[R] Confidence interval for Tau-a or c-index to compare logistic lrm (binary) models with each other.

[Frank E Harrell Jr]

Jan Verbesselt wrote:
> Dear R list,
> 
> How can confidence interval be derived for e.g. the Tau-a coefficient or the
> c index (area under ROC curve) such that I can compare the fitted lrm
> (logistic) models with each other. Is this possible?
> 
> The aim is to conclude that one model is significantly better than other
> model (a, b or c). 
> 
> y~a (continu)+ d(catergoric)
> y~b (continu)+ d(catergoric)
> y~c (continu)+ d(catergoric)
> 
> I will look at residual deviance, the AIC, c-index en Tau-a to compare
> different logistic models (lrm Design package).  All extra advice is mostly
> welcome!
> 
> Regards,
> Jan

You can only do this if you have an independent test dataset that was 
never used in model development or coefficient estimation.  Given that, 
look at the rcorrp.cens function in Hmisc.

Frank Harrell

> 
> _______________________________________________________________________
> ir. Jan Verbesselt 
> Research Associate 
> Lab of Geomatics Engineering K.U. Leuven
> Vital Decosterstraat 102. B-3000 Leuven Belgium 
> Tel: +32-16-329750   Fax: +32-16-329760
> http://gloveg.kuleuven.ac.be/
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
> 


-- 
Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University