Alvin Marcelo wrote:

> The LittleFish Project is using GEHR but it is the Nov 1999 version. See
> the link to Littlefish from Minoru's website.
> 
> In all your remarks about GEHR, were you referring to the old version?
> Please clarify this to the list..

O.K. I have learned now that GEHR 1.0 '95 has been updated and that
GEHR is in fact maintained. So, I was wrong saying that GEHR would
not be maintained. Interestingly the CEN TC251 EHCRA standard somehow
roots in GEHR but has developed differently. Although I still see a
lot of similarities.

Now, the most obvious difference between GEHR/HRAC/COAS and HL7 is
that HL7 defines more specific objects. Now, again, the model has
changed a lot and for a documentation of the EMR model of HL7 see

http://aurora.rg.iupui.edu/RIM/harmonization/usamp2a-print.pdf

Still, HL7 defines more specific objects, such as Medication besides
Observation.  HL7 objects still try to maintain a state, although
we have turned more towards a never-overwrite-anything approach.
There are clearly a lot of commonalities. Most importantly the 
"links" in GEHR. In HL7 the Service_relationship class allows one
to build both hierarchical structures (part-of) as well as threads
(revises, fulfills, etc.)

While GEHR/EHCRA define a transaction that contains data items of any
kind or structure GEHR relies on templates, i.e., separate pieces
of information to define what medical information items you can 
expect in a certain cluster. In HL7 we put fundamental classes and
their attributes into the model. For example, a medication is 
described with a fixed set of attributes for dose, form, route, etc.
Our information model standardizes these things, GEHR requires 
templates for it.  

In HL7 we came to assume an Action-oriented paradigm for any kind
of information, this allows us to reuse the same class, action names
to mean concept definitions, guidelines, plans, orders, documentations,
goals, risks, options, scheduling requests, reminders, and other
"moods" how we call it (from latin "modus", the mood of a verb not
our psychological mood.)   This is a very powerful construct.

Then, in HL7 we have the principle workflow constructs which we
can use to fully standardize the definition of both guidelines 
and care plans. We have conditional predicates (a mood of observation)
and triggers, triggered actions (either as reflex orders, or 
actions or as reminders.) We have elaborated somewhat on specifying
complex timing which is needed for drug prescription and other 
timing patterns. I don't see this directly in GEHR.

So, to conclude, I like the greater specificity of our HL7 model but
there are lots of similarities too. 

> True or false:
> 
> 1. HL7 2.x is substantially different HL7 3.0 so we should qualify all
> instances of HL7 in
> future posts to avoid misinterpretations.

When I say HL7 here I always mean HL7 v3, only if I said HL7 v2 I meant
v2. This is because here we discuss models not messaging interfaces,
and HL7 v2 didn't have an explicit model.
 
> 2. HL7 3.0 is a moving target (it changed since last month).

Yes
 
> 3. Yes, HL7 3.0 is a moving target but is stable enough to provide the
> model we need for the information model.

Yes ... especially since feed back falls on fruitful ground.
 
> Reviewing the posts for today, only two information models have been
> presented: HL7 3.0 and the GEHR (which has been revised Oct 1999). GUnther
> seems to be very supportive of HL7 3.0 (being an active participant in
> it). But who is going to make an argument for GEHR? Can anyone contact
> anyone from GEHR (Tom Beale perhaps)? My mail to Chris Fraser does not
> bounce but I have not received any reply.

To say something good about GEHR here: 

GEHR is actually implemented, and littlefish deploys it. So,
why should one not use GEHR in an open EMR project? Why not support
littlefish? May be that's the way to go for most of us, support
littlefish, use it, improve it. Meanwhile I will stick with HL7
and may be some day our paths will cross if not merge :-) 

To be fair however, I would be careful with maintaining that GEHR
is more comprehensive if it just lacks specificity. This brings up
my magic triangle again:

                        simplicity





        generality                      relevance 
        (& evolvability)                ("meaningful-ness")


Every two of these may go in opposition to the third and it is a human
art to trade off between all of these.

GEHR decided to trade relevance and meaningfulness (alias specificity)
for simplicity and generality. HL7 decided to become a lot more simple
and general, but still wants to hold on to relevance in its model.
We will not give in to modeling just an Item, Item_complex and Link_item,
like GEHR and EHCRA is doing, and we will not use the term Observation
metaphorically for every activity in the medical record such as COAS
is doing. We will hold on to a few specific classes and more specific
attributes to support general medically relevant constructs, such as
observation, medication, workflow, problem thread. To say that GEHR
encompasses all of HL7 would not be fair, since I can say the same 
thing about ASCII (I can express everything GEHR can say in ASCII,
can't I?).  XML was really bothering with such boldness it could 
represent everything and therefore nothing else than XML is needed.

 
> Gunther, you still have not responded to the question: can we take a
> subset of the HL7 3.0 information model and use it to to define a primary
> care EMR? If yes, can we see a schema of the model and where the subset
> lies?

Yes, follow this link: 

http://aurora.rg.iupui.edu/RIM/harmonization/usamp2a-print.pdf

Whenever you see an association line end in an arrow "to stakeholder"
"to location" it means that this refers into the big HL7 RIM. But
I believe that one can simplify that part of the RIM a lot so that
implementation becomes easier.

Thanks Alvin for continuing to stir things up and sort things out.

regards
-Gunther

-- 
Gunther_Schadow-------------------------------http://aurora.rg.iupui.edu
Regenstrief Institute for Health Care
1050 Wishard Blvd., Indianapolis IN 46202, Phone: (317) 630 7960
[EMAIL PROTECTED]#include <usual/disclaimer>
begin:vcard 
n:Schadow;Gunther
tel;fax:+1 317 630 6962
tel;home:+1 317 816 0516
tel;work:+1 317 630 7960
x-mozilla-html:FALSE
url:http://aurora.rg.iupui.edu
org:Regenstrief Institute
adr:;;1050 Wishard Blvd;Indianapolis;Indiana;46202;USA
version:2.1
email;internet:[EMAIL PROTECTED]
title:M.D.
fn:Gunther Schadow
end:vcard

Reply via email to