1
$\begingroup$

I created a Kaplain Meier disease free survival curve with the following statistics (in months):

      n  events  median 0.95LCL 0.95UCL 
  127.0    36.0    59.6    45.0      NA 

After subsetting, for the 36 patients with disease onset I calculated the follow-up distribution (in months):

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  2.104   9.485  25.940  25.260  35.660  59.570 

I understand that the median KM survival means the time to 50% disease onset, but I was wondering if my follow-up distribution make sense with this the KM analysis? If I presented both of these numbers how would I relate these two calculations?

$\endgroup$

1 Answer 1

0
$\begingroup$

The two distributions would only be equal if the censoring is uniform, which is a stronger condition than non-informativeness or independence.

As such, it does not mean anything. For instance, if you have a fixed censoring time for all the individuals, then the distribution of the subset which has an event will always underestimate the real distribution of the events, simply because you will not observe events that would happen later.

$\endgroup$

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.