Re: [Gnumed-devel] codes for narrative - how to store

[Ian Haywood]

Karsten Hilbert wrote:
> codes from coding systems (LOINC, ICD, ICPC, Ozcode, ...):
Permit me to digress briefly:
I would add ATC (drugs) and MBS (surgical procedures, in AU) too.
This covers soAp and soaP pretty much, except for referrals (easy
to come up with a list) and psychotherapy (not even sure if coding this makes 
sense)

As has been said, it would be interesting (although less important) to be able 
to code
Soap and sOap too. ICPC does have some subjective codes, but is non-free.
In the fullness of time, we could have a background system that trawls 
clin_narrative for the most
common uncoded phrases and submits them for consideration as "Gnumed codes"

>    - once-coded narrative always reusable without recoding
This is very useful. In fact vital if we expect real-life GPs to do any coding.
>    - avoids code-narrative association mismatch when narrative is changed
>    Disadvantages:
>    - cannot associate different codes to same narrative
>      depending on patient (is this needed ?)
IMHO, no, as the narrative entry is more specific than the code.
(so several different narratives may link to the same code)

> Problem with both approaches: They do not allow to specify
> a *part* of a narrative row a certain code is to stand for.
> Whether that is really needed remains to be seen. What *is*
> possible, however, is to associate several codes with a given
> narrative without regard for the fact that some may apply to
> part of the narrative only.

AFAIK in ICD-10 you sometimes use a code for the bacterial species, and a code
for the site of infection, e.g. S. aureus endocarditis, but I don't know ICD-10 
that well.
(so that's 2 codes for the one narrative, but it's obvious which code refers to 
which part)

In other circumstances, IMHO, the narrative should be broken up. There's no 
reason why
we can't have multiple entries in each SOAP category for a given SOAP note, 
each narrative
being the smallest logical phrase of the history.

> I certainly favour approach #1.
So do I.

Another suggestion is to let the codes table have an "ancestor" column,
(which can be NULL) which allows hierarchies of codes where this makes sense.
This allows us to extend ICD-10 by having broader clinical diagnoses covering 
multiple ICD-10 codes,
("chest pain FI") plus finer ones, i.e. (diabetes, type II, insulin-dependent, 
difficult-to-control")


Ian

signature.asc
Description: OpenPGP digital signature