Today's Electronic Health Record (EHR) offerings inhibit the ability to develop the next generation of clinical applications for providers to provide the best possible care and patients to become more engaged. Such offerings are designed as monolithic silos of storage and end-user experience that use proprietary methods for accessing key functionality. While large health care providers typically have physical control over their data, they lack in having functional control. This barrier makes it virtually impossible for data to be easily accessed by other applications without very costly and time-consuming migration strategies. As a result, the pace of innovation is greatly stymied by closed systems that appear to be all-too-prevalent in the healthcare industry. This session presents a strategy of vendor-neutral, public, open Application Programming Interfaces (APIs) and advocates for their use in developing open platforms for healthcare applications.
6. HIMSS MN 2014-2015 Event Calendar
Month Format
August 21st Webinar – Open Platforms for Healthcare Applications
September Webinar
October Expert Panel
November Midwest Fall Technology Conference (Chicago)
December Webinar
January Advocacy Event
February Webinar
March Networking Event
April HIMSS National Chapter Reception (Chicago)
May 14th Annual Spring Conference
June 3rd Annual Charity Golf Tournament
June Twin Cities Summer Social
July SE Minnesota Summer Social
August Webinar
Dates and topics TBD
Please take HIMSS MN survey to guide your chapter on topic selection!
Opportunity to win $100 Amazon gift card.
7. Today’s Speakers
Keith Toussaint
Executive Director
Collaborative Development
Mayo Clinic
Ken Bobis
Administrator
Chief Technology Office
Mayo Clinic
8. An Open Interface Approach
Presented to HIMSS MN
August 21, 2014
Keith M. Toussaint
Ken G. Bobis, PhD
9. Definitions
The problem
Solution vision
Business value
An example
Further research
10. Health Care Organization (HCO)
EMR Application
Interface / API / Messaging Standard
Ecosystem
Open Interface
11. Application that is integral to the
delivery of patient care, but not found
in the institution’s core Electronic
Medical Record
◦ Document Viewer, EMR Viewer, Pt.
Itinerary, smartplatforms.org
12. Interface
◦ A point where two systems, subjects, organizations, etc.,
meet and interact
Tape data transfer; interface engine
API
◦ Application Program (API) or Web Services (WS) Interface
A callable program interface
ValidatePatientNumber, ValidateDate
Messaging Standard
◦ Rules by which an interface is maintained
◦ An agreed upon message format
HL7, FHIR
13. (biological) A biological community of
interacting organisms and their physical
environment
◦ Rain Forest, Prairie, Coral Reef
(software) A set of businesses functioning
as a unit and interacting with a shared
market for software and services, together
with relationships among them
◦ Wintel, iOS, Android, Java
17. open (small ‘o’)
◦ May be used by others
◦ Control retained by the author
◦ Traditional view of “open”
Open (big ‘O’)
◦ Author-independent
◦ Emerging view of “Open”
Open Interface <> Open Source
18. Practice functionality needs are huge
◦ More than a single provider organization can satisfy
◦ Years to implement workflows in existing products
◦ New capabilities removed from “The Gemba”
Closed architectures
◦ No access to data sources and business logic
Product lock-in
◦ Product B may be more conducive but … switching
costs are prohibitively high
19. User Interface
Core Clinical
Informatics
Functionality
Data Storage
and
Computing
Infrastructure
EHR X
Monolith Y
Viewer 1
Tool 1
Viewer 2
Tool 2
Legend
:
Vendor provided, HCO Managed HCO provided and managed
20. Embrace Open APIs
◦ Author-neutral interfaces
◦ Reduce product lock-in
Layered architecture
◦ Separating application, platform and storage
functionality
New Value Network
◦ New problems require new solutions
◦ New business models with new value network
participants
Common infrastructure
◦ Available to all – but not required
21. User Interface
Open interfaces
Core Clinical
Informatics
Functionality
Open interfaces
Data Storage
and
Computing
Infrastructure
EHR
PoC Tool 1
PoC Tool 2 PoC Tool 3
HCO 1 Vendor 3
Vendor 1
Vendor 2
Legend Vendor provided and managed HCO provided and managed
23. Leverage common core capabilities
Minimize capital expenditures
Enable rich collaboration
Focus on differentiating technology
Reduce vendor “lock-in”
24. Common platform to meet MU
Establish common application frameworks
Improved Innovation on reporting and
analytics
Cost savings through shared physical
infrastructure
26. Enable new value network of
application innovation
◦ Shared core functionality
◦ New capabilities on existing ‘stack’
