[R] I have a question on nomograms.

[Marc Schwartz]

On Mar 25, 2010, at 9:51 AM, Frank E Harrell Jr wrote:

> Marc Schwartz wrote:
>> On Mar 25, 2010, at 8:54 AM, 笑啸 wrote:
>>> Dear volunteer:
>>>   I am a graduate student of medcine in china.And now,I am devoting myself to constructing a nomograms of bladder cancer.I want to do it with R-project.However, I do not know how to construct a nomograms with R-project.I want to get yours help,thank you! I wish you can tell me the operating procedure of the R-project.
>>>      And I apologize for my english,it is poor,sorry!
>>>                                                                                                                                                         truly yours
>>>                                                                                                                                                          ding ding  
>> If you are referring to a regression model based nomogram to be used for prediction, see the nomogram() function in the 'rms' CRAN package by Frank Harrell.
>> Also, if you go to rseek.org and search using the keyword 'nomogram' you will find this and some other functions that provide related functionality.
>> HTH,
>> Marc Schwartz
> 
> In addition, you will have to be expert in the type of regression modeling used to produce a reliable model that is then displayed in a nomogram.  The nomogram is the easy part; it is just a graphical display of a model.  It is not the model itself.


Picking up on the important points expressed by both Frank and Ted, I would point you to two excellent reference texts:

First, Frank's own:

Regression Modeling Strategies 
Springer, 2001
http://www.amazon.com/Regression-Modeling-Strategies-Frank-Harrell/dp/0387952322/
http://biostat.mc.vanderbilt.edu/twiki/bin/view/Main/RmS


and second:

Clinical Prediction Models: A Practical Approach to Development, Validation, and Updating
Ewout W. Steyerberg
Springer, 2008
http://www.amazon.com/Clinical-Prediction-Models-Development-Validation/dp/038777243X/
http://www.clinicalpredictionmodels.org/


HTH,

Marc