st: re use of age as a time scale in Cox Models

[Andriene Simone Grant <granta@email.arizona.edu>]

Dear Statalisters,

I am using Stata/SE 12.1 for Windows (64-bit x86-64) Revision 30 Jan 2013.

I am running a Cox Proportional hazards model to investigate the
hazard of death from a given exposure and would like to use age as the
time scale.  Unfortunately the dataset that I am using does not have
the exact date of birth or date of enrollment due to confidentiality
reasons, but has the age of participants when they were enrolled in
the study. They then have follow-up time until censoring/death. To
create an age timescale variable I  used the formula:

timeage = age at enrollment + (timeonstudy/365.25).

and then used the following stset command:

stset timeage, failure(death == 1) enter(age)  - where age would be
the age at enrollment

I have been browsing the archives and have seen some discussion on age
as the timescale and the use of the 'origin' and 'enter' options with
stcox, but they do not answer my question directly. I am just
verifying that the designation of the time scale should be
equivalent/similar  to the stset command from Paul Lambert's
introduction to stset command for using age as the time scale:

stset dateexit, failure(death == 1) id(id) origin(dateofbirth)
enter(studyentrytime)  scale(365.25)

I suspect so, since t for analysis: (dateexit-origin)/365.25 should be
similar to age at enrollment + (timeonstudy/365.25).

Forgive the elementary question. I would appreciate your feedback on
any unforeseeable issues with the method I am using, given the
constraints of the data.

Thanks

Andriene
PhD Student
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/