[Gnumed-devel] Re: GNUmed vaccination stuff

[Karsten Hilbert]

On Wed, Jan 04, 2006 at 10:42:30AM +0800, Syan Tan wrote:

> So far, in a prototype , I've got a subnotebook with 2 panels.
> 
> The second panel is for vacc regime registration, to enrol
> a patient into one or more vacc regimes. 
> 
> This is done via a tree, with root node hidden
> and the next level of nodes being the recommender Id.
> the next level is each regime vaccination ordered in min_time_due order.
> one the recommender nodes can be activated, and if double clicked,
> a dialog confirmation asks whether to enroll or unenroll the patient in the
> regime, depending whether the patient is already enrolled or not.
> unenrolled regimes are in red, and enrolled are in green , also marked
> with X and O  (? for the colour blind)
> when a confirmation is made, the regime vaccination appear in the
> first tabs  ""vaccinations scheduled" panel.

Ah, OK, that's of course another pretty-sounding way of
enrolling patients into schedules. Can we break out that
tree class into, say,

client/wxpython/gmVaccWidgets.py::class cVaccRegEnrolmentTree

?  That way others may be able to reuse it...

> the first notebook tab was as shown previously, with some modifications.
> the checkboxes select vaccinations generically, but when one presses
> add vaccinations, it attempts to match the one or more checkboxes selected
> with the indications of any vaccine stored. e.g. if dtp is selected, and ipv 
> is
> selected,
> and hib is selected, then one of the pentavalent vaccines would match,
> and the intention is to display the matching vaccines for confirmation
> (and possibly batch number entry ) in a dialog. in order to choose which
> particular
> vaccine to give. I am doing it this  way because vaccination is linked to
> vaccine.
I see. I would have thought (and that's the way *we* work)
that I would always know the name of the vaccine I just gave
(although Richard attested to the contrary in his workflow)
so that I could just select/type that and have the
indications/regimes calculated from there. But you better do
whatever your workflow requires.

>  If no vaccine indications matches all the indications  exactly, then a error
> dialog
> appears showing that the selected indications did not match any vaccine.
Hm. What about two vaccines being given in one session ?

> I can send you the diffs to the cvs, or commit, whichever you prefer.
I am entirely fine with you committing to the tree your
horstspace notebook panel. It shouldn't unduly impact much
of anything else as far as I can see. It's then one of the
panels that would be configured to be loaded in an AU
setting (or Syan setting, for that matter).

> I had some problem with egenixDateTimeDelta, as it did not return the interval
> fields properly,
I know, I know, mxDateTime does have quirks. Please do
document the problem as precisely as possible in the code.

> so I used select " extract( epoch from xxxx) / 60 /60 / 24 .." for the 
> interval
> fields to return float values
> in days , and had a interval prettifying function to shape into days, weeks,
> months, or years, in python.
Sure, if that's what it takes.

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