Re: [Gnumed-devel] Open Source Health Informatics, London, 30th Sep 2010

[Jim Busser]

On 2010-08-13, at 1:59 PM, Luke Kenneth Casson Leighton wrote:

> is it the intention to have lab requests "initiated" by gnumed,

*able* to be initiated, yes

I think that in Germany the clinician or helper would input

        clin.lab_request.request_id

based on bar coded labels provided by the path lab... I do not know how truly 
Germany achieves for these to be unique.

> perhaps even by generating an HL7 message to be automatically to the
> lab?  i assume clin.lab_request.pk would eventually end up in the ORC
> 004 in a response, under these circumstances?  if so then it really
> ought to be part of the code _now_, notes made to that effect:

agree, maybe

        default = auto-incrementing value

configurable (maybe gm-[incrementing value]) so as be compatible with 
externally-dictated values as might have to be manually entered e.g. in German 
bar codes?

> something like:
> 
>     if lr.placer_group_number:
>         req = gmPathLab.cLabRequest(pk=int(lr.placer_group_number))
>     else:
>            try:
>                req = cr.get_lab_request(req_id=lr.filler_order_number,
>                                         lab=test_org_pk)
>            except ValueError:
>                req = None
> 
> with of course a check to make sure it actually exists.  with perhaps
> some sanity "prefix" on the front to ensure that it's definitely a lab
> request expected!  perhaps even encoding the fk_patient in, as well.
> even something like...
> GM-{clin.encounter.pk}-{dem.identity.pk}-{clin.lab_request.pk) would
> be "ok, yep, that's almost certainly going to be for this patient".

yes

> perhaps even a praxis-unique identification prefix rather than just
> "GM-"

provision for such a value (as the praxis might take on, administratively, 
within its local health region) could be a good idea

... but bear in mind ORC 004 has only a length of 20! :-)

> (is there any serial number generated for each gnumed install?

no, and would anyway raise the problem you mention below

> hmmm, what if that changes by mistake, now you can't identify the
> expected incoming observational reports, hm, perhaps not a hot idea,
> unless it's generated at the time and then stored in clin.lab_request)

-- Jim