[EMAIL PROTECTED] wrote:
 
> The actual implementation of FreePM duplicates lots of information.

Yes it does.
 
> Each EMR's personal data is partially repeated in the Patient Account,
> then it gets in a nice percentage of cases repeated in the Guarantor
> Information.

Yep.
 
> I would like to have a Central Directory for addresses, other for
> persons (which would really be identically mapped with EMR's) and all
> the rest should be links to this id's.

This would be a sound relational implementation.  But it is one
of the keys to why FreePM's model will not break.  Granted you
duplicate information in MANY cases.  This was a concious
decision I made in order to containerize all aspects of a patient
record and a patient account.  The reason I did this was to
simplify (GREATLY) data analysis and extraction.  Your
description may be doable and may hold up to all cases.  I would
have to spend some time going back over my original notes since
we are talking about the VERY first decisions here. :-)
 
> What impact would a data-layout redesign have on the actual FreePM
> implementation.

On new ones, nothing.  On existing sites you would have to
provide scripts to move the objects around to their new locations
and create the linking information.
 
> Other question: What about getting this information from LDAP or from
> Postgres.  

LDAP support is something that I would have really liked to
have.  But it wasn't mature in Zope at the time and I didn't have
the time/knowledge to do it myself.  But this is one thing that
is on the list to implement.  Patches greatfully accepted. ;-)

> With only the basic FreePM and some fifteen EMR's my
> "server" goes **slow**, how about next years when there will be 2000
> EMR's?  Out"sourcing" the database could probably speed the thing
> quite up.

When I originally doubted that the ZODB could handle this amount
of data I first looked at some very large Zope installations. 
Then did some testing that created I believe 100,000 EMR's with
10 objects (granted that isn't very many) each.  FreePM ran
fine.   Where I did have a slow problem was when running FreePM
on a RH6.2 machine and using that same machine (PII 500MHZ 128MB)
as the client with either NS or Mozilla running in X.  What I
found was that it was actually the client-side that was slowing
down. With everything still running I can go to a Win98 IE
machine or another RH6.2 machine and do not have a problem.

The demo site is a 1GHZ 512MB Linux machine with FreePM, Apache,
MySQL, PostgreSQL and tons of other stuff running on it and the
only time I have ever hurt that machine is during an upgrade when
I import the new version move the EMR's & PA's over. Then delete
the old one.  Once I do this I pack the database.  That really
hits the CPU pretty hard so I generally do it at off-peak hours.
:-)
 

> A question about Laboratory Orders Managment:  Is there a reason why
> you do not delete a pending Lab Order in the moment a result is filled
> in and saved?

There was.  It was a paranoia thing about wanting to keep the
request around until specifically deleted. I do not know that it
has a real use for any type of auditing / accountability issue
though.  We can rethink this and I would appreciate thoughts from
the physician community on that.  BTW: THe Radiology Requests
work the same way.
 
____________________________________________________
Tim Cook, President - Free Practice Management, Inc. 
http://www.FreePM.com Office: (731) 884-4126
DOWNLOAD or DEMO: http://www.freepm.org:8080

_______________________________________________
FreePM Support List
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/freepm-discuss

Reply via email to