gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnumed-devel] About systems of codification


From: Karsten Hilbert
Subject: Re: [Gnumed-devel] About systems of codification
Date: Thu, 29 Sep 2011 10:32:10 +0200
User-agent: Mutt/1.5.21 (2010-09-15)

On Wed, Sep 28, 2011 at 11:33:59PM -0700, Jim Busser wrote:

> The traditional and current approach (even in GNUmed) is
> *first* to input free text, and then (optionally) to code
> it. Once I would have SNOMED I would want the fields
> reversed and would want to flow (input) multiple past
> history items through a single widget like the one below

There have been studies on this with the result that:

- important information will be left out because it cannot
  be captured or people are too lazy to search for the
  "proper" code

- patients will receive invalid and/or wrong diagnosis in the EMR

- upcoding

- downcoding

- miscoding

- inappropriate batching of patients (people get diagnosed by
  the coding system, not by the doctor)

A coding system can never replace free text in the EMR. It
can only serve precisely defined secondary purposes.

> I may be wrong, but predict Karsten will doubt the
> adequacy of SNOMED to replace free text.

It is not so much about SNOMED but about the *concept* of
"normalizing" medical narrative. It never works. SNOMED has
much developed since its inception.

> But maybe he will be open to the possibility that it could
> warrant to be the primary entry, with text as the optional
> supporting field?

No chance with SNOMED being a successful business model.

> A colleague who turns out to have worked a lot with SNOMED (and is in fact on 
> a Canadian committee) told me today:
> 
> "It was interesting that in the Infoway project that
> recommended SNOMED for clinical use that 20 or so clinicians
> seemed hostile towards SNOMED but as each subsequent meeting
> happened people were more and more in favour of SNOMED. In
> the end there was no question that SNOMED should be the
> terminology of choice for clinical purposes."

We are talking entire worlds of fantasy here:

- no one here uses SNOMED
- no one CAN use SNOMED as it stands
- no one here needs to use SNOMED
- we don't even have a way to support SNOMED (for what it
  seems to be advertised to be: a foreign language)
- we are trying to release

Karsten
-- 
GPG key ID E4071346 @ gpg-keyserver.de
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346



reply via email to

[Prev in Thread] Current Thread [Next in Thread]