See the thread "RE: data model, was RE: gnumed progress", and also below:
[John wrote]
>
> I guess that there is an enormous amount of work going into
> medical data
> models. Is that correct? Is HL7 a data model? Certainly, there are
> several XML initiatives in medicine (as well as in computing:
> .NET :-).
>
Yes, a lot of work is going int this, especially at research universities
with programs (Duke, Columbia, U. Washington). Check out medical
informatics sometime and also http://www.himss.org
[John wrote]
> Is that the principle hang-up, the data model, that condemns
> EMR to 10
> years of further development before there is a widely accepted and
> installed technology?
You're just now seeing the open source flavors of EMR being popularized.
But these projects
have the advantage of not having to rewrite code. The technology is
actually quite robust. Care Management Science (http://www.cmscience.com)
has some great
clinical benchmarking tools, and there are a myriad of vendors selling
systems.
Richard
>
> John
>
> At 02:15 PM 7/30/01, you wrote:
> > > -----Original Message-----
> > > From: John S. Gage [mailto:[EMAIL PROTECTED]]
> > >
> > > Nearly everything that is being done in electronic medical
> > > record keeping
> > > has as its common denominator with other projects a
> > > browser-esque interface
> > > connecting to an SQL back-end. One can put CORBA in the
> > > middle of that, or
> > > create one's own interface, but the essential elements remain
> > > the same.
> > >
> > > Is this correct?
> >
> >That's a very broad question. Here's the short answer: no,
> I don't think
> >so.
> >
> >yes, one *can* put CORBA in the middle of a GUI front-end
> and a SQL backend,
> >but client/server and n-tier models built around middleware
> are by no means
> >almost always common denominators in EMR systems. It seems
> to be in the
> >open source world so far though. Conceptually you can and
> do have any
> >arrangement in today's systems, including a centralized
> mainframe processing
> >on a legacy system with a proprietary storage scheme. In
> such a case the
> >GUI on the client desktop runs a terminal emulator.
> >
> >I defer to publications like Healthcare Informatics that do
> surveys on these
> >types of things for the hard facts, though. . .
> >
> >Now here's a broader *intuitive* explaination:
> >
> >You will find that as of late more and more EMR systems have GUI type
> >frontends and SQL backends, but the genesis for EMR systems is in the
> >mainframe where you have text based terminals and proprietary storage
> >systems. Currently you can actually find a mixture of system
> >infrastructures, and some I've seen have a GUI interface
> that speaks to the
> >back-end through a telnet connection - a GUI literally
> "slapped onto" a text
> >based interface. But, that being said most user
> workstations seem to be GUI
> >based workstations, and where the back-end system is text based, the
> >workstations just run a terminal emulator.
> >
> >As far a middleware goes (i.e. CORBA), you have a variety of
> implementations
> >as well, and there does not seem to be any great common
> denominator in
> >middleware. The only consistent factor I see in EMR systems
> is the presence
> >of TCP/IP protocols, which of course can be seen as
> infrastructure upon
> >which to build middleware.
> >
> >Backends come in a variety of shapes and sizes, with SQL
> being very popular,
> >but by no means a common denominator among ALL systems. You
> can still find
> >some systems built around other file storage schemes, such
> as that found in
> >HP/MPE, DEC OpenVMS, and IBM operating systems.
> >
> >It is correct to say that the common denominator between all
> the functional
> >EMR systems is an appropriate data model, and the data model
> drives the
> >interface and back-end requirements. That data model can be
> implemented in
> >a variety of environments and with a variety of middleware schemes.
> >
> >
> >Hope that helps. Questions, comments . . . ?
> >
> > >
> > > John
> > >
> > > At 01:06 PM 7/30/01, you wrote:
> > > >Those interested in the progress of the gnumed project may
> > > have a look at
> > > >http://gnumed.net/whitepapers
> > > >
> > > >Horst
> > >
> >
> >
> >
> >Richard Schilling
> >Webmaster / Web Integration Programmer
> >Affiliated Health Services
> >Mount Vernon, WA USA
> >http://www.affiliatedhealth.org
> >phone: 360.856.7129
>
