Hi all, I have a dataset of censored observations which contains 2 variables which I wish to account for in a time-dependent manner (tumour recurrence and subsequent treatment of that recurrence). Hence I need to convert the data to counting process notation. survSplit and the like do a nice job of doing this for a single event variable - is there anything that can do it for 2 or more events? Can survSplit be adapted or is there another way (which saves me having to write my own code)? Cheers Scott _____________________________ Scott Williams MD Peter MacCallum Cancer Centre Melbourne, Australia