Thomas Beale wrote: > Gerard Freriks wrote: > >> When medication is considered be aware that there is already a huge >> amount of requirements collection in HL7v2, HL7v3 space and in >> European standards. >> I think that with a limited amount of work the Clinical Models can be >> produced and discussed. >> In addition we need this all over the world and there is one European >> project needing it : epSOS. >> >> > * > I would suggest that the 2 key requirements today are: > a) being able to merge medication lists for a given patient from > multiple sources > b) being able to handle entries in the list that were prescribed (and > that are recorded in some structured, standard way) and medications that > the patient says they are on already, including non-prescription and > herbal etc. > > Being able to do this would generate a list that can actually be used > for decision support (interactions testing) for a new prescription. > > I don't think either of these is dealt with directly by any of the > current standards. > > - thomas beale > > * > > _______________________________________________ > openEHR-clinical mailing list > openEHR-clinical at openehr.org > http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-clinical > > > > How would the original sources of truth be monitored / recorded if you merge two or more lists from say different authors or sources? Would there be anyway for the decision support algorithm to know it was dealing with two different authors and then possibly a third for the final concatenated list. Certainly that would be needed for any medico-legal audit if errors in one or other of the originals were responsible for a negative outcome. How might you deal with two identical (or maybe slightly different) medications in each list - would you sum the amounts (not good), ignore one (which is the more correct and how would an alogorithm know), ignore both and seek advice (not optimal) or keep both with some form of flag back to originating list supplier and not publish)?
Interesting issues. David
