Re: [Gnumed-devel] gmClinNarrative: some questions

[Karsten Hilbert]

Ian,

> Firstly "is_aoe" and "is_rfe": what does it mean if these flags are true 
> for SOAP lines other han Subjective or Assessment respectively.
It does not have a meaning. There are constraints, however, to
prevent that from happening. I don't like this field being in
clin_narrative either but didn't feel a linked table would be
justified. Perhaps it is. Would just require a few changes to
two views and to some code in cClinNarrative. Frontend code
couldn't care less. Another possibility is to just declare
is_aoe/is_rfe to be attributes that belong into
clin_item_type. What do you think ?

> (i.e how can an Objective comment be a request for encounter?)
Not.

> Would it not be simpler to have 2 more SOAP categories, say "r"
> for where S is RFE and "e" for AOE (where an assessment is an AOE)
Hm. I am not so sure.

> Secondly, what is the difference between an assessment coded
> by lnk_code2narr and v_pat_diag_codes?
None.

> I don't like the fact
> that there are two ways of accomplishing the same thing.
There are no two ways about it. You always need to go through
lnk_code2narr for any soap row including soAp. There ARE,
however, two views that convey slightly different meaning:

v_pat_diag_codes
- this view has a row for each code-per-diagnosis-per-patient,
- hence one patient-diagnosis can appear several times, namely
- once per associated code

eg. this table lists codes attached to a diagnosis *for a
    particular patient*, there CAN be duplicates apart from
    the patient ID

v_codes4diag
- this view is a lookup table for locally used
- code/diagnosis associations irrespective of
- patient association, hence any diagnosis-code
- combination will appear only once

eg this view shows any codes associated with a diagnosis
   irrespective of which patient - duplicates filtered out

> IMHO the 5 diagnosis flags (chronicity, definitiveness, laterality, activity, 
> significance)
> belong in lnk_code2narr
No. Having a row in lnk_code2narr is not a requirement for
being a diagnosis. We shouldn't kid ourselves into thinking
that just because Great Buffalo Speaks of the Shoshoni tribe
does not use ICD-10 in it's March 2003 version he does not
generate accurate and useful diagnoses.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346