Op 22-3-2017 om 12:31 schreef Thomas Beale:
Just catching up on this conversation - I am unclear on why the
original solution Bert proposed here isn't correct. What this says is:
* for the ac0001 term constraint in the model, allow the term to be
from one of ETDA or ICD10
I think, the term-binding applies to myproblem with the DV_Quantity,
where we want to allow two units, according to to different LOINC codes
(one in mol, and one in g)
So we want to connect the term_bindings to the units in the DV_Quantity,
not to the DV_Quantity itself.
It is just to make sure: the value here entered is a value which must be
understood according this LOINC-code.
I understand that this is not possible.
(If you connect two LOINC-codes to one archetype-node, an error occurs
in the Template-Editor (when generating HTML, it does another check
through the structures (maybe it is the Template-Editor making a
mistake?))).
Since this is set at the archetype level, it is stated as true for all
instantiations of the archetype.
Now, you might say, what does this 'binding' really mean, practically
speaking? Because here we are talking about an external value set, so
the codes have to come from one of the bound terminologies / value
sets. THere is some potentially useful explanation of the meaning of
'binding' in the ADL2 specification
<http://www.openehr.org/releases/AM/latest/docs/ADL2/ADL2.html#terminology_integration_concrete_codes>
(I suggest the whole of section 8 is worth reading).
However, if the requirement is to limit the possible run time
/mappings/ recorded in the data for an original term to coming from
only 2 particular terminologies, then the suggestions made here about
TERM_MAPPING may apply. Note however that these don't solve the
question of how to control the coding of the central DV_CODED_TEXT.
If the Dv_Coded_Text would allow two defining_codes, we would have no
problem at all. We would write in the value (derived from Dv_Text) the
term we wanted to write (in the organization I work, this would be in
Hungarian, Slovenian, German, Dutch......) and we would code it twice,
once in EDTA and once in ICD10 so that the other member-organizations
would also understand what that data-set is about. Some
member-organizations work with EDTA, some with ICD10, so we need both codes.
I don't understand the philosophy behind the restriction in
Dv_Coded_Text to have only one defining code. But OK, it is like that.
So we wanted to find the solution in the Term_mappings, a list of
term_mappings in Dv_Text (like Heath (and some others) suggested).
It looks very familiar then to Dv_Coded_Text. It has a value to write
some text in (DvText-value), and it has multiple mappings as
Code_Phrase, like Dv_Coded_Text also has a Code_Phrase in defining_code.
So we would like to constrain these mappings to use only code from EDTA
and ICD10, and we want to check if the two codes are really used (this
is the requirement/problem).
And now we find that the archetype-editor does not like constraining the
Term_Mappings. So we cannot model it on the level where we do the
modeling, in archetypes and templates.
We must solve this in software-coding, and that create a breach in our
barrier between the software-builders and the data-modelers (which is a
kind of spaghetti).
Bert
_______________________________________________
openEHR-clinical mailing list
openEHR-clinical@lists.openehr.org
http://lists.openehr.org/mailman/listinfo/openehr-clinical_lists.openehr.org
