Lots of interesting comments being filed at the journal about this article. Here is one from a librarian (and hurray for her!).

Value of Fulltext Searching
Joy C. Kennedy
BMJ, 11 Nov 2006
http://bmj.com/cgi/eletters/bmj.39003.640567.AEv1#148956

This relates well to a front page piece in Sunday's (11/12) New York time. This article talks about Web 3.0, the "semantic Web."

Described as being in its infancy, Web 3.0 is visioned as being able to deliver "reasonable and complete answers to simple questions like: 'I'm looking for a warm place to vacation and I have a budget of $3000' ... You get the idea.

Companies mentioned in the article include IBM, Google, Radar Networks and Metaweb (where have I heard THAT name before?).

It seems like health care is a fertile field for this type of work. Would it not be slick to be able to pose a question such as "What is the best therapy for condition x with patient y with these qualities and limitations: age z, other disease conditions, etc.

Some vendors are presenting something like this, such as OVID, but I am wondering about something that would require serious computer processing. Underlying tools: Cochrane, POEMS, Guidelines, meta-analyses, but all gathered and evaluated by some underlying computer program capable of doing the work we see POEMS authors and our own residents and fellows do now to create meta-analyses and evaluations of studies. Where now Cochrane volunteers apply criteria to studies for inclusion or exclusion in bottom line recommendations, could this not be applied to an ai decision tree and automated to increase the availability of such evaluated work.

Surely this must be work already begin. Does anyone know who is doing such work??

Sure! Semantic web has been the buzz in the internet community for quite a few years now. Doing a little digging, it looks like around 2000, 2001. For who is tracking semantic web developments in health, check out this.

W3C Semantic Web Health Care and Life Sciences Interest Group: http://www.w3.org/2001/sw/hcls/

Scroll all the way down and look at the "Related Articles" section. Some very interesting directions being explored.

Also see the abstracts and slides from their November 6 2006 meeting:
http://esw.w3.org/topic/HCLS/ISWC/Workshop/Abstracts

There is an article that most of us should probably read, if we haven't already, from JMLA last Spring.

J Med Libr Assoc. 2006 April; 94(2): 198–205.
An introduction to the Semantic Web for health sciences librarians*
Ioana Robu, MSc, Library Director,1 Valentin Robu, MSc, Researcher,2 and Benoit Thirion, LIS, Conservateur-Chief Librarian3
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1435839

Here is a web search for people who want to explore more:

Google: WebSearch: "semantic web" (health OR medicine OR telehealth OR telemedicine OR medical OR drug): http://www.google.com/search?num=100&hl=en&lr=&newwindow=1&c2coff=1&q=%22semantic+web%22+%28health+OR+medicine+OR+telehealth+OR+telemedicine+OR+medical+OR+drug%29&btnG=Search

In that search you can find many reports of applications in development from a variety of institutions, covering drug discovery tools (and drug development and drug safety and drug ...), medical diagnosis, public health preparedness, and more. They seem to be thinking that the drug area may provide the "killer app" that will make the semantic web really take off.

The very interesting discussion on this article seems to be winding up. It has also been widely discussed in the popular press. The Author's reply was posted today, which is well worth reading in its entirety.

http://www.bmj.com/cgi/eletters/bmj.39003.640567.AEv1#150402

I would like to highlight this paragraph about librarians.

"We do not think that our article downplays the role of medical librarians in the practice of evidence based medicine. We have benefited enormously from workshops and help from librarians. Clinical questions arise whenever doctors meet patients [11-12]. Most of these questions can be answered using Pubmed, Cochrane Library, journals, UptodateTM et cetera. However if traditional sources of information have failed to answer these questions, then it would be unwise to ignore the web as a resource."