I wonder how acceptable it is, pros and cons, of using the censoring indicator with survival data as a binary outcome for ROC curves and logistic regression. One issue is if we have early dropout / incomplete followup then patients who drop out early have cens=0 (the good group) and those who die after years (cens=1) are in the bad group (even if they have actually survived well give the usual median survival time for a particular disease). Is Heagerty's time-dependent survival ROC a good alternative. I believe that also requires you to "pick" a time points (its a ROC curve as 2 years for example). Is Heagerty's methods used as much as we'd expect given surviva data ad ROC curves are common or is the alternative censoring indicator as a binary outcome approach ever used ?


Your Answer

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

Browse other questions tagged or ask your own question.