0
$\begingroup$

As my topic suggests, can someone explain me what tied events actually are?

I just read that there are different algoriths of handling ties differently in Cox-Regression, but so for I did not understand what "tied event" data actually is.

Can someone give me an example?

Thanks,

KS

$\endgroup$
1
$\begingroup$

Tied events are simply when two events occur at the exact same recorded time.

For example, if a study measures time until remission in months and subjects 1 and 2 both experienced remission at month 2, then their event times are tied.

This creates a small issue for the canonical Cox regression model, as it uses the ranks of event times to fit the model. However, this is not too difficult to handle and there are several easy ways to fix it up.

$\endgroup$
  • $\begingroup$ what are ways to deal with tied events in cox regression? $\endgroup$ – mnky9800n Sep 26 '19 at 16:19
  • $\begingroup$ @mnky9800n: I don't know all of the methods, but a cheap solution is simply to break them at random $\endgroup$ – Cliff AB Sep 26 '19 at 16:43
0
$\begingroup$

The tied events are the parameters that the different variables contain at the same time occasion.

There are about 5 methods to handle the tied events in Cox regression. Berslow, Efron, Exact, distinct, and random. according to http://soep.ue.poznan.pl/jdownloads/Wszystkie%20numery/Rok%202014/06_borucka.pdf the erfron method performes better in terms of time, fit statistics, and differences in parameters estimates.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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