-Meier Survival Estimator
- Estimate survival at distinct event times
- Handle right-censored observations
- Build a step table for plotting a survival curve
How the calculation works
At each distinct event time, survival is multiplied by 1 − deaths/at-risk. Censored subjects remain in the risk set through their censoring time but do not create a downward step.
S(t) = product over event times tᵢ ≤ t of (1 − dᵢ/nᵢ) Interpreting the result
Each row is time, cumulative survival, at-risk count, and event count. The curve changes only at observed events. Final survival can remain above zero when the last observations are censored.
Important limitations
- Event indicators must be 1 and right-censoring indicators 0.
- The tool does not produce confidence intervals, median survival, or group comparisons.
- Kaplan-Meier assumes censoring is non-informative with respect to future event risk.