Re: [Gnumed-devel] error in gmPG?

[Karsten Hilbert]

> I added code so it creates a allergy health issue if
> one doesn't exist for the current patient or use
> the existing allergy health issue, 
This is /maybe/ OK. Others please comment on the clinical
correctness.

> it gets the current clinical encounter , which is bound
> to the creation of a gmCurrentPatient object,
No. This would mean I get many different encounters for the
following scenario: a) the front-desk stuff pulls up the patient
for waiting list registration, b) the nurse pulls up the
record for height/weight measurments and blood sampling, c) I
pull up the patient for initial assessment, d) the X-Ray
assistant (Richard: still looking quite good :) pulls up the
patient for X-Rays, e) I pull up the patient again for final
assessment etc.

However, the real encounter is the entire thing (yes, I know,
the id_provider in clin_encounter suggests otherwise, this is
a known, still unclean issue. but the comment on
clin_encounter.id_provider hints towards that and the "real"
person entering the data is captured in audit_mark.modify_by).
So there needs to be business logic in the application which
goes something like this:

Upon activation of a patient a row in the table
"encounters_in_progress" (that table's still missing since I
haven't fully understood the implications yet) is created if
one does not exist for the patient. If one does exist some
heuristics determines whether that encounter's still in
progress: difficult to define that, though, may be based on
some timestamp-since-last calculations plus some "unique"
events [reading in patient chipcard here in Germany] plus some
user intervention...

(Syan, that's what I meant by "one needs to understand the
domain".)

> and it creates a new clinical episode , using the 
No. The app hasn't got the slightest clue whether that's a new
episode. And it cannot, either, currently.

Despite of the above let me say that your effort is
appreciated.

> When I go back to the demographics screen and select another
> patient, I get a portal NOT found error.
> I don't think closing the most recently opened cursor closes the
> connection , does it ?
It should not.

> >>>>>>>>>>>> I think the root error is here.
[...]
> [ERROR] 
> (/home/sjtan/gnumed/gnumed/client/business/gmClinicalRecord.py::execute:335): 
> Unable to check existence of id 5 in identity
Do you, by any chance, really do use different databases for
demographics and clinical ?

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