Re: [Gnumed-devel] vaccination indication input panel

[Karsten Hilbert]

On Thu, Jun 03, 2010 at 11:15:51AM -0700, Jim Busser wrote:

> Would the dialog be as follows where, with the existing:
> 
>       Vaccine: (this is a field; upon widget invocation, this defaults blank)
> 
>       Indications: (checkboxes, upon widget invocation, these default 
> non-selected)

That is correct.

> and the user can do as follows:
> 

> - begin typing a value into the Vaccine field, which will
> (or can) invoke a phrasewheel of existing vaccines. When the
> user selects a previously-created vaccine, the
> previously-associated checkboxes will be fetched, and the
> screen will be updated to show those checkboxes

Correct. That happens when the user leaves the Vaccine:
field. Upon which the indications-area is also disabled (but
still shows the indications associated with the vaccine).

> - will the checkboxes be always "live", so that -- after
> fetching an existing vaccine -- if a checkbox gets
> inappropriately unchecked (or newly checked), this change of
> state will be saved with this vaccine definition, and will
> affect all patients against which this vaccine had been
> registered?

Neither will this happen (because there's no code for that)
nor can it happen (because the user *cannot* select
indications in this situation).

> - alternatively, upon fetching an existing vaccine,
> prohibit the checkbox indications to be altered.

That is what happens.

> Only if the
> vaccine name would get edited should the checkboxes again be
> allowed to be altered. This stands to confuse the user
> *unless* the GUI will support to auto-dim (read-only) the
> checkboxes on "vaccine load" and auto-enable checkboxes
> immediately on any vaccine name edit.

Well, re-enabling happens when the user leaves the Vaccine:
field again and the string does not represent a vaccine
anymore (the field will also turn pink then).

> - if the *first* action in the dialog is not to input a
> name fragment into Vaccine but -- instead -- to checkmark
> one indication, it sounds like the user will be presented
> with suitable vaccines among which the fake
> single-indication one.

No. checkmarking does not interact with vaccine selection.

> If none of offerings are suitable, an
> Add button will be available in a no-match / multiple match
> dialog?

The Add button is always available, regardless of Vaccine:
or Indications:. Vaccines do not get auto-created nor will
the add-vaccine dialog be auto-shown from the vacciation
edit area.

> -- what if one or more indications *were* pre-checked
> before inputting a vaccine name? Then, if a name is
> supplied, this would be a "vaccine creation"

No. Upon leaving the Vaccine: field the indications are
updated to that vaccines indications and disabled.

> -- for any one Vaccine, will batch numbers be saved so
> that, upon a search, the user could maybe re-select an
> already-used (matching) batch, or would the user select just
> the Vaccine name but the last-used batch would auto-populate
> the field which the user could keep or change, or would the
> batch be left out altogether and require to be, each time,
> entered?

Currently the latter. However, this is a phrasewheel :-)

It now does this:

                context = {
                        u'ctxt_vaccine': {
                                u'where_part': u'AND pk_vaccine = 
%(pk_vaccine)s',
                                u'placeholder': u'pk_vaccine'
                        }
                }

                query = u"""
SELECT code, batch_no FROM (

        SELECT distinct on (batch_no) code, batch_no, rank FROM (

                (
                        -- batch_no by vaccine
                        SELECT
                                batch_no AS code,
                                batch_no,
                                1 AS rank
                        FROM
                                clin.v_pat_vaccinations
                        WHERE
                                batch_no %(fragment_condition)s
                                %(ctxt_vaccine)s
                ) UNION ALL (
                        -- batch_no for any vaccine
                        SELECT
                                batch_no AS code,
                                batch_no,
                                2 AS rank
                        FROM
                                clin.v_pat_vaccinations
                        WHERE
                                batch_no %(fragment_condition)s
                )

        ) AS matching_batch_nos

) as unique_matches

ORDER BY rank, batch_no
LIMIT 25
"""

> - what happens for unsupported indications?

One cannot currently enter such a vacciation.

> enter instead as a plain soaP line (pending a future release of GNUmed)?

Yes.

> - when an exact date of prior vaccination (based on patient report) cannot be 
> recalled and entered with certainty, likewise use the Soap approach?

Yes, or decide on an approximate date that's clinically
meaningful. It doesn't really matter whether the patient had
his second tetanus shot on March 3rd or 4th 1979. What
matters is that she had 3 shots between 1975 and 1980.

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