TrialCurveMedian survival calculator
Median survival is the time at which the Kaplan–Meier curve first falls to 50%. Paste your data below to get the median for each group with its 95% confidence interval. If the curve never reaches 50%, the median is "not reached".
Paste one row per patient as arm,time,event (event 1 = event, 0 = censored), or load the example.
How the median and its CI are calculated
The median is the first event time at which the Kaplan–Meier estimate is at or below 0.5. The 95% confidence interval comes from the pointwise confidence limits of the curve: its lower end is where the lower limit crosses 0.5, and its upper end is where the upper limit does. The limits use Greenwood's variance on the log(−log) scale, which is the SAS default (R's default uses the log scale, so results can differ slightly).
Why the upper limit is often "not reached"
When few patients remain at risk, the curve's upper confidence limit may never fall to 50% within follow-up. That is common and correct: the data cannot rule out a longer median.
Median survival from a published figure
Read where the curve crosses 50%, or better, reconstruct the patient-level data and compute it here with a confidence interval. On published trials, TrialCurve reproduces printed medians within 10% for 96% of curves (accuracy).
FAQ
What does 'median not reached' mean?
The Kaplan–Meier curve stayed above 50% for the whole follow-up, so the median survival time is longer than the follow-up and cannot be estimated yet.
Is median survival the same as mean survival?
No. The mean is usually not estimable with censored data; restricted mean survival time (RMST) is its well-defined counterpart. See the RMST calculator.
Why do R and SAS give different confidence intervals?
They transform the survival estimate differently before building the interval: R uses the log scale by default and SAS the log(−log) scale.
Reconstruct IPD from your own figure. Upload a Kaplan–Meier plot and get curves, numbers at risk, pseudo-IPD, hazard ratios and a validation pack. Three figures a month are free.
Try the live demo Browse trial data API docs