I don't think it is essential to export the information from one system
in such a way that it can be imported into another with anything more
than a straight listing of what happened.
IE it is not essential to import the linkage of a prescription to the
condition for which it was given.
THe concept itself of course is not as simple as it might appear, since
not all systems link diagnosis and treatment at all, and those that do
may be defeated by a GP reasonably wishing to provide one prescription
for one problem and another for another but not as two separate actions.
Also drugs may be given for two problems, say angina and hypertension.
I suppose both are in the same problem tree, but I will settle for an
account of what diagnoses have been recorded, and another of what
treatment has been given.
Bringing a Surgery Manager record into a POMR based system will pose
the (superable) problem of deciding what problems to attach the
treatments to, I suppose it can all go in as unlinked notes under old
problems.
However, if one is exporting such info from a system that does link
problem to treatments, I don't think dropping the contents of the
fields will do it.
You really need to go toa markup system, I think, so it becomes
something like:-
<patient record>
<Summary>
G2 Hypertension
G. MI
</summary>
<narrative>
<problem number=1 name="Heart trouble" code="G....">
<encounter date=#1 June 1998#>
Hypertensive, treatment increased
<prescription>
Atenolol tabs 100 mg take 1 daily [56]
<prescription>
</encounter>
</problem>
<problem number=2 name="Minor illnesses">
</problem>
</narrative>
</patient record>
I think we can all puzzle that out, and a little bit of cleverness in
the word-processing input part of the receiving system could make the
tags more easily read, but, importantly, could be left out and the
record just printed on paper without it being wholly unusable.
A system which for instance didn't link problems to encounters wuld
just ignore the problem containers, and put all the encounters in in
date order, one which didn't link prescription to problem would
similarly drop the prescriptions into the Old Rx History bucket where
it wouldn't matter much to us that they were not parsed out to reflect
the difference between text based systems that give the PPA description
in a single field, and database based versions that use separate fields
for drug, form, and strength.
One can make a case for either, I suspect the PPA would say that the
"Atenolol tablet 100mg" is a concept which corresponds to the
real-world object, and avoids offering atenolol capsules if they don't
exist...
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|