I understand 'when' we right censore a survival time. But why do we do so? And how does it affect analysis in survival context. I would appreciate it anyone could recommend some reference.
$\begingroup$
$\endgroup$
2
-
$\begingroup$ See also What exactly are censored data? & What is the difference between censoring and truncation?. $\endgroup$– Scortchi - Reinstate Monica ♦Oct 6, 2015 at 8:22
-
1$\begingroup$ Please note for future reference that titles like this are useless for others searching the forum. A title like "What is the justification for right censoring" would have been fine, except that the question has been answered already. $\endgroup$– Nick CoxOct 6, 2015 at 8:46
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Right censoring is done so that we can make use of the data that those subjects that are censored contribute with. Since we do not know if they have an event after the time of drop-out, we cannot consider them as having had an event. So we either censor them or we exclude them from the analyses.
If we have a subject that is censored at 1000 days, we know that the subject has survived at least 1000 days, and the subject will contribute to the survival model with data for these 1000 days. If all of the subjects have an event, this is not an issue, but in many cases, the event rates are low and we need to make use of the censored subjects to increase power in our analyses.
Note that right censoring relies on the assumption that it is non-informative, so that those who drop out should do so due to reasons unrelated to the study.
I can recommend Survival Analysis: A Self-Learning Text by Kleinbaum and Klein.
