|Date: ||Tue, 27 Feb 2007 21:58:47 -0800|
|Reply-To: ||David L Cassell <davidlcassell@MSN.COM>|
|Sender: ||"SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>|
|From: ||David L Cassell <davidlcassell@MSN.COM>|
|Subject: ||Re: Baseline Calculation in Proc Phreg|
|Content-Type: ||text/plain; format=flowed|
ppatrawala@INDUCTIS.COM wrote back:
>The data does have censored observations, and it is a typical Survival
>Even if, for arguments sake, my problem is not that of Survival Analysis, i
>still am interested in finding out the procedure to calculate the Baseline
>Function calculation for time dependent covariates in Proc Phreg.
>Perhaps you could post a solution to that effect. It would be greatly
Okay, I still don't see where you want to go, because:
 the point of PROC PHREG is that the proportional hazards model is
semi-parametric, with an *unspecified* baseline hazard function;
 proportional hazards regression doesn't work right unless you have
the proportional hazard assumption being met, and I don't see that
you have that;
 you may have some discrete regressors and that would lead you
from PHREG to the experimental PROC TPHREG.
If you really want survival analysis, instead of time series modeling
or queueing theory modeling, I think you should start with PROC
LIFETEST, and get a look at the baselines for each of your categories,
along with an evaluation of the proportional hazards assumption.
David L. Cassell
3115 NW Norwood Pl.
Corvallis OR 97330
Win a Zune™—make MSN® your homepage for your chance to win!