On Sun, 18 Mar 2001 22:25:00 philippe Ameline wrote:
>Hi,
>
>I think "best concepts" has been misinterpreted by some people on the list. I am
>french, and sometimes don't use the most proper words.
Hi Philippe,
I think I understand what you meant. "Best idea" from a project is very different
from best product or system for health care! I think you meant best ideas from each
project.
>My opinion is that we all have taken the (difficult) problem of building a medical
>system in a personnal way.
Right. We are all biased in our view of the problem and thus our solution.
>For example, many of you are centered on hospitals (maybe because you work in) ; this
>is not my case.
Where do you work exactly? If not hospital, then clinic? What specialty?
>Some of you think that a web browser is the good interface ; I work on Artificial
>Intelligence-able workstations.
That's me. I think web-browser is an adequate interface for most things.
>I believe in highly structured patient records ; some of you probably think it's an
>utopy.
I think structured records are very import but I think the system must permit the end
users (e.g. clinic/hospital) to modify the structure to fit the system to their needs.
Thomas describes the importance of this type of flexibility very well in his
"future-proof" paper.
>However, we all have developped good components that could certainly be re-used by
>someone else, even if he has different paradigms
>
>Thomas Beale and I both used the Lego brick comparison when describing our concept ;
>in the same way, there is lots of ordinary bricks in my system, but I also have smart
>bricks I can offer, and I know I miss some bricks you probably have.
I think Thomas and you are working from a model of extensible knowledge
representation. Each concept is like a Lego brick. From there, you develop a user
interface to manipulate those bricks. If the system can manipulate those bricks
according certain rules or according to other bricks, that would be called artificial
intelligence.
I am working from a different direction but hopefully reaching the same results. My
idea is to start with bigger Lego bricks that map more naturally to how clinicians and
researchers organize elementary concepts (if there is such a thing). So, rather than
starting with an abstraction such as a "concept", I am starting with a "form". A form
(i.e. as in a "paper" form) is a more concrete and functionally-defined object, I
believe. This approach is very similar to Thomas' "GEHR archetypes" idea except "OIO
forms" are explicitly defined as conceptually equivalent to paper forms.
A consequence of this forms-based approach is that two sementically identical concepts
(e.g. blood pressure) that reside on two different forms need not be linked to each
other unless there is a functional reason to do so. For example, if two data sets are
to be merged or if either could trigger an action in the workflow.
We can further discuss the advantages and disadvantages of bigger vs. smaller Lego
blocks, but in essence both GEHR and OIO permit flexible sizing of the Lego blocks by
allowing archetypes/forms with variable content (=constraints). So, the point is not
how big the Lego blocks are, but whether they are flexible.
The other idea from OIO is that the forms are visible plug-and-play web-forms. This
makes it easy for users to see and thus manipulate these Lego blocks. In addition,
this eliminates the extra "interface-building" step since the knowledge representation
and the interface are one and the same.
I am not sure what you mean by smart bricks but perhaps that is something that we
don't have. In OIO, the bricks (forms) are not smart. The smarter elements would
reside in the workflow components (e.g. forms packets, studies, pathway) and reports.
>Andrew's interactive repository of projects is a good starting point
Thanks for the vote of confidence :-).
>(I must confess the message with its URL is in my office, and since I didn't
>understand how it works in 30 seconds, and then telephon rang...).
O.K. It is at www.TxOutcome.Org. Remember that the OIO Library is not the OIO Server.
The OIO Library uses the same plug-and-play web-forms but does not manage patients.
> Could we work on it ? I am ready to start.
Yes! Could you explain who is using the Odyssee and what they are using it for? Also,
what are your plans for it?
>As a summary, I think that if we have to work on paradigms we probably must meet each
>others (medical open source congress - sponsored by Minoru ;o) ) - but we can already
>share components (at least describe our best components), and that is the first step
>to a genuine collaboration - and more if affinity.
Right. What are the best components of Odyssee project in your opinion?
Best regards,
Andrew
---
Andrew P. Ho, M.D.
OIO: Open Infrastructure for Outcomes
www.TxOutcome.Org
Assistant Clinical Professor
Department of Psychiatry, Harbor-UCLA Medical Center
University of California, Los Angeles
Join 18 million Eudora users by signing up for a free Eudora Web-Mail account at
http://www.eudoramail.com
