0
$\begingroup$

My question pertains to how to calculate censoring proportion in the perspective of recurrent event data in which an individual subject can have multiple survival times related to repeated occurrences of an event. As an example, in this data an individual can be censored throughout follow-up while others can experience one or more events and get censored afterwards and some have no censored observations.

When thinking about how to calculate censoring proportion in such data two approaches comes to mind. One is to divide the number of censored observations by the total number of observations and the next is to divide the number of censored observations by the number of individuals under study.

I am not sure which of these is appropriate. Any help on this is greatly appreciated. Links to any literature will also be helpful.

$\endgroup$
0
$\begingroup$

unless there are different types of events such that eg a patient may be censored for one type of skin cancer (no lesion detected) and uncensored for another (lesion detected), the proportion censored, whether you have recurrent events or not, is simply (no. of patients who experienced no events)/(total no. of patients). You could define exactly what you mean by censoring and then give the proportion that relates to that definition. However, there is a conventional definition and you might confuse your readers if you start to talk about the proportion censored within given time periods, unless it makes sense according to the design of the study

$\endgroup$

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.