Alan Midgley wrote
> From: [log in to unmask]
> Subject: Re: Re: Computer system problems and testing approaches
> Date: 28 Jun 1996 20:10:38 GMT
and produced a starter kit for what he (and I on your behalf) wants.
Can I ask GP-UK to provide help with putting some glosses on
this? And please PLEASE do not interpret my questions as challenges.
I am not trying to score points, but to indicate how some perfectly
reasonable requirements may affect HOW things are done, and may even
be mutually incompatible with other, perfectly reasonable
requirements, in which case the User Community needs to reach
agreement on what gets watered down.
> Mike Wells wants what we want it to do, not how:-
>
> I want a narrative of the entire history of the patient in order of
> happening, neatly typed with a little bit of laying out to make it easier to
> read.
Who will type it? And When? Is it free text, with markers defining
what sort of information it is, that are inserted by whoever
types in, (Gerard are you listening), or is it to be entered into
some sort of pre-defined sheet. Or would you like both?
If more than one person can do the typing ( and I cannot see how you
could manage if only one person can), what arrangements woud you like
for controllong access, and for resolving clashes when two or more
people wish to update the records simulataneously.
It matters you see, because this is a decision on what you want to
do, that will seriously affect how you do it?
>
> I want it to be foldable into an outline, displayable by topics or
> problems, searchable on free text.
I see no particular problems with any/all of that. There may have to
be debate about the speed of the search engine (or more accurately,
how long you are prepared to wait for a search to complete) and on
what you will do about spelling and fuzzy matching.
>
> When I open a patient's notes I wish to be presented INSTANTLY with the
> text of the last encounter, while this is being transferred to my brain I
> want the patient's summary to assemble behind it, with the problem list and
> current treatment appearing similarly behind and to the sides ready for
> attention by clicking on them (or otherwise activating thaose bits of
> record)
Fine.
>
> Somewhere about this whole lot I want the equivalent of a sticky note,
> which can be placed right on top to be seen first thing, and I want this
> note to be un-audit-trailed, ephemeral, unattributable, deniable,
> temporary. (but everything else in the system the opposite of those
> attributes)
I think there might be problems here! Who will be able to put the
'sticky note' there? How will you know whether to believe the note or
not? The reason for asking is that the mechanism for associating the
sticky note will need to be ever so carefully designed. The note
cannot go IN the record, because it has to be all the things that the
records are not, so it must be associated with the record, but not be
part of it.
>
> When I go home I want the system to be seething with thought and activity,
> looking for things to claim payment for, and thinking up hints for me about
> difficult medical problems, which it can deliver with coffee in the morning.
I like it. Of course, the engine that does all this will need to be
totally tamper proof, since presumably it has access to the records
in order to do its job. Who will be able to fire up this engine, and
who will be able to see its results. In particular, what steps are
needed to ensure that it is not possible to breach patient privacy,
by starting up lots of searches which home in on a single patient;
or is this seen as an accepatable risk?
>
> I want the ability to run ordinary programs alongside the clinical system
> (eg I want a non-dedicated word-processor, networked fax, e-mail, separate
> reference database)
That is what modern systems are supposed to be about. There will
need to be thought given to ensuring that these sub-systems do not
provide all manner of security loop-holes, and any access they might
make to the underlying data base will need to be controlled and
audited.
>
> Whenever I run a search to produce a list of patients with {a|b&&c|d} I
> want one click on a name to take me to the notes of that patient.
>
Sounds reasonable to me. Who is the 'I' who runs the search? How
important is it that there is control over the type of search, and
over the number of patients in the cells that result? Again, it
matters. If there is not control, then we have a security loop-hole.
the result
> That the sort of thing you want Mike?
Very helpful. We should do this more often!!
Mike Wells
>
==========================================================
Professor Mike Wells
Department of Physics, The University of Leeds
Leeds LS2 9JT, United Kingdom
Phone: 0113-233-2339 E-Mail [log in to unmask]
==========================================================
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|