s-news
[Top] [All Lists]

[S] coxph() output

To: s-news@wubios.wustl.edu
Subject: [S] coxph() output
From: James Song <jsong@bcm.tmc.edu>
Date: Tue, 19 Sep 2000 17:33:42 -0500
In-reply-to: <Pine.GSO.4.05.10009191740080.5008-100000@welles.harvard.ed u>
Sender: owner-s-news@wubios.wustl.edu
Dear S users,

I have a question about the x vector in coxph. The input x variable has
value 1 and 0, representing treatment group and control group. I have 50
subjects in treatment group and 100 subjects in control group. Using a
model such as

coxout<-coxph(Surv(time,status)~x,mydata)
coxout$linear.pred     returns a vector of 0.6021568 (each for 50 subjects
in the treatment group) and -0.3015784 (each for 100 subjects in the
control group), instead of a vector of 50 "beta estimator"s and 100 "0"s.

coxout$linear.pred is not equal to exp(beta*X). I wonder how splus handles
X variable. Since cumulative baseline hazard Ho is estimated as sum(# of
death/sum(exp(beta*X)))       (1),    the baseline hazard is also different
from the Ho calculated by  plugging the final beta estimate, X and # of
death into (1).

Thanks for your help!

James
 



-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu.  To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message:  unsubscribe s-news

<Prev in Thread] Current Thread [Next in Thread>
  • [S] coxph() output, James Song <=