[Marcos Almeida]

I guess Stata is fully correct.

The observation with event before the starting of the study should be excluded. I don't know the reason for having the same results at the present time, but maybe if this observation is - artificially - set to the day the study starts, well, maybe, you'll get the same results.

Again, as I remarked 4 years ago, Stata is not to a fault in this scenario. Quite on the contrary, it did exactly what should be done.

Please give us a feedback concerning this attempt to reach your goal.

[Stine Rognan]

Thank you Marcos for your quick reply.
To correct myself, I don't want to exclude the observation because the event happened on same day as the follow-up started. The patient was discharged and re-admitted on the same day.
Now, however, I found that by changing the time to event value manually from "0" to "0.01", I get the same result as in SPSS.

[Marcos Almeida]

Thank you for the feedback and for reporting the suggestion in #16 solved the issue. That said, as also, remarked, this is just to get identical results in SPSS as well as in Stata.