st: Puzzling wrong result with stptime compared to strate under special circumstances

["Jesper Lindhardsen" <JESLIN01@geh.regionh.dk>]

Hi All,

This may be obvious, but I cannot get my head around this problem, which
I ran into in the cause of some datamanipulation that required -stset-
twice.
Specifically, using the Stata generated underscore variables for the
second -stset- with the -origin- option gave different eventrates
depending on whether -strate- or -stptime- was used.
I find that -strate- and -stptime- give identical estimates before the
new stset, while -stptime- gives a wrong estimate after.
Please see example code with use of catheter.dta
(http://www.stata-press.com/data/r9/catheter.dta)

I am using an fully updated 11.2 on a Windowx XP box



Code begin:

sysuse catheter,clear
stset time,f(infect) id(patient) enter(time 10)
strate
stptime
stset _t,f(_d) id(patient) origin(time 10)
strate
stptime
g risk=_t-_t0
tabstat risk,s(sum)

Code End:

Any thoughts ???
BW,

Jesper

Jesper Lindhardsen
MD, PhD candidate
Department of Cardiovascular Research
Copenhagen University Hospital, Gentofte
Denmark


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/