Re: [Gnumed-bugs] GNUmed 0.9.3 measurements selector does not identify mixed within-cell signed states

[Jim Busser]

On 2011-04-27, at 6:25 AM, Karsten Hilbert wrote:

> On Wed, Apr 27, 2011 at 02:14:29PM +0200, Karsten Hilbert wrote:
> 
>>> Within Kirk, I selected to
>>> 
>>>     Review and sign 
>>> 
>>> but when given the opportunity to select from within the CRP, I decided to 
>>> select only one and then I took responsibility for all results except the 
>>> one CRP value.
>>> 
>>> The result is measurements updated to correctly display the abnormal state 
>>> (in red) and that a CRP result remains unsigned (screenshot), but it seems 
>>> that in this state the button to
>>> 
>>>     Your unsigned | All unsigned [Select]
>>> 
>>> does not correctly query the state therefore does not select the cell?
>> 
> Indeed. This is by design. The relevant comment in
> the code sayeth:
> 
>       # loop over results in cell and only include
>       # those multi-value cells that are not ambiguous
> 
> This logic is applied if the selection strategy included:
> "unsigned_only = True" which, indeed, it does given that the
> user initially pushed
> 
>       [SELECT]
> 
> after either of
> 
>       (*) Your unsigned       ( ) All unsigned
> 
> was chosen. One can still manually select the cell and
> effect signage. Note that this is about the selection
> strategy of *grid cells*, not about results.
> 
> Karsten

Ah so the priority is, in the case of a mixed cell, to avoid to re-sign one or 
more cell elements (granular results) what had already been signed, at the cost 
of failing to select the results that do still need signing.

My concern is that a cell with abnormal results might be left unsigned, so as 
to let the usual (most responsible) clinician sign and take responsibility for 
it, except that this result may get lost (may get difficult to see)  inside a 
large collection of already-signed results for this patient.

The clinician who is *supposed* to sign might:
1) in the inbox, be advised (unreviewed results for patient…)
2) double-click and be brought to the measurements viewer
3) select own (or all) unsigned… but the mixed value cell won't be selected!
4) it may not even be visible on screen if it is in a scrollable area below the 
screen or (in the case of many more recent columns) to the right of the columns 
currently visible.

So I am wondering how such values can be more easily brought into the 
clinician's view? Is it feasible / sensible to add (to the right of the 
"Select" button) a Filter button? The filter button might bring into view only 
those cells which contained one or more own-unsigned or all-unsigned values 
(even if the cell contains "mixed" values"). That way, after attending to  the 
unambiguous cells, leftover cells (containing mixed values) could more easily 
be "found" and properly signed individually?