Andrew
No doubt Dipak will tell you more about this but we had some lengthy debates about all of these ‘qualifiers’ or annotations at the meeting. It was the best attended and most productive EHR meeting I have been to - not absolute numbers, but intellectually. I will mention what happenned in these areas.
:
· negation warnings
It is clear that negation is important. Just how to do it is the issue. In our archetypes so far we have only 2 negation indicators - for Family history and adverse reaction. We have proposed an archetype for excluded diagnosis - present and in the past - as there are features of such as statement that are not the same as a positive finding. It also makes it quite clear that the diagnosis has been excluded.
An example is the difference between the person not having asthma now and the person never having had it. Clearly a person is more likely to have a recording in their notes that they do not have asthma today - when they have had it a lot in the past.
In complex structures negation becomes quite difficult - there are some key publications in the pipe line about this matter. HL7 has stated that the negation applies to all dependent act relationships - this is very difficult to be sure about and will not be sustainable I suspect.
The group agreed to drop this from the reference model and leave it to the archetypes.
· certainty warnings
There is no doubt people want to be able to express uncertainty - when it is large (even boundless!) - it is not resolved how to best cope with this need - often a ? or two (or ???) in the written record.
· urgency
This was relegated to archetyped structures - unanimously
· focus
Dropped
· role
Likewise
· potentiality
Likewise
The approach of being specific about these things in the domain concept models is seen as the best way to go…so things are getting out into the openEHR a little further in Europe!
Cheers, Sam