Date:         Sat, 5 Feb 2000 16:54:18 +0800
Reply-To:     Wang Xiaoshan <medp9194@NUS.EDU.SG>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wang Xiaoshan <medp9194@NUS.EDU.SG>
Subject:      Question on Medical Data Analysis
Content-Type: text/plain

Hi, all,

My colleague is analyzing a data from a cross-sectional survey about
asthma. Generally speaking, he should use LOGISTIC REGRESSION for
multi-factors analysis. But he isn't content with the OR results for
each factor from LOGISTIC REGRESSION. So he want to use ' COX
PROPOTIONAL MODEL'. His program as follow:

        DATA ASTHMA;
        INPUT X1-Xn asthdiag;
        time=1;
        CARDS;
        ...
        RUN;

        PROC PHREG;
        MODEL time*asthdiag(0)=X1 X2 X3.../ ties=breslow risklimits
        RUN;

The question is: Whether COX MODEL can be used like this? Thanks a lot!