Stream: ontology
Topic: Work todo
Grahame Grieve (Nov 04 2016 at 06:15):
GF#11451 - @Eric Prud'hommeaux @David Booth - need to work on this
Grahame Grieve (Nov 26 2016 at 18:46):
and do we want a connectathon stream or some other alternative to get us to focus on some tangible outcome from the all this RDF work? @Michael van der Zel @Eric Prud'hommeaux @David Booth
Michael van der Zel (Nov 27 2016 at 11:22):
It would be excellent to have a RDF stream. We have to come up with a use-case and some servers that implement it (e.g. Spark).
Grahame Grieve (Nov 27 2016 at 11:23):
my server implements the turtle format now
Grahame Grieve (Nov 27 2016 at 11:23):
what we need is a use case
Grahame Grieve (Nov 27 2016 at 23:41):
@David Booth - where are we with GF#11450?
David Booth (Nov 28 2016 at 02:24):
Last week I got the FHIR build working on my laptop, and got the module renamed to Linked Data module, and committed the changes to svn. In order to re-write the page content to focus more on "how to", I need to be able to reference examples of usage, so I need to collect some from people.
David Booth (Nov 28 2016 at 02:24):
@Michael van der Zel , @Eric Prud'hommeaux, @Harold Solbrig and anyone else, I need examples of FHIR RDF usage that I can draw upon for re-writing the Linked Data module pages (formerly called the Ontology module) to have a more "how to" flavor. Can you point me to some, or provide me with some? Who else should I ask?
Grahame Grieve (Nov 28 2016 at 03:37):
This is an old example from pre-FHIR days: "...calculated the HEDIS Diabetes quality measure on a population of over 400,000 patients real data" using an OWL ontology.
David Booth (Dec 01 2016 at 22:19):
Thanks, I'll ask Peter Hendler about that.
Michael van der Zel (Dec 07 2016 at 08:28):
Here <https://vimeo.com/141875367> is his presentation about it: Use of a RIM-lite OWL triple store to calculate HEDIS quality measures (Peter Hendler, KP)
David Booth (Dec 07 2016 at 19:51):
Thanks @Rob Hausam and @Michael van der Zel . I have written to Peter. It is unclear how much he could have benefited from FHIR RDF. I think it would have simplified translation to the analysis model that he used, which was Entities and Roles Participating in Acts, but I have no data to prove that. I asked if the RIM mapping available in FHIR would have helped, and he didn't think so, because he only used a simple E-R-P-A model, not the full detail of RIM.
Grahame Grieve (Dec 07 2016 at 19:53):
right. much of what he did would be extremely straight forward in fhir, because it's just index joins pushed onto the FHIR servers, along with terminology subsumption. But it's still closest use case I've actually seen
Last updated: Apr 12 2022 at 19:14 UTC
