Stream: terminology
Topic: ConceptMap $translate operation, understanding dependencies
Diederik Muylwyk (Apr 11 2018 at 14:50):
Hello!
I'm implementing operation $translate for ConceptMap:
https://www.hl7.org/fhir/conceptmap-operations.html#translate
Unfortunately, I don't understand the following input parameters: dependency, dependency.element, and dependency.concept.
What is the intent of the spec here? What is an example (or two) of how these might be used? What should I be considering when trying to implement this operation with respect to these input parameters?
If there is additional documentation available that might offer explanation in relatively simple terms, I'd love to read it. : )
Grahame Grieve (Apr 11 2018 at 15:18):
there's an example here:
Grahame Grieve (Apr 11 2018 at 15:19):
http://build.fhir.org/conceptmap-example-specimen-type.html
Grahame Grieve (Apr 11 2018 at 15:19):
this shows a mapping where you need information from other elements in order to translate the code correctly
Diederik Muylwyk (Apr 11 2018 at 15:22):
Thanks, Grahame! Will review - and possibly return with more questions. : )
Grahame Grieve (Apr 11 2018 at 15:22):
sure.
Diederik Muylwyk (Apr 11 2018 at 15:49):
Hi, Grahame.
Using the link you provided, can you identify a line item in that mapping that includes a dependency that is required to properly translate the source concept? This has yet to click for me.
Under the Destination Concept heading, there are four columns. Would the first column (Code) be the destination concept returned in match.concept and the remaining three columns be the dependencies...?
Grahame Grieve (Apr 11 2018 at 16:00):
ah I was careless
Grahame Grieve (Apr 11 2018 at 16:04):
this concept map, the codes in table 0487 actually produce multiple codes in different elements. take, for example, the code "BRSH". This produces the matching Specimen.code 258415003 (Biopsy sample), and in addition, produces a code in Specimen.collection.method of 439336003 (Brush biopsy). if we reversed the direction of the mapping, then the Brush Biopsy code would be a depedency, not a product
Grahame Grieve (Apr 11 2018 at 16:06):
you might argue, looking at this concept map that:
a/ it's not at all clear that the the snomed code 246380002 implies Specimen.collection.method - and it's not at all clear, though it could be made a little bit non-obscure by at least stating that relationship here: http://build.fhir.org/specimen-mappings.html
b/ the concept map itself is pretty confusing. And indeed, I think it needs some real work; it's an example that was prepared outside the FHIR project early on and I picked up as a good structural example of using concept map. But it's not a good example concept map
Grahame Grieve (Apr 11 2018 at 16:06):
hope that clarifies a little
Diederik Muylwyk (Apr 11 2018 at 16:20):
Okay. That makes a lot more sense to me now. Thanks Grahame!
To make sure I understand:
I have the following ConceptMap:
ConceptMap.group.element.code identifies BRSH
ConceptMap.group.element.target.code identifies 258415003 (Biopsy sample)
ConceptMap.group.element.target.dependsOn.code identifies 439336003 (Brush biopsy)
If I want to map from BRSH to Biopsy Sample, my input parameters to operation $translate should identify BRSH in either code, coding, or codeableConcept, and optionally Brush Biopsy could be identified in dependency.concept.
In the output parameters, I should get a match.concept with Biopsy Sample, and a match.product.concept with Biopsy Brush.
Is that correct? I feel as though my $translate operation would still turn up the right results but providing the dependency.concept input will narrow the results.
Last updated: Apr 12 2022 at 19:14 UTC
