cfNRI

[Natalie Scherrer]

Hello
I would like to calculate the category free NRI with stata. I tried a lot of programs already but didn t manage to calculate id. Could somebody help me?

Thank you very much for your reply.

Kind regards,

Natalie Scherrer

[Friedrich Huebler]

Which programs did you try? Which problems exactly did you encounter?

Please also take a few minutes to review the FAQ for advice on how to ask questions on Statalist.

[Rich Goldstein]

as with Friedrich, you don't really supply enough information; here, however, is instructions for obtaining incrisk, a program I have used for this purpose

.net from http://research.fhcrc.org/content/dam/stripe/diagnostic-biomarkers-statistical-center/files/stata/ 

You will be prompted with a list of packages you could net describe. Click on risk_prediction. This will currently install two commands, incrisk and predcurve. A program description is available by using the help facility, e.g., help predcurve opens a window that explains the syntax and options for the predcurve command. PDFs of the help files can be obtained here for predcurve and incrisk.

[Natalie Scherrer]

Thank you Friedrich and Rich.

I am using Stata 13.1 for mac. I want to calculate the category free net reclaffication of a new marker of a dataset with 81 variabels and approximately a million of observations. I tried the following:

[idi from http://personalpages.manchester.ac.uk/staff/mark.lunt]
[net from http://www.ucr.uu.se/en/images/stories/downloads]

The first gives me cNR not cfNR. From the second I received some results but I am not sure if its the cfNR or the cNR. I search now a program which I can install in Stata to receive cfNR.

Thank you for your support.

[Rich Goldstein]

well, I gave you instructions for a different package - did you download and try that?

[Natalie Scherrer]

Yes, it worked. Thank you very much Rich. I just have a questions about the p-values. What is the p-value of the NRI and IDI in this example in my screenshot? Is it p-value=0.65 for both the NRI and IDI?


Thank you for your reply

[Rich Goldstein]

the p-value shown (but not easy for me at least to read) is ONLY for the H-L got test; no p-value is shown for either NRI or IDI; the CI's for each are shown

[Natalie Scherrer]

Thank you Rich. You helped me a lot.

[Natalie Scherrer]

Do you know how to calculate the p-value of these NRI and IDI results? Thank you.

[Rich Goldstein]

I have never looked at that; you have the CI so you can easily decide whether the result is statistically significant (if the null is within the CI limits, then, no, it's not; if the null is outside the CI limits then it is statistically significant at the 0.05 level (because these are 95% CIs); if you want an exact p-value, someone else might be able to help

[Natalie Scherrer]

Thank you Rich. Do you know how to calculate the NRI and IDI based on a cox model with this program?

[Rich Goldstein]

no, I don't; further, I don't see what NRI or IDI would mean for a survival model - they estimate change in risk at a particular time - this is not what survival models estimate

[Steve Samuels]

The FAQ that Rich asked you to read ask that you not use photographs or screen shots. Please in the future show all code and results between opening and closing CODE delimiters, which are (with spaces and quotes omitted) " [C O D E]" and "[/C O D E]".

[John Moran]

I came to this thread a little late.
If you do have survival data, and for instance you are looking at the effect of a biomarker (over time, not cross-sectional), then on way to assess this is with Hajime Uno's "survIDINR" R package (see also Statistics in Medicine: 2013, 32(14): 2340-42).
If R is not useful for you, then an alternative measure, which some have strongly recommended, is "Decision curve analysis" by Andrew Vickers, which is available for data in survival format ("stdca"; see http://www.decisioncurveanalysis.org)

john moran

[Natalie Scherrer]

Thank you John