I have data of patients surviving or not at 90 days after an event. Event happened on different days. Should I create a separate column in my data calculating surviving days for each patient and using this in my analysis?
1
-
1$\begingroup$ Yes, the quantity you're interested in is time after the event happened. That the events occur on different days is no big deal, unless when the event occurs somehow influences survival time. $\endgroup$– Demetri PananosCommented May 1, 2022 at 3:43
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
It's more informative to use actual survival times rather than a 90-day survival cutoff. That also allows you to use information from patients whose last data were from a time less than 90 days after the event that you are specifying as time = 0 for each patient. A patient who hasn't yet died is treated as having a right-censored survival time at the last observation time, whether that's at 90 days or earlier. That's readily handled by standard survival analysis.
The actual date of a patient's event might be interesting to include as a covariate, particularly if the study extends over a long period of time during which there might have been changes in things like standard therapy for the disease in question. If this study only covers a short extent of calendar time, including the date probably won't help and you would be better off including other covariates expected, based on your understanding of the subject matter, to be associated with outcome.
