|
|  |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: bland altman
I recently had some similar questions, maybe more related to methods
comparison studies, and how to do them well.
In addition to Nick's very useful program _concord_, which does Bland-
Altman analysis (but leaves out some confidence intervals), there is
a kind of orthogonal regression available by one of Stata's employees:
findit deming
will findit for you.
Note the data set used in the '86 Bland and Altman paper is available
by Stata, e.g.,
use http://www.stata-press.com/data/mlmus/pefr.dta, clear
concord wp1 wm1, loa
You might find this terse review useful:
http://www.biopharminternational.com/biopharm/data/articlestandard/
biopharm/032002/7276/article.pdf
If you found the analysis in the 1986 B & A paper interesting, that
same data <http://www.stata-press.com/data/mlmus/pefr.dta> can be
analyzed by a three level random effects model. The analysis is
presented on pages 215-229 of Rabe-Hesketh and Skrondal, "Multilevel
and longitudinal modeling using Stata", <http://www.stata.com/
bookstore/mlmus.html>.
A book I borrowed, "Statisical Evaluation of Measurement Errors:
Design and Analysis of Reliability Studies", 2004, by Graham Dunn,
also goes into detail about and promotes measurement error models,
including more advanced treatments by latent variable models on
replicated data with greater than two methods, using GLLAMM and
Mplus. The author is a Stata user, apparently.
-Dave
In addition,
search concord
findit pairplot
Nick
[email protected]
Scott Merryman
> -findit bland- points to several resources.
Decebal Gabriel Latcu
> > 1. I have two number series (same parameter measured by two
> different
> > methods). I'd like to do a Bland-Altman analysis. How to do
> it in Stata?
> > Please provide practical answer for a non-programmer!.
> >
> > 2. Is there a mean of concluing on the precision difference
> between the
> > two
> > methods type variance analyse? (the quantitative variables
> are continuous,
> > age-type : ex 1 year to 60 years).
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
