Cox works on the rank order of the survival times, not the survival times themselves. Using begage and endage won't change this rank order.
If you have the date of birth (dob), the date of entry(denter) into the study and the date of exit (dexit, either death or censoring date), you could try something like this:
stset dexit, enter(denter) scale(365.25) failure(died) id(idnum)
stsplit agegrp, at(0(5)100) after(dob)
The scale option will calculate the follow-up time in years and the id option identifies each person and is needed for the stplit.
I think this code is OK; I can't check it at the moment.
This will split the follow-up time by age in 5 year intervals, creating a separate record for each age-specific period at risk and a new variable, agegrp, that identifies age at risk.
______________________________________________
Kieran McCaul MPH PhD
WA Centre for Health & Ageing (M573)
University of Western Australia
Level 6, Ainslie House
48 Murray St
Perth 6000
Phone: (08) 9224-2140
Fax: (08) 9224 8009
email: [email protected]
http://myprofile.cos.com/mccaul
_______________________________________________
The fact that no one understands you doesn't make you an artist.
From: [email protected] [mailto:[email protected]] On Behalf Of uri goldbourt
Sent: Monday, 8 September 2008 5:51 AM
To: [email protected]
Subject: RE: st: AGE at event occurrence as the time variable in mortality analysisrather than time between entry into follow up an event occurrence
Thank you. In the meantime I tried stset endage, enter(time begage) failure(died==1)and obtained almost identical results. I am trying to figure out why.
UG
-----------------------------------
-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Svend Juul
Sent: Sunday, September 07, 2008 10:21 AM
To: [email protected]
Subject: Re: st: AGE at event occurrence as the time variable in mortality analysis rather than time between entry into follow up an event occurrence
Uri wrote:
�
In an epidemiological study where we follow up very old individuals the
importance of time till event (say, mortality) pales in comparison to the
actual age of the individual when the endpoint event occurred. Using stcox
(for example) we would seek to estimate the event hazard ratios associated
with several predictors. In a cohort which spans over a meaningful age range
the age when the event occurs rather than how long it took to occur since
that person entered follow up which will matter. So instead of the later the
better we would like to say the older the better.
�
What is the accurate way to stset the data so that age at event(assuming of
course it is included for each event) become the time variable?
�
===============================================================
�
You can make age the time axis:
�
��� stset endage , enter(time begage) failure(died==1) id(id)
�
where -begage- is the age at entry and -endage- the age at death
or censoring.
�
hth
Svend
�
________________________________________________________
�
Svend Juul
Institut for Folkesundhed, Afdeling for Epidemiologi
(Institute of Public Health, Department of Epidemiology)
Vennelyst Boulevard 6
DK-8000 Aarhus C,� Denmark
Phone, work:�� +45 8942 6090
Phone, home:�� +45 8693 7796
Fax:���������� +45 8613 1580
E-mail:������� [email protected]
_________________________________________________________
*
*�� 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/
*
* 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/
