Hi,
I have a propensity score matched dataset which was created as follows:
Each observation in the treatment group represents one subject.
Each treated observation was then matched to an observation from a control group, where several treated observations were matched to the same control subjects (although to different observations at different points in time).
I would like to run a logistic regression on the matched data. I understand that I need to account for the matched data (and thought about conditional logistic regression using proc logistic with the strata statement), but I also need to account for the multiple observations for some of the control subjects (which I understand that I do with proc genmod with the repeated statement).
However, the proc logistic doesn't allow to account for the repeated observation and the proc genmod doesn't allow to include the strata command.
Any suggestions how to do this?
Thanks!