Re: st: RE: Interpretation of tvc in stcox

[Enzo Coviello <enzo.coviello@alice.it>]

Hi,

> Thank you for your interest.  To answer your 
> questions, yes, the texp(_t) is how the variables
> change with time. It is the default so did not 
> need to be included in the stcox statement.
>
> My data is set up so that there are multiple 
> records per individual, with different rows for each
> visit.  Using the 'id' option in the stset 
> statement, Stata will know how to identify records
> within an individual, ie. 360 subjects, 1577 records:
> Behavioral variables were measured at each visit 
> (time varying), and were specified with the tvc
> option.
I believe that a distinction should be done 
between time-varying covariates and time-varying effects.
The purpose of the options tvc() and texp() is 
addressed to model time-varying effects.

In other words you can have time-varying 
covariates and model hazard ratios that are fixed in the time.
Using tvc() and texp() hazard ratios varying 
along the time can be modeled. This can be useful 
if proportional hazards assumption does not hold.
Time-varying effects do not require time-varying covariates in the dataset.

Best wishes

Enzo


Enzo Coviello
Unit� di Epidemiologia e Statistica ASL BARI
Piazza V.  Emanuele 14
70054 Giovinazzo (Bari)
Italy
tel./fax +39 080 3357867
mobile +39 347 5016 016
home +39 0883 695055


*
*   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/