Comments

• 

  Navah Stein October 21, 2019 at 1:16pm

  Working session notes 10/18/19

  Action Items:

  • Identify and reach out to additional experts
  • Connect with Jeff Jonas about MPIs  https://senzing.com/ (Paul Wormeli- Connection)
  • MiHin leaders
  • Connect with Dan Chavez about patient matching/proprietary 
  • Choicemaker (open-source matching product) - Contact Rick Hall, rphall@choicemaker.com
  • Start discussion on hub for topic of Patient Matching POC
    • identify/share resources, experts etc. 

  Action items from 10/11 that still need attention:

  • Note the mention of Sequoia's RCE kickoff call this week
  • Need common definition of Interoperability 

  Updates: 

  • Michigan Health Information Network- dominant standards, will join call starting 10/25
  • Workgroup update coming soon 
    • IT industry + cross-domain information-sharing 

  Working Session: 

  • Catalog of services
  • Who are the users of services?
  • What’s missing:
    • Data scientist (local or federal level) - looking at from warehouse perspective for data coordination, real time
    • Genome data sets
    • Washington State:
      •  Department of Health - Pop’n health analysis
      • Department  of Social Services - building MPI 360 view of a citizen
  • Patient Matching Project:
    • Cross domains and logical next steps
      • Use case with basic building blocks in healthcare, education, housing
      • Added social services and juvenile justice
      • There is not just one approach
      • Exploratory stage continues
    • What about undocumented/non-citizens?
    • "unite" may be confused with unite.us, also in this space https://www.uniteus.com/ NOTE: Our project is UNIFY
    • MPI request response (multi-factor matching)
      • Comes back with possible matches
      • Confidence interval 
      • Protects against identity theft 
      • Probabilistic matching
      • National/global ID use as problem
      • MiHin token use- bundling of separate sets of info that belongs to same person (pseudo-identifier bundling)
      • Face recognition - many cities prohibiting
      • When to have algorithms be public?
      • Terms can be domain specific using metrics like- Precision and Recall
      • Match to highest level, instead of lowest
    • APIs
      • Everything is proprietary 
      • Exception of NY state, which is open
        • Tom Silvius
        • Ivy Pool
      • https://api.hslynk.com human services open API with match retrieval
        • https://en.m.wikipedia.org/wiki/Pluggable_authentication_module
      • FHIR - no patient matching 
      • Service Discovery concept 
      • Open source matching product mentioned was Choicemaker. Contact is Rick Hall, rphall@choicemaker.com
      • Choicemaker - used in NYC immunization registry 
        • https://oscmt.sourceforge.io/
  • How to structure POC LGT meetings moving forward
    • Agree as group on action items of POC
    • What to include, what to exclude 
    • Determine core team and “observers”

  

  Senzing - The First Real Time, Plug n Play AI for Entity Resolution

  Senzing delivers the first AI product for entity resolution to find who is who in data. Affordable, plug-and-play, and available as a desktop app or…