W3C home > Mailing lists > Public > public-grddl-wg@w3.org > April 2007

Re: "Killer" use-case for GRDDL Primer

From: Harry Halpin <hhalpin@ibiblio.org>
Date: Fri, 06 Apr 2007 21:51:26 -0400
Message-ID: <4616F91E.3070607@ibiblio.org>
To: John Madden <john.madden@duke.edu>
Cc: public-grddl-wg <public-grddl-wg@w3.org>

John,
    There was some concern from Chime that the sort of "saving patient
life" use-case might bit a bit unwise from a medical standpoint. Chime
suggested, can we discover if we have an "X-ray indication" attached to
the HL7 information model:

http://lists.w3.org/Archives/Public/public-grddl-wg/2007Mar/0126.html

    However, I don't know if that's compelling enough as an example.
Anyways, would like to address this XML/OWL example in Primer, and keep
it short, simple, and catchy.

                -harry


John Madden wrote:
>
> Harry,
>
> Yes, I am putting suggested revisions to this use case together now
> and will post this weekend together with comments.
>
> John
>
>
>
> On Apr 4, 2007, at 1:48 AM, Harry Halpin wrote:
>
>> John (cc'ing the GRDDL WG list),
>>
>> As per our last conversation, I was hoping if you can help us with the
>> "killer" GRDDL use-case you were talking about earlier. In particular,
>> the use-case I believe has a patient going to the doctor twice, and so
>> filling out two different HL7 CDA documents, both in XML. On his first
>> visit, he records that he is allergic to a particular family of drugs.
>> On the second visit, he forgets that he is allergic to this family of
>> drugs, and so his second HL7 CDA document does not record this. A doctor
>> prescribes him a drug on the second visit, unknowingly a drug that is a
>> subclass of the family of drugs the patient is allergic to by virtue  of
>> its properties.
>>
>> Yet the hospital can automatically catch this error and save the
>> patient's life by converting both HL7 documents via GRDDL to RDF, and
>> merging them. Therefore the old data about the patient being allergic is
>> not lost, but discovered.
>>
>> Furthermore,  because the family of drugs is kept in an ontology, some
>> sort of simple OWL entailment can show the drug the doctor prescribed is
>> a subclass of the drug the patient said they were allergic to.
>>
>> John - this piece would be a rewrite of piece of the Primer Chime
>> already wrote. I'm happy to write the prose if you can produce another
>> HL7 document that describes this sort of use-case, and feel free to
>> modify anything , including data files, (all files are linked from the
>> primer):
>>
>> http://www.w3.org/2001/sw/grddl-wg/doc29/primer.html#hl7
>>
>>
>>
>> --        -harry
>>
>> Harry Halpin,  University of Edinburgh
>> http://www.ibiblio.org/hhalpin 6B522426
>>
>
>


-- 
		-harry

Harry Halpin,  University of Edinburgh 
http://www.ibiblio.org/hhalpin 6B522426
Received on Saturday, 7 April 2007 01:51:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:11:48 GMT