Re: semantic web for EHRs

On Wed, 2010-03-10 at 15:38 -0800, conor dowling wrote:
> Is anyone on the group involved or know about adding a Semantic
> Endpoint to an EHR system? I'm working on a SPARQL "profile" called
> FMQL (see: http://www.caregraf.org/semanticvista#fmql ) for VA's
> VistA, the most widely deployed EMR in the U.S. (and open-sourced to
> boot). You can see/code-to the current "linked data" support for VistA
> at http://vista.caregraf.org .

We certainly use SPARQL endpoints for patient records in the Cleveland
Clinic's SemanticDB project:
http://www.w3.org/2001/sw/sweo/public/UseCases/ClevelandClinic/ 
However, at present the data is only heart and vascular related -- not a
patient's entire medical record -- and it is only used for research and
quality reporting -- not for real time patient care.
> 
> Luckily VistA's built around graphs - if you look at a patient like
> the unfortunate Christopher Jones
> ( http://vista.caregraf.org/rambler/2/9 ), you see that his "record"
> is a patient node referenced by nodes for vitals, orders,
> medications ... The VA has over ten million patients in graphs like
> this. 

Kudos!  That's a big step in the right direction.

> Unluckily, their graph store - called FileMan - has no remote query
> language. Which is where SPARQL, OWL et al come into play.
> 
> Is such EHR system "exposure" common outside the U.S.? (There's little
> here from what I can see - the interoperability push is around SOAP).
> Any work in Britain, Germany etc.? Or Finland (where they use a
> similar system to VistA)?

In my view, SOAP is the wrong direction.  It is just adds complexity and
contributes to "babelization":
http://www.w3.org/2003/Talks/0717-semweb-dbooth/slide10-0.html 

I think you're on the right track with RDF, SPARQL and RESTful
interfaces, as my talk at last year's SemTech conference explains:
http://dbooth.org/2009/stc/dbooth-SemTechConf-2009.ppt 

> 


-- 
David Booth, Ph.D.
Cleveland Clinic (contractor)

Opinions expressed herein are those of the author and do not necessarily
reflect those of Cleveland Clinic.

Received on Friday, 12 March 2010 04:12:26 UTC