I have asked Andrew Ho a question about use of OIO-Zope to enact guidelines for preoperative testing. My question, slightly edited, is posted below, along with Andrew's interpolated responses. We think the rest of the group might be interested.
For anybody who wants to see the actual table of guidelines that I am talking about, write to me individually and I will mail you a copy.
Gary Kantor, MD
University Hospitals of Cleveland
> We have a set of guidelines for preoperative testing. It's intended for
> use by surgeons or other non-anesthesiologists. It's in a large table
> with tests along the horizontal axis, and conditions (age, medical conditions, medications,
>surgical blood loss) that prompt testing along the vertical axis. Because of its size, many
>find it unusable.
It depends on what you need it to do. As a psychiatrist, I find it quite
informative and concise. :-) Did you receive different feedback from the
intended users?
> I would like to build a web-based form which would allow selection of
> applicable patient conditions and surgical procedure, and produce as
> output a list of recommended tests as well as an indication of whether a
> patient should be seen preoperatively by an anesthesiologist.
Right. This is called "guideline enactment". Ideally, you will be able to
author the "guideline" through a nice GUI tool and automagically generate
the requisite forms and conditional output. The guidelines should be
portable (e.g. XML document) and the "enactment engine" should be able to
grab patient information from a medical records system.
> Ideally the form would be easily modifiable to reflect changes in
> guidelines.
Or, the forms will be automatically changed when you change the guidline!
> We have experimented with an Access database-driven approach but I have not
> been satisfied with the user interface or the backend.
Part of the motivation for adding complex workflow capabilities to the OIO
system is so that we can support guidelines enactment. This need is
clearly present across medical specialty.
> I feel that such an "application" would be generally useful provided there
> were easy tools for a group to change the underlying guidelines.
Yes, I agree completely. The "forms" currently modeled in the OIO system
is the first step towards having an automated/semi-automated
"guidelines"-capable system. The fundamental building blocks must be
sufficiently flexible and easily customizable/generalizable.
It is easy to custom-build an "application" for each specific guideline.
However, interoperation between guidelines will be a nightmare and
hand-crafting each guideline application by hand will be quite costly.
(The is my understanding of the state-of-art.) *This is especially true
since guidelines must be frequently updated to incorporate new medical
advances and discoveries.*
> Any suggestions for the architecture?
We started to discuss this on the open-outcomes-general list.
See: http://www.geocrawler.com/archives/3/5445/2001/9/0/6654759/
and http://www.geocrawler.com/archives/3/5445/2001/9/0/6668921/
If you have a specific need to enact a particular "guideline protocol",
perhaps I can start implementing it. Nandalal's group may have similar
needs for multi-step workflow enactment.
> If you think this issue might be of general interest to the openhealth group
> (which seems pretty quiet these days) let me know and I'll post there.
Sure (you are welcome to copy my response if you like). There were some
general discussion about workflow modeling last year - but not much action
either. My view is that the first step is to model the data/metadata
in a flexible manner. So far, GEHR and OIO are the two projects working
towards that. With that foundation, it becomes possible to build flexible
workflow on top.
Hope that makes sense,
Andrew
---
Andrew P. Ho, M.D.
OIO: Open Infrastructure for Outcomes
www.TxOutcome.Org
