If it's possible.
Over the years I have looked for advice on interactions between three or
more drugs - and been told the information doesn't exist.
E.g. lithium, indomethacin and warfarin.
Not sure how the results would be represented in a humanly
understandable way, either!
I suspect hospital prescribing is not "mechanically checked" (do you
mean subjected to decision support?) because much of the medication is
not linked to individual patients - hence their problems with TTOs in
EDLs (Electronic Discharge Letters).
In message <[log in to unmask]>, Adrian Midgley
<[log in to unmask]> writes
>The interactions between drugs are best described in a space of at least
>as many dimensions as there are drugs.
>
>The most usual presentation of interactions between drugs is on a sheet
>of paper, allowing pairs of drugs to be checked.
>
>An approach something like this appears to have been taken to designing
>the interaction checkers in GP computer systems.
>
>Hospital prescribing is not mechanically checked.
>
>
>An additional use of a space of large dimensional order is seeking
>interactions whether good or bad between drugs and the medical
>conditions, genotype etc of the patient.
>
>This approach should be adopted in implementions of medication automation.
>
--
Mary Hawking
|