enter image description here

Specifically, the Cumulative Proportion Surviving at the Time. I've tried reading elsewhere but can't seem to wrap my head around it.


1 Answer 1


That's just the standard Kaplan-Meier survival estimate, perhaps a bit disguised with the way the data are presented. It's the fraction of the original group that is estimated to survive longer than the indicated time, with censoring taken into account.

The trick with this type of data format is to consider the number of cases still at risk just before each event time. Just before time = 2 there were 78 at risk, of whom 2 had events at time = 2. A fraction of 0.02564 (2/78) of those at risk had the event then. No one had the event previously, so a fraction of (1-0.02564) = 0.97436 survived past time = 2.

Just before time = 3 there were 72 still at risk. Of those, 1 had the event at time = 3, or a fraction 0.01389 of those still at risk then. A fraction (1-0.01389) = 0.98611 of those at risk just before time = 3 survived past that time. As a fraction 0.97436 of the original group was estimated to have survived past time = 2 up to just before time = 3, 0.97436 * 0.98611 = 0.96083 is the estimated fraction of the original group surviving past time = 3.

I suspect that the corresponding standard error estimates are based on Greenwood's formula but I haven't checked directly.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.