|
Decision support
Compliance
CD regulations
Dispensing
Pharmacy assistants
Packaging
Alcohol metabolism
Capacity planning
Reciprocity
The Charter
Letters to the Editor
|
Decision support
Decision support systems do exactly what they say they do
From Mr S. J. Goundrey-Smith, MRPharmS
May I remind John Wilson (PJ, 24 January, p86) and others who are concerned
about the number of drug interactions they are warned about by their
pharmacy computer system that the term “decision support” means
exactly what it says.
Pharmacy and prescribing computer systems provide a logical checking
functionality for drug interactions, according to an accepted rule-base,
in order to support practitioners in their clinical decision making.
However carefully structured it is to reflect clinical importance of
interactions, a logical checking system such as this can only operate
according to a rule-base. It cannot itself determine exceptions and variations
to that rule-base. The advantage here is that every drug interaction
specified in the rule-base is flagged up to the practitioner to act upon,
or not as the case may be, and the practitioner’s actions will
be based on a consistent drug interaction data record. The disadvantage
is that interactions will be specified even if they appear superfluous.
As Robin Glasspoole and
Hillary Judd point out (PJ, 3/10 January, p15),
what may be an interaction of minor significance in one patient may be
of greater importance in another patient, depending on the patient’s
medical history — something that only a practitioner is in a position
to judge. To quote Mr Wilson’s example, there is a drug interaction
warning on his system when furosemide 20mg tablets are added to a patient
record with furosemide 40mg tablets already on it, simply because there
is a clinical effect when a patient’s furosemide dose is increased
from 40mg daily to 60mg daily. In terms of adverse effects, this may
be of little clinical significance in most cases, but only the practitioner
can make the final decision on this, with knowledge of specific patient
factors.
As someone who works in the area of IT development and drug data support
for electronic prescribing systems, I have been alarmed recently by comments
made by a number of pharmacists concerning decision support in electronic
systems.
On the one hand, there are those such as Mr Wilson who believe that only “interactions
of clinical significance” should appear on their computer screens.
The problem with this approach is: what is the cut-off point in incidence
rate for clinically significant interactions? When faced with legal action
from a patient who suffered injury due to a rare drug interaction that
was not indicated by his pharmacy system, no doubt Mr Wilson would be
quick to contact his software supplier concerning this omission. With
this in mind, I am sure that most pharmacists would prefer their software
supplier to err on the side of a comprehensive rule-base.
On the other hand, there are those pharmacists who want a computer system
to warn of every possible drug interaction and deal with all possible
risk situations. But this approach goes well beyond the remit for any
computer. As stated, a computer can only work exactly according to a
certain rule-base; it cannot formulate exceptions and variations. Only
human input — in the form of practitioner decisions, or the formulation
of procedures appropriate for the working environment — can provide
the appropriate support for clinical risk minimisation. And if there
is negligence arising from a “failure to warn” situation,
the patient will not sue computers; they will sue human beings, ie, the
clinical practitioners involved.
So let us not dismiss the term “decision support” as a “pseudotechnical
term”. Let us think about what it actually means.
Stephen Goundrey-Smith
Banbury,
Oxfordshire
|
The
Pharmaceutical Journal
* PDF files on PJ Online require Acrobat
Reader 4 or later.