◦ Enabling consistent tools across the spectrum
of care
Reduce innovation friction
29. HL7 Fast Health Interoperability Resources
Fast and easy to implement
Implementation libraries
Free for use
Interoperability out-of-the-box
Path from HL7 Version 2 and CDA
Web standards compatible
Support for ReSTful architectures
Human-readable message formats
http://www.hl7.org/implement/standards/fhir/summary.html
30. Is it big “O” open? Yes!
◦ Non-proprietary standards
◦ Enables local extension of the standard
◦ Well-defined process for evolving the standard
What is missing?
◦ Definition of layered architecture
◦ Clear description of “separation of concerns”
◦ Example of such an architecture:
A Robust Health Data Infrastructure prepared by
JASON for ONC:
http://healthit.gov/sites/default/files/ptp13-
700hhs_white.pdf
31. User Interface
Core Clinical
Informatics
Functionality
Data Storage
and Computing
Infrastructure
33. Cloud-based deployment architecture
Catalog of required of services
Flesh out business value
Operational policies & procedures for
member HCOs
Protocols for EMR & Ecosystem
interactions
Common local & public architecture
requirements
34. Current State is sub-optimal
Loosely-coupled architecture
Open APIs
◦ Understanding the difference between small ‘o’
and big ‘O’
Now possible to make a shift
Can be driven by providers
Opinions invited on how to proceed
35. Today’s Speakers
Keith Toussaint
Executive Director
Collaborative Development
Mayo Clinic
toussaint.keith@mayo.edu
Ken Bobis
Administrator
Chief Technology Office
Mayo Clinic
bobis.kenneth@mayo.edu
36. An Open Interface Approach
Presented to HIMSS MN
August 21, 2014
Keith M. Toussaint
Ken G. Bobis ,PhD
Notas do Editor
Ken
Ken
Ken
Be sure to highlight the sub-bulleted when defining …
Interface / API/ Protocols
FHIR
ReST
KMT: Took out the “Open Innovation” bullet point. Concerned that it added too much to the discussion. Open Innovation is a positive, but not a *requirement* for an ecosystem (Apple appstore ecosystem is a good example of a viable ecosystem that is *not* an Open Innovation environment
Ken
Smartplatforms.org
FHIR used in integrate to Cerner and Epic, Seen in HIMSS - 2014 Interoperability showcase, read-only
Ken
Keith
Definitions:
Biological Ecosystem:
http://www.oxforddictionaries.com/us/definition/american_english/ecosystem
Software Ecosystems:
David G. Messerschmitt and Clemens Szyperski (2003). Software Ecosystem: Understanding an Indispensable Technology and Industry. Cambridge, MA, USA: MIT Press. ISBN 0-262-13432-2.
Keith
Definitions:
Biological Ecosystem:
http://www.oxforddictionaries.com/us/definition/american_english/ecosystem
Software Ecosystems:
David G. Messerschmitt and Clemens Szyperski (2003). Software Ecosystem: Understanding an Indispensable Technology and Industry. Cambridge, MA, USA: MIT Press. ISBN 0-262-13432-2.
Keith
Definitions:
Biological Ecosystem:
http://www.oxforddictionaries.com/us/definition/american_english/ecosystem
Software Ecosystems:
David G. Messerschmitt and Clemens Szyperski (2003). Software Ecosystem: Understanding an Indispensable Technology and Industry. Cambridge, MA, USA: MIT Press. ISBN 0-262-13432-2.
Keith
Definitions:
Biological Ecosystem:
http://www.oxforddictionaries.com/us/definition/american_english/ecosystem
Software Ecosystems:
David G. Messerschmitt and Clemens Szyperski (2003). Software Ecosystem: Understanding an Indispensable Technology and Industry. Cambridge, MA, USA: MIT Press. ISBN 0-262-13432-2.
