PSMATCH help is needed I have a data set where "pregnancy" is the treatment. As it is a time-dependent variable I would like to know if anybody knows whether this is applicable to psmatch. My data setup is the following: The baseline date is the date of diagnosis where none of the included women have children - then some have children later on and others do not. At baseline there are the usually variables I would like to include in my ps score. The case is this: Women X has a child at day 132 after diagnosis - she needs a match with someone without a child at this day or later on to avoid conditioning on the future as I would if I did it by logistic regression based on "pregnancy at baseline" - as this only happens after baseline - and at different timepoint furthermore. I can make proc phreg make a ps_score using the xbeta command including the time-dependent nature of the pregnancy - but then I am lacking all the plots of the variables that I mostly appreciate. I do have a ps_score and the day of the pregnancy eg. 132 from the ph reg if that is applicable somehow into psmatch procedure? Does anybody knows how to fix this problem - if possible?
... View more
