On 9/22/2010 10:26 AM, Djulbegovic, Benjamin wrote:
> Thanks for using our SRs- it is always good that someone can find some
> use of the research one does. However, I am not an advocate of data
> mining of medical records- I am only pointing to the exponential growth
> of this type of research. Surely, we ought to understand its strenghts
> and weaknesses because it will only continue to be performed at ever
> accelerating pace. Perhaps, EBMers have already said what needs to be
> said (as implied in your message and summarized by Roy and Steve in
> their seperate messages- that this research should not be trusted and
> perhaps not even done?)
It's always hard to write with the proper sense of nuance. I did not
intend to imply "never trust" or "never do". Rather, I think I was
saying "be careful" especially if there was no input from a professional
statistician. I like the idea of a Users Guide for Data Mining Findings
and would love to help draft such a document for publication.
--
Steve Simon, Standard Disclaimer
Sign up for The Monthly Mean, the newsletter that
dares to call itself "average" at www.pmean.com/news
|