Skip to content
Omnikalc

Kaplan–Meier survival calculator

Enter the number of patients, and the events and censored patients in each of five intervals, to get the survival curve and its 95% confidence interval.

Updated on

Educational tool, not medical advice: this page does not diagnose and does not replace a doctor, nurse or pharmacist. It has not been reviewed by a healthcare professional. Editorial and medical policy

Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Value out of range

Results

Survival at the end of interval 5

72.6 %

Survival at the end of interval 1
92.0 %
Survival at the end of interval 2
85.9 %
Survival at the end of interval 3
80.6 %
Survival at the end of interval 4
76.1 %
Lower 95% limit at interval 5 (Greenwood)
63.5 %
Upper 95% limit at interval 5 (Greenwood)
81.6 %
Patients still at risk at the start of interval 5
65
Median interval (0 if survival stays above 50%)
0

The Kaplan–Meier survival calculator gives the survival at the end of each of five intervals, the 95% confidence interval at the last interval, the patients at risk and the median interval, from the events and censored patients in each interval.

How the calculation works

  • Enter the number of patients at the start.
  • Enter the events and the censored patients of each of the five intervals.
  • Read the survival at each interval and the confidence interval at the last one.

Example

From 100 patients, the survival is 92.0 %, 85.9 %, 80.6 %, 76.1 % and 72.6 % at the end of the five intervals. At the last interval, 65 patients are at risk, and the 95% interval runs from 63.5 % to 81.6 %. The median code is 0.

Why it matters

The calculator helps students of biostatistics, epidemiology and clinical research who learn survival analysis, and readers of studies who want to reproduce the logic of a curve and see the effect of censoring on the estimate.

Good to know

This is a simplified life-table version. Events and censoring are assumed at the end of each interval, and at least one patient must remain at risk. Exact Kaplan–Meier curves use each event time, and a log-rank test or a Cox model compares groups, which this tool does not. Censoring must be non-informative, that is unrelated to the risk of the event, or the estimate is biased. Confidence limits are wide when few patients remain at risk, so look at the numbers at risk below the curve.

The median survival calculator gives the median from a constant hazard, the events needed in survival trials calculator plans the size of a study, the Weibull distribution calculator models lifetimes and the probability calculator handles basic chances.

Frequently asked questions

What is the Kaplan–Meier estimator?

It is the standard way to estimate the probability of surviving beyond a given time from data in which some patients are lost to follow-up or are still alive at the end of the study. At each interval, the survival is multiplied by the share of patients at risk who did not have the event, and the product gives the curve.

What is censoring?

A patient is censored when the event has not been seen by the end of the follow-up, because the study ended, the patient moved or dropped out. A censored patient counts in the risk set until the moment of censoring and is then removed, without being counted as an event. Ignoring them would bias the survival estimate.

How is the confidence interval computed?

The calculator uses Greenwood’s formula for the variance: Var(S) = S² × Σ d / (n (n − d)), where d is the number of events and n the number at risk in each interval, and it gives S ± 1.96 standard errors, limited between 0 and 100 percent. For small numbers or survival near 0 or 1, other methods, such as a log-log transformation, are preferred.

How is this simplified compared with software?

A full Kaplan–Meier analysis uses each event time, while this tool groups events into five intervals and treats all events and censoring as happening at the end of the interval. It gives a good idea of the method and approximate values, but a trial report should use the exact times and a statistical package.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.