Proc lifetest : how to make assumptions about when the population is exposed to the risk ?

[sossolafrip]

Hello ! I’m wondering about the proc lifetest. I’m studying the risk of having a first child using answers using retrospective data from a survey. My « basic » time event analysis would be :

proc lifetest;

  time duree*event(0);

  run;

with « duree » being the time variable and « event »  being the censoring one, and now I want to do the time event analysis while considering that women are exposed to the risk of having a first child only between ages 25 and 35 (it’s just a question to practice our understanding of time event analysis in class, so it doesn’t have any « real » study implication).

I actually know what the code would be using another software (stata, where it would be : Stset duree, failure (event) origin (time 25) exist (time 35)), but I haven’t found a way to do the same thing on SAS.

Are there any options in the proc lifetest that can be used in order to make assumptions about when the population is exposed to the risk ?

Thanks in advance ! (And apologies for the potential mistakes, I’m French).

[greveam]

You can use the delayed entry function to restrict comparisons to 25-35 years:

https://support.sas.com/documentation/onlinedoc/stat/ex_code/141/phrcpa1.html

However, when you are using parity as the outcome, you should also account for pertinent competing risks (at least all-cause mortality):

https://support.sas.com/rnd/app/stat/papers/2014/competingrisk2014.pdf