Re: st: Time-varying covariates

[Steve Samuels <sjsamuels@gmail.com>]

You asked the same question in January, and after some back-and-forth
 I answered it at: 
http://www.stata.com/statalist/archive/2012-01/msg00997.html

Should waiting time be a time-dependent covariate? No. It should
be the fundamental time variable for the survival analysis.

Steve
sjsamuels@gmail.coms


 

On May 10, 2012, at 2:30 PM, Kyleigh Schraeder wrote:

Hi,

I am running a cox regression model in survival analysis and I have a
question about time-varying covariates.  I have a covariate that
describes subjects total waiting time (in days) by the end of their
observation period.  At the beginning of the study (t=0), all subjects
were put on a wait-list.  I have two groups of subjects: 1) still
waiting by the end of the study and 2) waited and received help (i.e.,
stopped waiting during study).  I am interested in whether subjects'
waiting time impacts probability of the event of interest occuring.

I am wondering whether I should treat this waiting time covariate as
time-varying or not.  If I treat this as a time-varying covariate, I'm
looking for the appropriate syntax.  I tried the following syntax from
the example in the Stata manual under 'Cox regression with discrete
time-varying covariates' but I'm not sure if it's appropriate for my
question.
Target event = Event_A
EventTime_A = Study time (time to event)
Waiting_Status = grouping variable

stcox   Wait_Time_A, tvc(Wait_Time_A) texp(.1*(_t))

Example of my datasetup in Stata:
id_1	Waiting_Status_A	Wait_Time_A	Event_A	EventTime_A
16	2	120	0	176
29	2	180	0	528
31	2	267	1	121
34	1	113	0	113
68	1	281	0	281

Thank you for your help!
K
*
*   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/