|
Hi all Decisions support systems in health are aimed at reducing the probability
of harm and increasing the probability of positive outcomes. While there are
certain events in medicine that are 100% deterministic – decapitation causes
death, the majority of health issues, diagnostic tests and treatment options are
based on probabilities. The very basis of medication therapy is probabilistic if one looks at
the whole clinical trials process. The fact that NPS has decided that there is
an apparent inadequacy of drug-interactions knowledge base doesn’t
guarantee that the probability of outcomes will improve if one doesn’t
address some pretty basic building blocks and that the work needs to be mindful
of what is happening now and how they can quantifiably improve that situation. Yet has the NPS looked at how they will improve the sensitivity and
specificity of what they are trying to achieve with this exercise. How will the medications be coded and classified in the knowledge base
and how will the clinical systems talk to this knowledge base? There is no
standard medication code set at the product level that is then classified at
generic, therapeutic class etc. that all vendors are using. NEHTA and others
are obviously keen to get this sorted. Apparently there is no one knowledge base that appears up to the task
hence the need to review the literature and synthesis of experiences to provide
a clearly articulated risk value. I am not sure whether what is being
articulated is clearly identifying the inadequacies in terms of quantification
of poor outcomes, risks etc. It appears based on individual opinions and focus
groups rather than extensive quantifiable research. It is a little bit alarming that the PIs have been questioned to their usefulness,
when isn’t that a fundamental requirement under TGA type rulings? Shouldn’t
PIs be better kept up to date? Should the CMI also be reflective of changes in
evidence? There is evidence that drs don’t record drug allergies,
diagnoses/history or keep complete medication record so whether you have a
super drug interactions data base is not going to help much if drs don’t
record these things. Drs don’t record these things for a number of
reasons. Shouldn’t we be improving our understanding of these things and
improve this area. It could be argued if these were better recorded the relevant
increase and decrease in probabilities would be more effective than a super
knowledge base and would have better flow on effects to other areas in
healthcare. I think you will find that a dr is more likely to get litigated against
as a result of keeping poor records rather than whether they were using one knowledge
base over another. I would like to know how many drug interactions are happening that will
not happen if this super knowledge base is going to come in to play. Of course
various people are saying it is optional and given that not everyone agrees
with NPS this will also affect its usefulness and uptake. We have seen debate
on 1st line treatment for hypertension most common problem in
general practice but only occurs about 6 per 100 encounters. There are hundreds
and thousands of other condition combinations that people have. Keep in mind always that decisions support systems SUPPORT clinical judgment.
Many drs would argue of course that clinical independence affords them the
right to treat the patient given the patients specific circumstances in a way
that they feel will provide the best probably outcomes for that patient in
consultation with the patient. I believe that this is largely evidenced based
and does give the patient a good chance of a positive outcome, but I am sure we
can improve the odds. Geoffrey Sayer HealthLink Ltd From:
[EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of James Reeve NPS is convening an expert
panel to review existing drug interaction decision support in prescribing and
dispensing software in NPS is also investigating the
feasibility of offering drug interaction decision support via a Web service,
using the drug interaction reference source identified by the expert panel. The
Web service would be maintained by NPS, while the drug interaction information
would continue to be maintained by the developers of the information. By
offering drug interaction information as a Web service, NPS hopes to create an
opportunity for interested vendors to link prescribing or dispensing software
directly to a single source of quality drug interaction information. Users can
then choose whether they use existing drug interaction decision support, or use
the content delivered by NPS (or both). The rationale for this
project stems from consultation in 2002 with general practitioners (GPs) using
5 different prescribing software packages. GPs thought that drug interaction
alerts were not always relevant to the prescribing decision. One reason for
this is the inclusion of many theoretical and minor drug interactions. NPS
reviewed drug interaction information offered by 4 commonly used prescribing
packages in James Reeve The
contents of this email and any attachments are confidential and |
_______________________________________________ Gpcg_talk mailing list [email protected] http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
