"Eyes on FHIR" - project introduction & objectives

Backdrop

• FHIR's popularity and maturity is accelerating as a community driven and implementer-focused modern flagship of healthcare standards to support interoperable healthcare data exchange. 
• This is in part due to the 2020 ONC “Final Rule” to the U.S. 21st Century Cures Act, which legislated that health technology vendor commmunity must support 2 open APIs that enable access to health data, which should follow HL7® FHIR® standard (see graphic below for associated timelines) [1]
  • 1: SMART on FHIR® API - primarily provides access to data on individual patients
  • 2: SMART/HL7 FHIR Bulk Data Access enables access to patient-level data across a population[2]

• In parallel, diverse organizations and projects have endorsed and developed strategies to support FHIR-based activity, such as:
  • US federal healthcare agencies - 
    • NIH
    • FDA
    • CMS / medicare / medicaid
    • ONC USCDI (United States Core Data for Interoperability); this defines the minimum data standards required to be supported
  • Technology giants; developed cloud-based FHIR APIs
    • Microsoft
    • Amazon
    • Google
    • Apple
  • HL7 - sponsored accelerator programs: designed to assist communities and collaborative groups across the global health care spectrum in the creation and adoption of high quality FHIR Implementation Guides or other standard artifacts to move toward the realization of global health data interoperability - 
    • Argonaut
    • The CARIN Alliance: Various healthcare stakeholders working collaboratively with government to overcome barriers in advancing consumer-directed exchange across the U.S
    • CodeX (Common Oncology Data Elements eXtensions) - building a community to accelerate interoperable data modeling and applications leading to step-change improvements in cancer patient care and research.
    • Da Vinci: Goal - to help payers and providers to positively impact clinical, quality, cost and care management outcomes.
    • Gravity: Goal - to create and maintain a consensus-building community to expand available SDOH core data for interoperability and accelerate standards-based information exchange by using HL7® FHIR®.
    • Vulcan: Vulcan will bring together stakeholders across the translational and clinical research community in order to bridge existing gaps between clinical care and clinical research, strategically connect industry collaboratives, maximize collective resources, and deliver integrated tools and resources

"Eyes on FHIR" - purpose

• HOWEVER - real-world applicability to clinical specialties currently has limited utility; this is due to a core problem – there is no universal implementation guidance (IG), which is required to enable truly systemic interoperability.
• HENCE -  the "Eyen on FHIR" overarching mission is to develop the necessary international standards to bring FHIR's value to enhance the quality and outcomes of ophthalmic healthcare delivery. 
• To do this, "Eyes on FHIR" aims to represent the comprehensive ophthalmic clinical lexicon in FHIR ® format.
  • This will follow routine HL7 democratic procedures requires multidisciplinary and broad stakeholder engagement and active participation
• Process:

1. Document a series of use cases that define real-world interoperability problems in ophthalmology in the traditional semi-structured "use-case" standards development format
   • The format should consider:
     • Actors
     • Pre-condition
     • Workflow
     • Post-condition
     • Exceptions (unique scenarios that, for example, may not follow routine workflow patterns)
     • Data elements
       • Regional differences and the associated implications for terminology use and application 
     • Necessary commentary to provide addition clinical and/or technical guidance for either the whole scenario or components of it
   • As FHIR is intended to support sharing data in a computable manner, determine the data elements, vocabularies and technical artefacts etc. requried to represent these scenarios, eg:
     • FHIR Resources 
     • Clinical terminologies, eg: SNOMED-CT, LOINC, ICD, DICOM, CPT, RxNorm (also noting regional variation)
   • Identify any scenario-relevant gaps and / or limitations in:
     • Clinial terminiologies
       • Eg - current SNOMED-CT international contains codes for the majority (but not all) of intraocular pressure measurement devices
     • Base FHIR profiles (which may require modification / extension etc.)
2. Collate and document the above information, including code systems, individual codes etc.
3. Convert this information into FHIR Profiles, configured as required
4. Develop an Implementation Guide (IG) - the set of rules of how a particular interoperability or standards problem is solved through the use of FHIR resources. 
   
