thomas.beale at openehr.org
Tue Apr 11 10:39:23 EDT 2017
I don't disagree that we may need a more subtle analysis in the future
that deals with the supersmart machines that are now mixing pure
measurement with interpretation. For now, I would suggest that the best
dividing criterion for choosing Observation or Evaluation is to ask the
* is this information (only) 'about' the individual patient?
o YES => Observation
* or does it contain inferences made by comparing individual data
against knowledge of clinical categories?
o YES => Evaluation
if it seems to be mixed, e.g. a path result containing both the raw
result e.g. microbiology 'organism = giardia' and an interpretation
'probable giardiasis', you have to consider using both an Observation
and an Evaluation, or that only one of them is really relevant.
There is a good philosophical and practical reason to distinguish
between Observation and Evaluation - Observations can be expensive to
make, but are generally reliable. In difficult cases, Evaluations can be
wrong, and need to be revisited. The EHR needs to be able to show the
observational data distinct from subsequent interpretations so that an
investigation for a difficult case can proceed efficiently.
On 11/04/2017 02:58, Bert Verhees wrote:
> Saying this, it comes to my mind that often complex devices, also
> supported by computers, AI-algorithms, etc, not only observe but also
> So what comes out of the machine can be a mixture of observations and
> evaluations, hard to distinguish, and also rather academical to
> Maybe the reference-model is in need of another term, that can be
> partly observation and partly evaluation. And when we have that term,
> it is questionable if that term shouldn't have been there at the first
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the openEHR-clinical