On Saturday 07 December 2002 23:59, Eric Browne wrote:

> There is a huge leap in functionality 

_And how_.  Still, the story of Linux and of some of the classics of Open 
Source describe hackers thinking "It can't be that difficult to ..." and lo 
and behold 10 Christmases later there is an implementation that is getting 
common use.

I'll pick out a couple of areas to comment on.

1.  One of the workflow things is a "work token".  This is a key to  a 
general area of difficulty  in moving medical systems from paper to computers 
but is also an obtainable advantage from doing so.
In paper, if the patient record is no my desk then I have a job to do.  The 
notes are being used as a work token, and until I have done that job and 
moved the result (the work token: the notes) on to someone else, the buck 
visibly stops with me.

Electronic systems are easier for people to ignore, particualrly if thye have 
the option of some paper systems to concentrate on as well.

Of course, the visibility of that work token is only moderate within the room 
in which I might work, if there were a dozen of me, and we each had a pile of 
20 sets of notes and their sticky paper work tickets on them, then someone 
outside would have a hard time finding the notes or deciding who had the 
token: for whom we are waiting.

And one can only work on one job at a time that way, hence the combinatorial 
explosion in delay generated by increasing sub-specialisation and team work.

Eletronic work-token tracking can improve this...

2.  I pick up from Andrew's posting a feeling that there is more "packaging" 
of medical services, than a full workflow system available in relatively 
short time here.

Everyone who is referred with new central chest pain suggestive of Angina 
should have a Cardiology consultation, which should be preceeded by an ECG 
with that data presented at the consultation.  These are likely in the same 
building...
Everyone bar a few should go on to an exercise ECG, and everyone should have 
some of a set of drugs which are backed up by a set of knowledge resources 
for the dotor, and for the nurse, and for the patient.

In other words, there is a single example of a fairly well-ordered set of 
ations that will improve the efficiency of the whole system if they can be 
grouped together and run in the right sequence, and which it hepls if we have 
in a _package_ rather than bokking individually.

There are lots more, including things like the need to measure the blood 
pressure before deciding upon the anti-hypertensive medication and the time 
of the next review in general practice, or taking the blood and getting the 
results before the doctor review of Diabetes rather than at/instead of it or 
afterward.  Both of these are more difficult to run _smoothly_ than they seem 
and both can generate benefits to patient and doctor.

So in the same slightly sloppy way I think of expert systems (they are really 
hard to do, but I can run up an opinionatd system that'll give a hint on a 
certain area much more easily, and you na grow lots of those over time 
andrefine them) I think there is merit in generally understanding the 
workflow approach and concepts as Eric very kindly has appeared to contribute 
to and in using some of those to inform some work on a manageable scale 
wherever there is benefit from them.

Trying to do an all-singing all-dancing system from scratch is very likely to 
result in a gargling stumbling system.
-- 
From one of the Linux desktops of Dr Adrian Midgley 
http://www.defoam.net/             

Reply via email to