Survival analysis longer than follow-up

[Leandro Blas]

If this is not the correct place to post I apologyze in advance.

Hi, I am trying to make 5 and 10-y conditional survival analisys.
Is it ok measure 5-year survival if the follow-up is 30 months? I have read papers reporting 10 year survival rates with follow-up of 4 years but I cannot find a way to justify it. Can anyone kindly help me?

[Clyde Schechter]

I think if you read those papers carefully, you will see that they are reporting 10 year survival rates with a mean or median follow-up of 4 years. If the entire study cohort is only followed for 4 years, then, no, it is not possible to estimate 10-year survival rates. But if some of the cohort has passed 10 years of follow-up, and if the censoring up to that point has been uninformative, then, yes, you can get an unbiased estimate of the 10-year survival probability. Of course, if the number that have been followed out to 10 years is very small, that unbiased estimate may still be wildly imprecise and way off the mark! So you need to look carefully at the details to know what to make of it.