15:03:23 <RRSAgent> RRSAgent has joined #hcls
15:03:23 <RRSAgent> logging to https://www.w3.org/2018/10/30-hcls-irc
15:03:28 <dbooth> rrsagent, make logs public
15:03:36 <dbooth> Meeting: FHIR/RDF
15:03:40 <dbooth> Chair: David Booth
15:05:32 <hsolbrig> hsolbrig has joined #hcls
15:48:17 <dbooth> Topic: FHIR/RDF
15:49:14 <dbooth> harold: Big thing right now is graph DBs and knowledge graphs.  Conjecture: RDF could be viewed as the bridge between FHIR and these knowledge graphs.
15:50:50 <dbooth> david: FHIR/RDF is top hit when searching google for fhir knowledge graph: https://www.google.com/search?q=fhir+knowledge+graph&sa=Search&sitesearch=
15:53:08 <hsolbrig> https://findwise.com/blog/tinker-knowledge-graphs/
15:55:03 <dbooth> harold: We need to become more knowledgeable about [other] graph DBs.
15:56:18 <dbooth> eric: What do people want to do with these knowledge graphs?
15:57:04 <dbooth> ... They strip down the data to the bits they care about.  E.g., clinical encounters at the same facility, therefore MRSA probably followed this vector.  Not a wholesale import to KB.
15:58:05 <dbooth> harold: They're using it for exactly what the Linked Data community does.
15:59:51 <dbooth> ... At the moment they're using OBO ontologies, as unifying factor.  They're either creating new OBO resources or mapping existing OBO resources to the new facts coming in.
16:01:19 <dbooth> ... They're looking at can this be transformed to disease ontology, LOINC code.  They want to map it into low/high measures that disease ont has (hypo/hyper).
16:04:53 <dbooth> david: FHIR/RDF is already stepping up a level from  bare RDF -- a domain-specific language based on RDF.
16:06:05 <dbooth> harold: FHIR/RDF is a first exmaple of RDF representation is available from the beginning.
16:06:40 <dbooth> david: Action items?
16:07:09 <dbooth> harold: 1. Possible Yosemite presentation.  2. positioning of RDF vs knowledge graphs.
16:07:47 <dbooth> s/positioning/relationship/