- From: John Madden <john.madden@duke.edu>
- Date: Tue, 14 Feb 2006 19:16:58 -0500
- To: Dan Corwin <dan@lexikos.com>, HCLS <public-semweb-lifesci@w3.org>
- Cc: jluciano@predmed.com
Joanne & Dan, Thanks for the words of encouragement and interest. I feel stupid that my automatic email rules engine misfiled both your responses and I just noticed them today! Anyway, the tie-in to HL7 is slightly secondary until I get the SNOMED part under control. But I agree with you, Dan, that figuring out how to semantically interoperate with HL7 RIM classes in the context of constrained use cases would be a big payoff. A funny thing about the HL7 RIM is that the abstract RIM model is done in a dialect of UML, and thus the progenitor Act class is a full- fledged object. But HL7's current development methodology, for historical reasons, assumes that the implementation language is XML schema (this has to do with the legacy HL7 messaging model), so much of the three-dimensional object quality of the Act class (and other RIM classes) gets "flattened" in the course of implementation. I think Act is a good class and is well done in the abstract HL7-RIM model, but it would be much more useful in the semantic web context if it were implemented as a Java class, or a web service class, instead of as a schema fragment. You could actually imagine some clever interoperation with an NLP service over the web if this were the case. John On Feb 10, 2006, at 1:42 PM, Dan Corwin wrote: > Hi John - > > Thanks, Joanne, for thinking of me. I do follow HCLSIG > emails (at your kind urging), but not as a paid up member. > I merely lurk, seeking new sorts of fascinating ideas. > > This thread in particular really caught my eye due to my > recent SBIR proposal on using HL7 to guide NLP dialogs: > > http://www.lexikos.com/charts/usaftop.htm > > I just came off a contract where I got exposed to HL7. It > impressed me especially with its professional Acts, a class > of situations applicable far beyond health care realms. > > As a basis for "logical forms" (that NLP semantics are often > expressed in), Acts seem quite exciting. So does the notion > of adapting MODELER to work with the UMLS. Hence the above. > > Please keep me in mind if I can help out in any way. > > Dan Corwin > > > > Joanne Luciano wrote: >> I figured you'd had read them, but just in case :-) >> Your right, I send it for the benefit of the community. >> I forgot to mention that I'm interested in the project and >> also will copy Dan Corwin (dan@lexikos.com) who's been looking at >> HL7. He >> may not be on the public semweb list - so this will >> serve as his <HINT> <HINT>. I believe Dan is on the job market so >> maybe >> something will come of this, who knows. >> I'm envious you're getting to go to Stanford for the Protégé >> workshop. I always benefit from those and I know I'm far from >> a Protégé power user. >> Joanne >>> -----Original Message----- >>> From: John Madden [mailto:madden.jf@gmail.com] On Behalf Of John >>> Madden >>> Sent: Friday, February 10, 2006 9:05 AM >>> To: jluciano@predmed.com >>> Subject: Re: HL7 and patient records in RDF/OWL? >>> >>> Joanne, >>> >>> Already done. I agree these are great resources, and should be >>> known to all who work in this area. >>> >>> John >>> On Feb 10, 2006, at 8:32 AM, Joanne Luciano wrote: >>> >>>> All, >>>> >>>> Before attempting to OWLize anything first read this: >>>> >>>> Matthew Horridge's Protégé OWL Tutorial: >>>> http://www.co-ode.org/resources/tutorials/ProtegeOWLTutorial.pdf >>>> >>>> Then this >>>> >>>> Alan Rector et al: >>>> OWL Pizzas: Practical Experience of Teaching OWL-DL: >>>> Common Errors & Common Patterns >>>> http://www.co-ode.org/resources/papers/ekaw2004.pdf >>>> >>>> I speak from experience. You won't be sorry :-) >>>> >>>> Joanne >>>> >>>> Joanne S. Luciano, PhD >>>> Lecturer in Genetics, Genetics Department, Harvard Medical >>>> School Visiting Research Fellow, Computer Science, University of >>>> Manchester >>>> US: +1 617.440.4364 >>>> UK: +44 (0) 161 408 4096 >>>> Email: jluciano@genetics.med.harvard.edu Home Page: http:// >>>> arep.med.harvard.edu/~jluciano/ >>>> >>>> >>>> >>>> >>>> ________________________________ >>>> >>>> From: public-semweb-lifesci-request@w3.org >>>> [mailto:public-semweb-lifesci-request@w3.org] On Behalf Of >>> John Madden >>>> Sent: Wednesday, February 08, 2006 11:37 PM >>>> To: public-semweb-lifesci@w3.org >>>> Subject: Re: HL7 and patient records in RDF/OWL? >>>> >>>> >>>> Dan, >>>> >>>> This is also a big interest of mine. >>>> >>>> I'm co-chairing the HL7 Anatomic Pathology SIG (and I'm in HCLS >>>> courtesy SNOMED). With SNOMED's support, I'm embarking on >>> OWL-izing a >>>> subset of pathology content to (a) support some circumscribed >>>> inferencing on pathology report content (b) demonstrate >>> SKOS mappings >>>> to legacy SNOMED and >>>> (c) investigate OWL embedding in pervasive (non-xhtml) medical >>>> document formats (HL7-CDA, but maybe also OpenDocument, DocBook). >>>> >>>> Many issues. I'd love to discuss, whenever. >>>> >>>> John >>>> >>>> >>>> >>>> >>>> On Feb 8, 2006, at 10:37 AM, helen.chen@agfa.com wrote: >>>> >>>> >>>> >>>> Hi, Dan >>>> >>>> It is an area of my interest too. >>>> >>>> You can check out the previous work on a RIM owl model: >>>> http://www.ics.mq.edu.au/~borgun/RIM-HL7All.zip >>>> >>>> If you use N3 to write your ontology, you can >>> also get a N3 version >>>> here http://www.agfa.com/w3c/hl7/rim/rim.n3 >>>> >>>> We have a graduate student currently looking >>> into this model and see >>>> how it can be used in real EPR . >>>> >>>> Do you have any sepecific task/charter in mind? >>>> >>>> Regards. >>>> >>>> Helen >>>> >>>> http://www.agfa.com/w3c/hchen >>>> >>>> >>>> >>>> >>>> >>>> >>>> Dan Brickley <danbri@danbri.org> >>>> Sent by: public-semweb-lifesci-request@w3.org >>>> >>>> 02/08/2006 10:09 AM >>>> >>>> To >>>> public-semweb-lifesci@w3.org >>>> cc >>>> >>>> Subject >>>> HL7 and patient records in RDF/OWL? >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> Hi all >>>> >>>> I'm looking into use of RDF for healthcare >>> apps, specifically >>>> patient records. I'd be interested to hear this >>> group's sense of the >>>> current state of the art (eg. HL7-in-RDF etc). >>> Was this discussed >>>> much at the recent F2F meeting? Anyone >>> interested in collaboration, >>>> testing and interop? >>>> >>>> cheers, >>>> >>>> Dan >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>>> >>> >>> >
Received on Wednesday, 15 February 2006 00:20:11 UTC