OHDSI Home | Forums | Wiki | Github

CDM V 6.0 To FHIR Mapping

cdm

(Umesh Yadav) #1

Can somebody provide me the source link for mapping between CDM v6.0 to FHIR?


(Frank DeFalco) #2

To my knowledge there is not a current ETL utility for CDM 6.0 to FHIR.


(Christian Reich) #3

Should be the other direction anyway, shouldn’t it? FHIR is the source.


(Umesh Yadav) #4

I think there is something avilable on hl7 website. Will that be not useful?


(Sebastiaan Van Sandijk) #5

I don’t think there is a OMOPv6.x to FHIR mapping available either. A complete mapping - full version to full version in whichever direction - is not readily available anyway: there always will be coverage for some resources/domains while other domains/resources have not been mapped (yet). Much depends on the available source data. That’s probably why cdm6.x has not been mapped to FHIR and why also FHIR R4 (current stable version) to OMOPv5.x is mostly work-in-progress rather than complete.
As to the direction of the mappings, there are some use cases for OMOP to FHIR mappings. GeorgiaTech ( OMOPonFHIR therefore works on bi-directional mappings. Why do you ask about CDM6 to FHIR mappings @UmeshClinakos?


(Sebastiaan Van Sandijk) #6

you mean the Common Data Models Harmonization, project?
On the HL7 FHIR website?


(Umesh Yadav) #7

yes.


(Umesh Yadav) #8

Because we want to convert our FHIR datamodel into CDM. So that we can do some analysis on atlas by using that. Will that make sense ? @Sebastiaan_van_Sandi


(Sebastiaan Van Sandijk) #9

Yes, it does. That is in line with Christian’s comment: your FHIR data are the source for an OMOP-ETL, apparently into CDMv6.0. I would say that you will need to build up your own mapping/ETL-design; however, I expect you can build on top of existing v5.3-FHIR mappings. Depending on where you are and the capabilities of your FHIr server (if you use one?) there may be interesting good practices for you.


(Sebastiaan Van Sandijk) #10

The CDMH project worked on mappings between different research-oriented CDMs via BRIDG as an intermediate. This is probably too complex a route for most practical use cases. Also, the CDMH IG is rather US oriented building its mappings on top of US Core- and DAF profiles. So I would not necessarily use these mappings as your foundation - depending on where you are, of course


t