   • Include relevant/necessary associated content and commentary to support and clarify the usage (from both a clinical and technical standpoint)
   • This should consider and account for relevant usage rules that pertain to a specific artefact or scenario, such as: required resource elements, constraints, binding strengths, cardinality, structural usage Information such as "Slicing", mappings to other specifications etc
5. Publish the implentation guide (see here for the "Eyes of FHIR" IG), which is published by HL7 International and the Patient Care WG (the github repository can be found here).
6. Participate in sequential tri-annual technical connectathons (in partnership with real world implementers) to demonstrate use cases, collaborate with the wider FHIR community, test and validate the IG technical integrity, real-world applicability and clincal utility, as well as identify problems and aim to solve them.
   • Incirporate learnings into IG iteration 
7. Participate in sequential tri-annual HL7-FHIR ballots to formalize the IG publication and community acceptance of the IG, ensuring openness and fairness of approval processes
   • Iteratively enhance and broaden the utility and 
   • Respond to 
8. Translate validated IG FHIR material into real world implementations to actually solve the problems/use cases that informed their development
   • Providers and implementers to collaborate in translationg proof concept demonstrations in connectathons to pilots (eg - in a real world setting test environmnet) through to integration into routine care

In parallel - it is imperative to build strong and broad community stakeholder engagement and collaboration to advance the content relevance and technical proficiency for implementers and drive adoption by ensuring clinical relavance, accuracy as well as implementation feasability and desirability. These include:

• FHIR community - expertise, online forums, synergistic IGs and accelerator programs
• Early and ongoing collaboration with the implementing vendor community
• Clinicians, service proving institutiont (eg - hospitals, practices) and peak clinical bodies (societies, colleges, academies etc.)
• Academic researchers and research organizations
• Life science industry
• Relevant technology enterprises (eg - large cloud providers, FHIR server / integration services etc.)
• Federal agencies (eg - ONC/ USCDI, CMS, FDA, NIH etc)

Desired Outcome

• "Eyes on FHIR" project
  • Ensure sustainability
  • Establish governance, quality control and ethical principles
  • Form committees with relevant and diverse representation from the entire ophthalmic / healthcare / technology ecosystem to oversee development, roadmap, priorities, 
  • Maintain structured administrative and communication process
    • Meetings
    • Committees
    • HL7 administrative and milestone requirements
  • Retain community-driven participation and other core values that have enabled FHIR's success in engagement and adoption
• Cinical and research impact
  • Serve and benefit all stakeholders in the ophthalmic community (ophthalmologists, optometrists, general practitioners, other specialists, patients, life science industry etc.) to enhance the experience, quality and outcomes of eyecare healthcare delivery.
  • Mobilization of healthcare data and enablements of its interoperable API-driven communication throughout the healthcare system
  • As the first clinical specialty endeavoring to undertake this effort - encourage and enable others to do so by:
    • Demonstrating clinical etc. benefits
    • Archive materials to be repurposable (technical, protocols/procedures, engagement processes etc.)

2021 activity summary

• Use case development
• IG development - linked here
• May '21 connectathon participartion - see output report (linked in IG) & below for more details
• Submission of first IG for ballot building of connectathon success
• Planned RW pilot implementations at US AMCs with vendor engagement following ballot (pending outcome)

May 2021 FHIR Connectathon, summarized:

• This was a pivotal demonstration of numerour world first acommplishment solving real world use-case driven data exchange problems in ophthalmology
• With leading vendors (see below and report) the IG profiles were validated, and showed how their utilization could enable FHIR - mediated bidirectional sementically interable messaging between EHRs/EMRs and ophthalmic diagnostic instruments/PACS
• This undertaking and outcome involved lengthy collaborative use case/IG authership and specific technical connectathon preparation from the following vendors and the dedication, work and inpur from their representatives:
  • Diagnostic device manufacturers:
    • Heidelberg Engineering
      • Alexander Gogol
    • Zeiss
      • Regis Deshayes
      • Marita Bergman
  • EHRs / EMRs
    • Oculo 
      • Warren Oliver (connectathon technical lead)
      • Ash Ramsay
    • Medisoft
      • Leah King
      • Adam Child 

Project formal HL7 engagements and reviews / progress since inception

• PSS (link) submitted 
• PC WG (29/9) - approved and commited to be the primary project sponsor (by vote in weekly meeting)
• EHR WG (30/9) - approved and commited to co-sponsorship (by vote in weekly meeting)
• FMG reviewed (15/10) - approved
• CSQ - approval - no ongoing involvement required
  • Recommended review of retinal exam: CMS 142 (available at https://ecqi.healthit.gov/). Although the modelling is in the CMS model QDM but there are parallels to FHIR and a mapping from QDM to FHIR in the QI-Core IG http://hl7.org/fhir/us/qicore/qdm-to-qicore.html.  
• CSD approved
• TSC - no issues noted
• May '21 connectathon participation; May 2021 Ophthalmology track prep link
• Report out (here); IG amended addressing issues identified
• NiB sumbitted for Sepetmber '21 ballot 
• IG reviewed by PCWG (July 29 meeting) - provided feedback, subsequently incorporated into into IG
• Migrated contant to HL7 international github repository 
• Ahead of final content submission to FMG priot to community for ballot;
  • QA/QC, content revision/update and technical issues addressed
• IG submitted (by S. Chu - PCWG co-chair) to FMG by August 8 deadline for review
• IG - "HL7 FHIR® Implementation Guide: Ophthalmology Retinal, Release 1 - proceeds to ballot