Keith: find document that spells out distinctions between open interfaces/source/formats/protocols
There are various definitions of open that might be employed here
Open API
Not to be confused/conflated with “Published” API. For example if you go to open.epic.com, you will see what Epic defines as an ‘open’ API. By our definition, this in no way meets the requirement of Open API. An open API is explicitly not owned by any one entity. It is freely available to use (like OpenStack)
Open Source
Open Format
For the purposes of our endeavors in Mayo Clinic IT, our definition of Open should be centered on Open APIs.
https://collab.mayo.edu/team/TSAT/Architecture%20Working%20Group/Working%20Documents/Clean%20and%20Open/Defining%20Open%20and%20Clean.docx
Ken
Ken
The current state of EHR systems and clinical software in general is characterized by the ‘monolithic silo’ of data storage, core clinical informatics functionality and user interface as depicted above.
At the time that the mature EHR systems were developed, this approach was considered state-of-the-art. However, there are numerous disadvantages to this approach, among them being vendor ‘lock-in’ and significant impediments to innovations based on rapidly evolving clinical needs.
In the intervening years, however, modern computing architecture has emerged that is predicated on layered architecture employing standard computing interfaces between those layers.
Eric Schmidt translation:
The current state of EHR systems and clinical software in general is characterized by the ‘monolithic silos’ of homogeneous data storage (living in non-virtual compute infrastructure) proprietary interfaces to core clinical informatics functionality and sub-optimal user interfaces designed not by the those delivering care, but by the infrastructure/technology companies.
At the time that the mature EHR systems were developed, this approach was considered state-of-the-art. However, there are numerous disadvantages to this approach, among them being vendor ‘lock-in’ through opaque architecture and proprietary interfaces and significant impediments to user interface innovations based on rapidly evolving clinical needs.
In the intervening years, however, modern computing architecture has emerged that is predicated on REST principles that call for layered, stateless architecture employing standard computing interfaces between architectural layers. Further, the ability to leverage the emerging Business of APIs requires adherence to these architectural principles.
Ken:Describe the term “new value network”
Keith
The current state of EHR systems and clinical software in general is characterized by the ‘monolithic silo’ of data storage, core clinical informatics functionality and user interface as depicted above.
At the time that the mature EHR systems were developed, this approach was considered state-of-the-art. However, there are numerous disadvantages to this approach, among them being vendor ‘lock-in’ and significant impediments to innovations based on rapidly evolving clinical needs.
In the intervening years, however, modern computing architecture has emerged that is predicated on layered architecture employing standard computing interfaces between those layers.
Eric Schmidt translation:
The current state of EHR systems and clinical software in general is characterized by the ‘monolithic silos’ of homogeneous data storage (living in non-virtual compute infrastructure) proprietary interfaces to core clinical informatics functionality and sub-optimal user interfaces designed not by the those delivering care, but by the infrastructure/technology companies.
At the time that the mature EHR systems were developed, this approach was considered state-of-the-art. However, there are numerous disadvantages to this approach, among them being vendor ‘lock-in’ through opaque architecture and proprietary interfaces and significant impediments to user interface innovations based on rapidly evolving clinical needs.
In the intervening years, however, modern computing architecture has emerged that is predicated on REST principles that call for layered, stateless architecture employing standard computing interfaces between architectural layers. Further, the ability to leverage the emerging Business of APIs requires adherence to these architectural principles.
Ken
Ken
Ken
Ken
Ken
Keith
Keith (AR): Go into some detail about FHIR (separate slide?), incorporate ReST into the FHIR discussion
HealthKit: Note that it is not an open standard, Walled Garden
Samsung:
… other new ones
Examples:
Providers could develop or purchase at limited cost solutions that meet their clinical needs
They about your own case of what applications you are still waiting for
Keith
Keith (AR): re-do this slide (and previous?)
FHIR does not specify how the layers are fit together
Keith
Does this look familiar (show previous architecture)
Keith
Keith/Ken
Ken/Keith
Current State is sub-optimal for doing all we need to do meet the needs of the health care system of the future
Disparate departments
Population health needs
Complex revenue cycle
Complex integration between organizations
Need to move to
Now possible? Why?
Rest architecture proven at scale
Driven by providers?
Technology industry has been reluctant to drive major change
HCOs have the best incentives to drive change
Time is now to move forward?
Crossroads in health care
Replacing emr shift is happening (1st to 2nd generation)
Time to rearchitect how we do things
Small percentage of automation done today …
adapt or die (agility)
How do we do it?
Consortium
Open marketplace
Need to hear your views