I think LOINC has been trying to put their data into "categories". This
effort would be of enormous value for the Heterogenous (aka Healthcare)
Information Locator Service we have been working on. In it we want to
find what servers have information of a certain type on a specified
patient, without having to ask all servers out their what they have.
This is not practical if the granularity isn't fairly coarse. (Anyone
who wants to help in specifying this service is invited to help out).
Dave
John S. Gage writes:
> Thanks, Wayne, for the comment. I do not believe however that I have
> gotten across what I am trying to describe.
>
> The DTD for "Medical Terminologies" would define <transducer
> data>. *However* you would not have actual transducer data in a document
> created by this DTD. Not at all. Instance documents created from this DTD
> (and there would in reality only be one of them), would resemble the following:
>
> <transducer data> systolic blood pressure, arterial systolic blood
> pressure, intracranial pressure, [etc.] </transducer data>
>
> That's the beginning of a thesaurus, but I don't quite think it's the same
> as the UMLS, because UMLS doesn't try to decrease granularity. I only
> envision there being at most 50 tags (probably far fewer) like <transducer
> data>.
>
> John
>
