R help - May 2011 - logistic regression lrm() output

Hi, I am trying to run a simple logistic regression using lrm() to calculate a 
odds ratio. I found a confusing output when I use summary() on the fit object 
which gave some OR that is totally different from simply taking 
exp(coefficient), see below:
>
dat<-read.table("dat.txt",sep='\t',header=T,row.names=NULL)
> d<-datadist(dat)
> options(datadist='d')
> library(rms)
> (fit<-lrm(response~x,data=dat,x=T,y=T))
Logistic Regression Model
lrm(formula = response ~ x, data = dat, x = T, y = T)

                      Model Likelihood     Discrimination    Rank Discrim.    
                         Ratio Test            Indexes          Indexes       

Obs           150    LR chi2      17.11    R2       0.191    C       0.763    
 0            128    d.f.             1    g        1.209    Dxy     0.526    
 1             22    Pr(> chi2) <0.0001    gr       3.350    gamma   0.528
max |deriv| 1e-11                          gp       0.129    tau-a   0.132    
                                           Brier    0.111                     

          Coef    S.E.   Wald Z Pr(>|Z|)
Intercept -5.0059 0.9813 -5.10  <0.0001 
x          0.5647 0.1525  3.70  0.0002 

As you can see, the odds ratio for x is exp(0.5647)=1.75892.

But if I run the following using summary():
> summary(fit)
             Effects              Response : response 

 Factor      Low    High   Diff.  Effect S.E. Lower 0.95 Upper 0.95
 x           3.9003 6.2314 2.3311 1.32   0.36 0.62       2.01      
  Odds Ratio 3.9003 6.2314 2.3311 3.73     NA 1.86       7.49

What are these output? none of the numbers is the odds ratio (1.75892) that I 
calculated by using exp().

Can any explain?

Thanks

John
	[[alternative HTML version deleted]]

Why is a one unit change in x an interesting range for the purpose of
estimating an odds ratio?

The default in summary() is the inter-quartile-range odds ratio as clearly
stated in the rms documentation.
Frank

array chip wrote:
> 
> Hi, I am trying to run a simple logistic regression using lrm() to
> calculate a 
> odds ratio. I found a confusing output when I use summary() on the fit
> object 
> which gave some OR that is totally different from simply taking 
> exp(coefficient), see below:
> 
>>
dat<-read.table("dat.txt",sep='\t',header=T,row.names=NULL)
> 
>> d<-datadist(dat)
>> options(datadist='d')
>> library(rms)
>> (fit<-lrm(response~x,data=dat,x=T,y=T))
> 
> Logistic Regression Model
> lrm(formula = response ~ x, data = dat, x = T, y = T)
> 
>                       Model Likelihood     Discrimination    Rank Discrim.
>                          Ratio Test            Indexes          Indexes
> 
> Obs           150    LR chi2      17.11    R2       0.191    C       0.763
>  0            128    d.f.             1    g        1.209    Dxy     0.526
>  1             22    Pr(> chi2) <0.0001    gr       3.350    gamma  
0.528
> max |deriv| 1e-11                          gp       0.129    tau-a   0.132
>                                            Brier    0.111
> 
>           Coef    S.E.   Wald Z Pr(>|Z|)
> Intercept -5.0059 0.9813 -5.10  <0.0001 
> x          0.5647 0.1525  3.70  0.0002 
> 
> As you can see, the odds ratio for x is exp(0.5647)=1.75892.
> 
> But if I run the following using summary():
> 
>> summary(fit)
>              Effects              Response : response 
> 
>  Factor      Low    High   Diff.  Effect S.E. Lower 0.95 Upper 0.95
>  x           3.9003 6.2314 2.3311 1.32   0.36 0.62       2.01      
>   Odds Ratio 3.9003 6.2314 2.3311 3.73     NA 1.86       7.49
> 
> What are these output? none of the numbers is the odds ratio (1.75892)
> that I 
> calculated by using exp().
> 
> Can any explain?
> 
> Thanks
> 
> John
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
> 

-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context:
http://r.789695.n4.nabble.com/logistic-regression-lrm-output-tp3533223p3533278.html
Sent from the R help mailing list archive at Nabble.com.