Kaplan meier curves are usually calculated for survival probabilities over time. Is it valid to calculate kaplan meier curves for survival probabilities over other continuous variables e.g. over height, over levels of a biomarker? Thanks


The thing about time is you experience each value of it, and death/failure (or whatever other one-off event is considered) is a risk at each point.

With levels of a biomarker you're presumably not experiencing each lower level between 0 and the current value along the way to experiencing the current value.

If you mean to follow single individuals over time, you could perhaps substitute height for age, at least up to the point where height no longer changes.

Similarly one might consider length along a wire to the first fault.

  • $\begingroup$ Thanks, I see your point about experiencing each point in time in contrast to other variables. I've got 200 patients, a continuous biomarker for each, and a binary event for each. I'm trying to calculate/plot event probability over the values of the biomarker. Initially I thought this would be pretty straight forward, but I can't figure out how to do it. Any help would be greatly appreciated. $\endgroup$ – Scarper Sep 10 '17 at 11:24
  • $\begingroup$ Perhaps if you describe the circumstances and what you want to find out in a new post. I'm not clear on what "event probability over the values of the biomarker" is (at least not currently clear enough to suggest a model). $\endgroup$ – Glen_b Sep 10 '17 at 16:32

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.