100% Correct are your remarks.
Most developers start selecting the target machine, the database system,
the suite of tools to develop the applications.
How many start thinking about the processes a machine might best be used for?
What infrastructure is necessary?
Most are satisfied if: something has been concocted that can be recognised
by GP's as a CPR, the data can bee stored and retreeved. Perhaps a nifty
something is added, like a handy thing to write prescriptions or produce
bills and voila a SYSTEM.
And most systems work in the little island only. Not able to exchange the
data meaningfully between other users, even if they are using the same
system.
In short: these crippled systems are not able to participate in the GENERAL
exchange of data between the many actors in the Health System.
What's needed is a kind of INFRA-STRUCTURE standards:
-a safe way to exchange data (security: time stamping, authentication,
non-repudiation, etc, etc),
-a general way of coding the meaning of the data (e.g. a 'virual medical
file'). Using something like SGML.
And then systems (any system) which use the INFRA_STRUCTURE to exchange the
data, operate up on the data, archive the data.
And off course the User-interface is important. Not something to be put in
at the last moment by the apprentice. It's a thing for experts and the
second on the priorities list after the Infra-structure.
I have a dream.
Greetings
Gerard Freriks
At 11:12 on 15/9/96, Andrzej Glowinski wrote:
> Prit Buttar writes....
>
> >I am involved in a project to see how difficult it is to devise a GP
> >computer system using existing commercial software; it is proposed to
> >produce a system using either Access or FoxPro, running under Win95.
> ......
> ......
> >I have my own list of must-haves and desirables; please send me
> >yours!
>
> I believe the major difficulties lie not in the choice of what
> "commercial software" to use but in designing (and then building) a
> usable system itself. There are several crucial questions to ask, but
> unfortunately few get asked; they need to be answerd if there is to be
> any significant advance on the first generation of systems being used at
> present. There are at least four levels at which such questions may be
> asked. This is not a complete list!
>
> The requirements/design stage:-
> - What tasks (clinical and others) should the system support?
> - What aspect of these tasks will it help me with? (what will it DO?)
> - HOW will the system help me do what I want?
> - What are the implications for using and integrating the system in my
> day-to-day life in the surgery, and for other staff?
>
> Aspects of infrastructure:-
> - What are the information requirements of these tasks? (i.e. what
> information will the system need to hold, and what will I have to enter
> to get it to do things?)
> - What sort of information infrastructure is needed to support these?
> (What representation of clinical information is needed?; are existing
> ones good enough? do they support the tasks I want to carry out?)
>
> Implementation of the system
> - Is it possible to implement the system easily and RELIABLY? Securely?
> - If so, what software platform should be used to ensure all our
> requirements are met? (some common commercial software on PCs can't cope
> with the infrastructure needs)
> - How do we make the information entered portable
> (a) across systems and hardware platforms
> (b) over time, especailly as medical ideas and software capabiilities
> change
> (c) between practitioners (so I don't misinterpret what YOU have
>said)?
> - How do we build the necessary communication capabilities into the
> system to cope with
> (a) clinical communication - individual patients
> (b) administrative data
> (c) epidemiological data (NB some people want to interrogate your
> database remotely)
>
> The interface
> There are too many questions to list - but the interface has to be
> comprehensive, intuitive, easy tu use but hard to foul up, etc.
> Unfortunately an area that loses out every time because people would
> rather think about whether a 486 or a Pentium is needed.....
>
> Lastly, the hardware. OK, the target has to be something that is either
> already in place or likely to be so in the near future. But remember
> that the computers NASA used to get people to the moon and back made an
> IBM PC with an 8088 chip an 16kB of memory look like a Cray.
> My experience tells me to get the design right first, then worry about
> the implementation; at worst, not all of it will be implmentable today
> but the design won't be out of date tomorrow.
>
> >>>>>>>>>>>>>>>>>>>>>>>>>
> Andrzej Glowinski
> [log in to unmask]
Gerard Freriks,huisarts, MD
C. Sterrenburgstr 54
3151JG Hoek van Holland
the Netherlands (31) 174-384296/ Fax: -386249
ARS LONGA, VITA BREVIS
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|