Kaplan Meier curve for recurrent event data

#1
Hi,
Can anyone help. I would like to do Kaplan Meier curves for some recurrent event data i.e. each patient has multiple events. Can I use a Kaplan Meier curve for this or do I need to restrict the data to one observation per patient i.e. a sort of time to first event endpoint? Can anyone advise as to what would be the correct approach if I have recurrent event data with two treatment or more groups?
Thanks
w
 

Link

Ninja say what!?!
#3
Kaplan Meyer is really only good when you want to control for a dichotomous variable. Once you start getting more complicated with more variables or multiple values in a variable, Kaplan Meyer starts becoming insufficient. If you want to control for multiple variables, I would suggest a Cox-proportional hazards model.