Persisting FHIR Resources

From HL7Wiki
Jump to navigation Jump to search

Development of this AID whitepaper is discontinued. See FHIR Server Design.

---


Lead author/editor: Andy

Summary

  • From discussions during the September 2012 WGM
    • There are three approaches that have already been taken for persisting FHIR resources.
      1. Just each resource by itself with links
      2. Enough of a them linked together to be equivalent to a clinical statement or what we previously referred to as a SMIRF
      3. Entire CDA like documents that contain many clinical statements within them.
    • The experiences of persisting FHIR resources in any of the “clusters” above in the various “NoSQL” databases such as MongoDB (Ewout), CouchDB (Gordon), ?? (PHI), XML (Van Der Zel).

(potential) authors/contributors

  • Eve Marcelina Rubillos emfrubillos@gmail.com
    • I would want to volunteer in writing or creating this whitepaper. I am quite new to FHIR, it has only been a little over a month since I have been joining conference calls whenever my time permits, thus, I would prefer working with a team.
  • david hay david.hay25@gmail.com
    • I'm happy to be a reviewer / contributor - especially in the noSql space - but there are better qualified lead authors out there..
  • alex.zupan@tbsgroup.com alex.zupan@tbsgroup.com
    • Alex can provide info on our O-R (Hibernate) approach, starting from the class diagram down to the physical db model (tested on Oracle and MySql).
  • Charles McCay charlie@ramseysystems.co.uk
    • I may be able to offer some example content from the semantichealthnet project
  • George de la Torre delatorre.george@gmail.com
    • Something is missing here, "event sourcing", we should be open to any strategies, the white paper should demonstrate the benefits of options. Also, familiar with the other persistence methods, ORM, NoSQL, Graph, JDBC, SQL, and even Mumps (Cache) :)
    • Certainly, at least, I can contribute to writing up the "event sourcing" section, with FHIR as regard to RIM (missing here).
  • Jan Wittenber jan.wittenber@philips.com
    • Comment: There's a data type in ISO/IEEE 11073, 'Persistant Metric Stoe (PMS)', that might be of interest.
  • Nikolay Ryzhikov niquola@gmail.com