Hello Andrew,

Wow, a little bit of discussion on OpenHealth and my mailer starts to groan!
Just think what an inbox would look like on the SolvingWorldHunger list. :)

From: Andrew Ho [mailto:[EMAIL PROTECTED]]
>Hi Todd,
>  I disagree. If the so-called "future-proof" concept is of any value, it
>ought to be applicable to systems such as VistA. The value of open source
>approach is that we should be able to incorporate improvements developed
>elsewhere into VistA.

I believe that seatbelts save lives everyday but I don't think that they can
be easily retrofitted into 1909 automobiles.  (Yes, I know that is a bad
example, but it has been a long day and it is all that I can come up with on
the fly.)  My confused thoughts are that VistA has clearly been a relatively
static model with some room for growth.  This clearly is not the
"future-proof" direction that any of the advanced systems on Openhealth have
been trying to promote.  A somewhat better example might be that I read many
books that are written in the '20s and they are full of racial and social
stereotypes that are inappropriate social behavior today.  I don't assume my
current attitudes about those issues are relevant to enjoying a story that
is full of what was socially acceptable behavior in those days.  I feel that
VistA is what is and that we should try to extend it to the best of our
ability, but that it might not be possible to make it "future-proof"

>I do feel that VistA also contains highly functional features that other
>system still cannot match. The challenge is to find ways to synthesize
>new systems that improve upon existing systems.

I really believe that it is a combination of factors, some of which we have
no control over.  IMHO, the choice of underlying technology has been a very
wise and extendable platform.  The highly modular design with published and
well enforced interface schema has been invaluable to allow end-user and
nonuser third parties to adopt new technologies.  A long-term commitment
that so far, politics and money have not been able to sway.  I know some of
the history of the VA and DHCP but the "underground railroad" is hopefully
still going strong.  I think that the time element has been one of the most
important factors, since only through long-term use can a product truly be
debugged enough to really work well.

>If VistA is useful in its current state, can we improve it by adding
>"future-proof" features? What if we introduce GEHR archetypes or OIO forms
>into VistA?

I have read Thomas Beale's answer and defer to his wisdom concerning his
product GEHR.  I don't believe that we can add GEHR style archetypes without
massive modification.  It might be possible to create run-time templates and
layer something like OIO as a UI for the M datastore.  I don't know, but I
think it any case it would be serious work to "future-proof" a static data
model.

>Existing schema may remain "static" - however, I believe it is reasonable
>to extend existing schema through a "future-proof" module.
>Thereafter, the static schema can be incrementally replaced with
>"future-proof" schema - as necessary. Do you see what I mean?

Andrew, I don't really follow you because I don't see how a "future-proof"
module can remain "compatible" with old modules and still maintain
"future-proofness".  Hidden assumptions are the lifeblood of static or even
semi-static data models and those hidden assumptions will break any
"future-proof" module as soon as it comes online.  

>True. However, as Thomas Beale nicely pointed out in his "future-proof"
>paper, customization / modification are huge challenges. So, as long as
>you use VistA at a VA Medical Center and your users are 100% satisfied
>with what the system offers (or can be simply shooed away), VistA is a
>perfect system. (Of course, VistA is not the only system that employ a
>static schema).

This is exactly the point that I have pitifully trying to make.  Customizing
VistA for our shop is going to be huge undertaking, nit impossible, but very
large task nonetheless.  It clearly has issues and some of them are pretty
major, but it is the only open-source suite currently that can run a
complete hospital.  OIO and GEHR and others can run pieces of the puzzle but
not enough to implement here.

>The OIO system has recently been introduced into a hospital (Nandalal's
>site in Sri Lanka). The data schema is nowhere near as comprehensive as
>VistA - yet. However, the data schema is so easy to extend and modify
>(="future-proof") that the size of the data schema is almost a non-issue.
>Rather, the system evolves as staff at the hospital decide what
>information they like to maintain electronically (and how).

I wasn't aware that you had a installation away from your immediate area.
Congratulations are in order and I hope that it is but one of many
installations.  However, a small number of hospital installs doesn't provide
adequate comfort for senior executives trying to plan for the future.  I
would respectfully disagree with the size of the data schema being a
non-issue.  As far as I am concerned it is a major issue and one that OIO
doesn't take fully into account.  

A certain critical mass is needed before hospitals will use any information
system product and OIO doesn't have that critical mass yet.  At present, it
seems to me to be well-suited for niche installation, not to run a
full-blown vitals, radiology, pharmacy etc.  It will be years before you
have assembled enough "packages" to even be considered a viable alternative.
It is at this point that the schema problem rears it ugly head.  Commercial
vendors have "solutions" today and that is what management will see.  There
is no way that I could talk anyone here into slowly growing either GEHR or
OIO or anything else for that matter.  An opens-source solution is
considering risky enough that it would get exactly ONE chance to prove
itself.  It has to cost nothing and run on throw-away hardware and work 110%
in the first 30secs of startup.  That double-standard is what will kill most
open-source medical solutions in hospitals.  Only something that has
survived the wars like VistA for almost 30 years is going to be even
remotely considered.  Selling VistA is damn near impossible enough, trying
to go for a gradual solution is unacceptable.

>I think it would be helpful to consider how best to improve VistA and
>learn from VistA. I doubt any given system has a monopoly on all the good
>ideas :-). We really should take full-advantage of the open-source
>methodology to study each system, to clarify the design choices and
>intellectual contributions. This means challenging each developer to
>discuss their work in public :-).

Obviously, VistA allows you to study the source code in depth and any good
ideas are welcome for the taking!  I just don't know how much that you can
retro-fit into VistA from other systems.

> I hope that over time OIO and GEHR and others might be able to run a
> hospital and then the "best" system would be the one that meets our
> needs the closest.

>My vision is somewhat different. I believe the most useful system would be
>highly flexible so that end users can easily mold it to fit their specific
>needs.

If and only if, there is adequate numbers of modules available and the
perceived risk is low enough that risk is not a problem.  Otherwise, it will
not be accepted and the best that can happen is the "underground railroad"
history repeats itself. 

Todd Smith <[EMAIL PROTECTED]>

Reply via email to