Process for project → IG balloting

See HL& calendar here and FHIR specific calendar here

See confluence link here for details

Upcoming milestones

HL7 calendar

Prep for Sept '21 ballot

Note - Sept '21 ballot link here

Next:

• Build off May connectathon activity:
  • 1- 2x RW pilot implementation at large US AMCs (panding ballot seccesss)
• September connectathon
  • Intention to have collaborative working sessions with:
    • Gravity project
    • FAIR4FHIR
    • Imaging Integration
  • Building off May connectathon activity success - intention is to focus of end-to-end workflows that required to implement ophthalmic AI, which combines clincal, imaging and other critical data elements (eg SDoH - Gravity Project) foor the development and deployment ofimageing-based AI predictions.
• Jan '21 ballot
  • 22 sept - PSS TSC deadline (24th appeal)
  • 31 OCT proposals approved
  • 31 OCT - IG NiB deadline
• Jan '21 connectathon:
  • 16 NOV connectathon proposals due

Use cases

Example categories:

#1 - Routine clinical care communication; the patient journey & connecting care

• eg - longitudinal patient journey / collaborative care / asynchronous communication / remote devices

#2 Physician clinic tools; SMART apps (here) / CDS hooks (here) / integrating remote monitoring

• eg - amalgamating information for consolidated visualization and increased speed on decision making +/- decision support +/- links to resources

#3 Registry data collection; EHR > registry (minimize redundant duplication & increase population

• eg - AAO's IRIS /  Save Sight Institute's FRB! (Fight retinal blindness) international registry

  • See SMART's bulk data API
  • Nature's publication: "Push Button Population Health: The SMART/HL7 FHIR Bulk Data Access Application Programming Interface" - open-access link

#4 Clinical workflow optimization

• eg - Automating payments / prior authorization / audits / administration / patient flow

  • CMS

#5 Facilitating life Science & research 

• eg - facilitating data harmonization / real-time trial recruitment

• See: 
  • NIH "Fast Healthcare Interoperability Resources® Initiatives" - here
    • Announcement: (here) "to encourage NIH researchers to explore the use of the Fast Healthcare Interoperability Resources (FHIR®) standard to capture, integrate, and exchange clinical data for research purposes and to enhance capabilities to share research data."
  • Vulcan accelerator program - link to confluence - "Vulcan brings together stakeholders across the translational and clinical research community in order to bridge existing gaps between clinical care and clinical research, strategically connect industry collaboratives, maximize collective resources, and deliver integrated tools and resources."

Use cases

• Topic outline tabulated below
• See link to gdoc for approach and detailed material

Note: additional use case

• AI / imaging workflow (DICOMweb + FHIR)
  • 1 - capturer needs to be able to trigger the AI model
  • 2 - how to retrieve data they need to run model (from EHR)
  • 3 - how to report back into EHR / PACS for results and downsteam workflow triggers
  • See workflow diagram below or linked here

AI use case - workflow #2:

Enabling best practices for generating “AI/ML-ready” FAIR datasets minimizing bias

and confronting generalizability: a proposed workflow combining FHIR ® and DICOM ® synergy.

• FHIR & DICOM enabled workflow incorporating data elements from other FHIR IGs

Large thanks to the HL7-FHIR WGs and community who expertise and guidance contributed towards this endeavour:

• Stephen Chu - much appreciation - co-chair of HL7 international Patient Care WG
• Rob Hausum - co- chair of Vocabulary and Orders and Observations (OO) Work Groups
• Imaging Integration // DICOM WG 20 - provided domain specific guidance on enabling FHIR and DICOM to intersect with clinical and imaging workflows respectively.
  • Co-Chairs, Imaging Integration Working Group, HL7 International; Co-Chair, WG-20 Integration of Information Systems, DICOM)
  • Jonathan Whitby , Vital, a Canon Group Company
  • Christopher Lindop, GE Healthcare
• Also to Elliot Silver
• Grahame Grieve - overarching navigational guidance and support

Endorement and support is deeply appreciated from:

• NEI (National Eye Institute), NIH, USA
  • Michael Chiang, MD, MA (BMI - bioinformatics)
  • Kerry Goetz, Associate Director, NEI Office of Data Science and Health Informatics
• Save Sight Institute, Sydney Australia
• Ophthalmic Vendor community collaborators - noted specifically and individually in the May '21 connectathon output report; these ongoing colllaborations are representative of FHIR's translational capacity 

