Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Roctab for classvar in descending order


From   alvine.bissery@chu-lyon.fr
To   statalist@hsphsun2.harvard.edu
Subject   st: Roctab for classvar in descending order
Date   Thu, 06 Nov 2008 14:45:35 +0100

Dear all, 

I need to study specificity and sensitivity of measures of lymphocytes.
The gold standard is measures of CD4

A subject is diseased if CD4<200 and non diseased if CD4>=200
We want to study specifity and sensitivity for different threshold of
lymphocytes
Generally, if lymphocytes is <1200, subject is diseased.

Using the command

"roctab CD4_bin lymphocytes"

With CD4_bin=1 if CD4<200  and CD4_bin=0 if CD4>=200 

I obtain graph and table for sensitivity/specificity for threshold of
lymphocytes in ascending order.
Or I'm interested in descending order: a patient is diseased if lymphocytes
is <1200, and not >=1200

If I use 

gen inv=-1*lymphocytes
roctab CD4_bin inv

It's OK, but the table is not useful because of negative values of
lymphocytes

Thanks for help


A.Bissery

*****************************************************
Alvine BISSERY
Ingénieur statisticien 
Service de Biostatistique des Hospices Civils de Lyon 
162 Avenue Lacassagne 
69424 Lyon Cedex 03 - France 
tél: +33 4 72 11 57 55 
fax: +33 4 72 11 51 41 


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index