Hi Jennifer,
I wanted to ask:
You have specified which variables are time varying but havent specified how they change.
ie. Is the texp(_t) bit optional?
You also mention that you split your data into multiple record.s Do you have variables measured at these different time points.
How does stata know when to apply the updated covariate which has changed over time?
Dianna
Dr Dianna Magliano, BAppSci(Hons), MPH, PhD
Senior Epidemiologist
International Diabetes Institute
250 Kooyong Road
Caulfield Victoria 3162
Ph: 03 9258 5931
0425 706 637
Dianna Magliano BAppSci(Hon) MPH PhD | Epidemiologist
International Diabetes Institute | 250 Kooyong Road | Caulfield | VIC 3162. Australia.
p +61 (0)3 9258 5931 f +61 (0)3 9258 5090 e [email protected] | www.diabetes.com.au
The International Diabetes Institute and the Baker Heart Research Institute are merging during 2008
This message and any attachments are confidential and for the addressee only. If you have received it in error, please delete it immediately and notify the sender. Any views expressed in this email and any files transmitted with it are those of the individual sender, except where the sender specifically states them to the views of the International Diabetes Institute. The International Diabetes Institute does not represent or warrant that the attached files are free from computer viruses or other defects. The user assumes all responsibility for any loss or damage resulting directly or indirectly from the use of the attached files.
-----Original Message-----
From: [email protected] on behalf of Jennifer Evans
Sent: Thu 08/05/2008 7:46 AM
To: [email protected]
Subject: st: Interpretation of tvc in stcox
Dear statalist,
I have a dataset with multiple records per person with time varying covariates. I am running
survival analysis with the tvc option in stcox and have found 2 variables with significant time
interactions. My questions are:
(1) Am I setting up the stcox statement correctly (see code below)
(2) Regarding the interpretation- Is it correct to say that the variable var1 has a hazard of 0.35
with the relative hazard then decreasing by a factor of 1.00 with each unit of _t?
Any comments or suggestions would be appreciated!
Best regards,
Jennifer Evans
. xi: stcox yrsinj_b age_b gender_b var1 var2 var3 var4, nolog noshow tvc(var1 var3)
Cox regression -- Breslow method for ties
No. of subjects = 360 Number of obs = 1577
No. of failures = 104
Time at risk = 226907
LR chi2(16) = 117.25
Log likelihood = -467.29121 Prob > chi2 = 0.0000
------------------------------------------------------------------------------
_t | Haz. Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
rh |
yrsinj_b | 1.033031 .029971 1.12 0.263 .9759283 1.093476
age_b | .9916956 .0319112 -0.26 0.796 .9310822 1.056255
gender_b | 1.114194 .2912036 0.41 0.679 .6675631 1.859641
var1 | .3489511 .1266303 -2.90 0.004 .1713457 .7106501
var2 | .3903011 .1237031 -2.97 0.003 .2097093 .7264101
var3 | .2174367 .0915801 -3.62 0.000 .0952409 .496412
var4 | 1.562432 .3349837 2.08 0.037 1.026371 2.378471
-------------+----------------------------------------------------------------
t |
var1 | 1.001592 .000593 2.69 0.007 1.00043 1.002755
var3 | 1.001354 .0005982 2.26 0.024 1.000182 1.002527
------------------------------------------------------------------------------
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
<<winmail.dat>>
