using FHIR for CDS ontology

claude: need to look at terminology first
... wanted to represent concepts and collections of concepts in skos
... have been aligning terminology concepts in FHIR and SKOS
... HL7 captures some new concepts, e.g. a value set, that are not in skos
... when you think of a value set, you might think "it's a collection"
... but there are more semantics
... e.g. it can be a (hierarchical) schema
... or an extensional definition of a set

<Kerstin> Great point bout Value sets this is a hot topic in CDISC as well http://kerfors.blogspot.se/2013/10/the-future-of-cdisc-cts.html

claude: or an expression of the intensional creation of a set (e.g. restriction from another value set)
... skos captures collection and defintion of scheme
... also looking at broader, narrower and related
... for clinical terminologies, we need to specialize them a bit
... e.g. part-to-who vs. is-a
... skos encourages this distinciton

<Kerstin> 4257

claude: for related, it's a broad concept in skos but the docs tell you how to specialize it
... skos has the notion of broaderThan, narrowerThan plus the transitive variants

ericP: if you want a transitive related, do you make a subprop and say it's owl:transitiveprop?

claude: yep

ericP: could you create an ontology for the intensional definitions, i.e. restrictions? would it be worht it?

claude: first priority is just the collections
... but could be valuable to capture the construction expressions
... valuable to say "here's an expression which provides a value set"
... intensional example "the set of 3 terms that are captured by X"

ericP: could you make put the construction into an opaque URL and just have refs to the value sets?

claude: yes but the endpoitns that can give you the whole collection aren't available today
... nice to give the rules for creating the value set
... they aren't intersting to the user but they are to a terminologist

aacc is charlie

TA modeling