[R] Hmisc improveProb() and PredictABEL reclassification () function and continuous NRI

Frank Harrell f.harrell at vanderbilt.edu
Mon Jul 30 14:56:34 CEST 2012

If you type improveProb at the command line you'll see all the code, which is
quite simple.  I'm glad you are not attempting to classify risk as this is
almost always futile.


Svingen, Gard Frodahl wrote
> Dear Sirs.
> I am working with the  R packages Hmisc and PredictABEL to make NRI
> estimates from my Cox models with and without a specific biomarker.
> According to Pencina et al (Statistics in Medicine 2010, DOI:
> 0.1002/sim.4085 ), a continuous/non-categorical NRI (NRI>0) is to be used
> when there are no obvious reason to categorize risk, such as the risk of
> future cardiovascular events in patients with established cardiovascular
> disease.
> My question is therefore: Which value(s) are to be used in the calculation
> of continuous NRI from the output in Hmisc or in PredictABEL? Does
> continuous NRI equal total NRI in the output?
> Yours sincerely
> Gard Frodahl T. Svingen
> PhD student
> the University of Bergen
> Bergen, Norway
> 	[[alternative HTML version deleted]]
> ______________________________________________
> R-help@ 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/Hmisc-improveProb-and-PredictABEL-reclassification-function-and-continuous-NRI-tp4630259p4638347.html
Sent from the R help mailing list archive at Nabble.com.

More information about the R-help mailing list