Dear all,
I would appreciate your insight on this.
I am comparing two variables (X, Y) to see which one is better to
identify the "bad" actors from the "good actors".
The two variables are highly correlated, that's the criteria that
they are using are very similar, with two elements different.
I had built two separate cox models, including other covariates, one
with variable x and one with variable Y.
so now i have one hazard ratio for X+ vs X-, and one hazard ratio for
Y+ vs Y-.
Is there a way that I can program in S-plus to compare how
significantly does it improve using variable X instead of Y? such as
comparing the two hazard ratio? I can calculate the abosulte increase
of survival, but it would be awesome if I can say the difference is
"significantly" better with a p-value reported.
Your help and inputs are very appreciated
yours,
Maggie
mcheang@mac.com
|