References,  terminologies & toooling:

• Vocabularies

  • ICD - 10 / ICD - 9 (ICD - International Statistical Classification of Diseases and Related Health Problems, Whorld health organization)
  • SNOMED-CT 
  • LOINC
  • CPT (Clinical Procededual terminology - Copyright 2014 American Medical Association. All rights Reserved.)
  • Rxnorm

  • HCPCS (Healthcare Common Procedure Coding System)

  • Non US-based
    

    • Australia

      • MBS (mdeicare benefits schedule) - use for procedures

      • PBS (pharmaceutical benefits scheme)

    • UK

      • Note: UK Core -

        • i. As of June 28 2021 “UK Core Profiles (are) in development and should be considered as "work in progress"”

      • DMD (Dictionary of medicines and devices)

      • MIMS - “the essential prescribing and clinical reference for UK general practice.”

    • 
      

• Tools

  • shrimp ontoserver (SNOMED browser)

  • iMagic mapper link

  • UMLS meta-thesaurus link

• IHE reference documents folder

  • IHE General Eye Exam (GEE)

  • IHE Eye Care Technical Framework Key Measurements in DICOM® Encapsulated PDF Revision 1.1 – Trial Implementation April 29, 2019 

• DICOM ophthalmology conformance statements (from WG-09)

  • Folder with documents

Community awareness / involvement

• Publications
  • 2021 ARVO abstract - accepted for 2021 conference (link - FHIR arvo abstract.pdf)
    • Poster link

Contacts

Project lead:

• Dr. Ashley Kras, MBBS (BMI - bioinformatics), FRANZCO
  • Consultant ophthalmologist
  • Senior Clinical Lecturer, Sydney University
  • Postdoctoral fellow, Massachusetts Eye and Ear Infirmary Retinal Imaging Lab, Harvard Medical School, MA, USA
  • Honorary fellow, Moorfields Eye Hospital

Clinical FHIR Liaison 

• Stephen Chu (Patient Care Working Group co-chair, HL7 international)

Technical leads:

• Brett Esler (Oridashi)
  • Board director, HL7 Australia
• Warren Oliver (CTO, Oculo)

Primary Clinical Advisors:

• Mark Gillies, MBBS, MD., Ph.D, FRANZCO
  • Head, Medical Retina Unit, Sydney Eye Hospital
  • Professor, Director of Research, University of Sydney, Save Sight Institute
  • Founder and director - Fight Retinal Blindness! project (Save Sight Rigistries), Save Sight Institute, Sydney University
  • Chair, working group for the International Consortium for Health Outcomes Measurement (ICHOM) for macular degeneration (2014-2016)
    • Rodrigues IA, Sprinkhuizen SM, Barthelmes D, et al. Defining a Minimum Set of Standardized Patient-centered Outcome Measures for Macular Degeneration. Am J Ophthalmol. 2016;168:1-12.
• Thomas Hwang, MD
  • Chair, Medical Informatics and Technology Committee, American Academy of Ophthalmology
  • Professor of Ophthalmology, Casey Eye Institute-Oregon Health and Science University
  • Residency Program Director
  • Chief, Vitreoretinal Division
• John Miller, MD 
• Assistant Professor of Ophthalmology, retina faculty, Massachusetts Eye and Ear Infirmary - Harvard Medical School (HMS) affiliate medical centre
• Director, Retinal Imaging Laboratory, HMS
• Associate Director, HMS Vitreoretinal Fellowship
• HMS Ophthalmology Residency Retina Curriculum Advisor
• Nigel Morlet, MBBS, FRANZCO, FRACS
  • Professor clinical epidemiology, School of Population and Global Health University of Western Australia
  • Consortium member in the development of ICHOM for cataracts Standards Set
• Hemal Mehta, MBBS MD (Cantab.), FRCOphth, FRANZCO 
  • Advisor to RCOphth National Age-related macular degeneration Audit (UK)
  • Senior researcher and steering committee member, Fight Retinal Blindness!, Save Sight Institute, Sydney University
  • Consultant ophthalmologist, Sydney Eye Hospital
• Mike Mair, MBBS
• Pearase Keane, MBBS
  • Professor of Artificial Medical Intelligence at UCL Institute of Ophthalmology (IoO)
  • Honorary Consultant Ophthalmologist at Moorfields Eye Hospital NHS Foundation Trust

Technical Advisors

Tobias Elze (Schepens Schepens Eye Research Institute of Mass. Eye and Ear; Harvard Ophthalmology AI Lab)

Marco Garcia (Save Sight Registy)

Contact

Please contact ashleylkras@gmail.com for any feedback or questions related to the IG development or this project in general.