Welcome to the hotseat. We've prepared a guide if you'd like to read more about how it works.

People exposed to risk (when calculating Kaplan-Meier Estimate)

+3 votes
asked Apr 11, 2017 in BUS 3018F - Models by rohin_jain (1,040 points)

Hi there.

Unfortunately I cannot copy a question here since I do not have an electronic copy of ActEd (apologies for this). But this is related to Question 8.18 (page 36 of chapter 8).

It is a question about 12 patients who have an operation and time 0 is the start of the period of investigation. t is the time since having the operation. 

Some patients only have the operation after the investigation starts (say the 4th week, for example). My issue is this: 

Why are these individuals (ones who have operation only at time 4) included in Nj when calculating the estimates of the hazard function (for time periods before time 4)?


1 Answer

0 votes
answered Apr 11, 2017 by nadeem_fakier (240 points)
selected Apr 12, 2017 by Rowan
Best answer


I think its important that you consider what the question is asking. As you have mentioned, t is the time since having the operation. This means that time 4 refers to 4 weeks after the operation for all the patients. You should not think of time 4 as a date (as in the 4th week since the period of investigation in this case). I hope that this clarifies matters.

commented Apr 12, 2017 by Nikhil (110 points)
If the variable of interest were survival after the onset of illness, for example, would late entry to the study then affect the number of people under observation?
commented Apr 15, 2017 by nadeem_fakier (240 points)

This would affect the number of people under observation as there is a larger sample of individuals being exposed to the form of treatment described.