How can ACMI best contribute to the widespread adoption of EHRs in the U.S.

1
How can ACMI best contribute to the widespread
adoption of EHRs in the U.S.?

• W. Ed Hammond, Ph.D.
• Duke University

2
What I would like to do today

• Share some ideas of what I think we are trying to
  do
• These ideas are suggested as starting points. We
  will not all agree, but perhaps we can come to a
  point of agreement
• Identify some of the issues
• Suggest some recommendations that ACMI need to
  publicize, support and promote
• We will want to discuss, debate and modify
• We need to know how we will promote the
  recommendations

3
The EHRs time has come

• After more than 35 years of pursuing a thing we
  now calling the Electronic Health Record, that
  goal has captured the worlds attention. We are
  now challenged to create standards and define a
  process to make it happen.
• We are discovering that the details of what it
  is are elusive and the boundaries are very fuzzy.
• What should be standardized and what should be
  proprietary is open to debate.
• There are more views of what we must specify to
  move forward than we can accommodate.

4
How many EHRs are there?

• Inpatient medical record
• Outpatient health record
• Primary care medical record
• Disease-specific record (cardiology,
  hypertension, diabetes, etc.)
• Intensive care record
• Emergency department record
• Nursing record
• Nursing home record
• Billing/claims record
• Research record

5
The EHR

• It is not a clinical repository.
• Its purpose is to enhance the health and enable
  the care of the individual. Its contents are
  solely justified for that purpose. When data
  ceases to contribute, it is removed.
• Much of the data in the inpatient setting has
  limited persistence - usually the more intense
  the care, the shorter the persistence.
• There are other repositories a data warehouse
  that does contain and retain everything.
• The EHR documents maintenance of care, diagnostic
  and treatment processes, health status.

6
Additionally

• There will be more than one type of EHR and the
  purpose of each type will be different.
• The site and purpose of care are importance
  features.
• The required functionality will be somewhat
  different for each type of record.
• The persistence of data will vary with site and
  purpose.
• The core components of all records must be the
  same in terms of the data elements and how they
  are represented.
• All types of records are interconnected.

7
IOM HL7

• Health Information and Data
• Results Management
• Order Entry/Management
• Decision Support
• Electronic Communications Connectivity
• Patient Support
• Administrative Process
• Reporting and Population Health Management

• Direct Care Health Information
• Work Flow and Operations Management
• Communications
• Records Documents and Views
• Clinical Support
• Measurement, Analysis, Research, Reports
• Administration/Finance

32 second level functional recommendations
60 second level functional specifications
8
The HL7 Ballot

• 221 persons actually voted the largest number
  of people voting on an HL7 standard ever
• HL7 charged 100 administrative fee for
  nonmembers to vote (negative reaction)
• Approximately 50.2 of the votes were negative.
  The largest number of negative votes came from
  the vender community.

9
Ballot Results
10
Feedback from voting

• Structure vs functional description
• Vague and ambiguous terms and concepts
• Wanted clear and explicit definitions
• Variations in granularity
• Free text vs templates
• Clinical vs infrastructure terms
• Want specific data elements to be defined
• Confusion on Essential/Desired Tags

11
From an other perspective

• Multiple views of the EHR must also exist
• The institutional/provider view that serves the
  need of the institution in patient care, service
  management, workflow management, and billing.
  This view provides the source of data for other
  views, since it is in this setting that the
  patient/provider encounter takes place.
• The composite view that represents a complete
  summary view (the patient-centric view) of the
  persons health. This view also serves needs of
  health surveillance, bioterrorism surveillance,
  and global epidemiology.
• The personal health view that is customized to
  each individual and their current and future
  health needs.

12
ELECTRONIC HEALTH RECORD
COMMON TERMINOLOGY
MASTER SET DATA ELEMENTS
RIM DATA TYPES
CMETS GPICS ARCHETYPES
TERMINOLOGYSERVER
DATA ELEMENTSERVER
TEMPLATES CDA
ELECTRONIC HEALTH RECORD
ARCHITECTURE STRUCTURE
FUNCTIONS
CONTENT
13
ExerciseLog
TravelLog
Device, SensorInput
AccessControlList
PersonalInput
DietLog
Specialists
Specialists
Specialists
PersonalHealthRecord
ChatRooms
ChatRooms
Clinics
Clinics
Clinics
Physiciane-mail
Inpatient
Inpatient
Inpatient
InternetWeb
InternetWeb
DecisionSupportAlgorithms
Pharmacy
Dental
Dental
Pharmacy
14
EHR Interoperability Diagram
Billing/Claims
Profile
EnterpriseData Warehouse
PersonalEHR
PersonalEHR
Profile
PatientEncounter
Provider EHRDatabase
Institution EHRDatabase
Disease Registry
Profile
LongitudinalEHR
ResearchDatabase
Profile
Profile
Profiles contain business rules
15
Download Process
Sensitive DemographicData
DoubleEncryptionSiliconEncoder
ID
ID
HL7 Message
Identifying Data, name, address, etc.
Encrypted ID
Identifying Data, Translated (e.g. Zip).
Aggregated SummaryLongitudinalEHR
Summary Data
16
Summary Longitudinal Record
Patient controlled access
Summary Longitudinal Record
Access log
Access list permitby provider group clinic othe
r
Feeds PH surveillance, patient
safety, epidemiology
17
Regional Linkages
18
State-wide Linkage
UHI Pointer
Population 12 M
19
National Linkage
N
UHI Pointer
20
Business Linkages
Hospital Intensive Care EmergencyDepartment
NursingHome
Ambulatory Care Clinic
21
What must be in place

• Data standards
• Reference Information Model
• Common data elements
• Common data types
• Common terminology
• Clinical templates
• Ability to share data and knowledge
• Data interchange standards
• Common content architecture standards
• Common minimum set of functions for the EHR
• Infrastructure to support required connectivity
• Common methods of knowledge representation

22
Shared repositories at national or international
levels

• Master set of data objects or elements
• Common, integrated terminology sets
• Registered clinical documents using a standard
  architecture
• Registered clinical templates
• Registered clinical guidelines using standard
  format
• Registered decision support algorithms in
  standard format
• Registered disease registries in standard format

23
Everybodys doing it

• United Kingdom (11B)
• Canada (1B)
• Australia (millions)
• New Zealand
• Germany
• Denmark
• Sweden
• Finland
• Korea
• Taiwan
• Japan
• US (Can we? A bigger challenge)
• Others

Goal is to have nation-wide clinical,
interoperable electronic health record systems by
the end of the first decade of the 21st century.
24
Issues

• Vendor community
• Dont know installed base
• Legacy and costs too much to change
• Capacity and scalability
• Confused as to market demands
• Cant stay state of the art
• Dont really understand the problem
• Providers
• Wont pay to go or stay state of the art
• Are not motivated to change
• terminology
• Dont know what to change to
• 100 per month for ambulatory system

25
Issues

• Open source
• Continuity of Care Record
• Too many choices
• Who will make the case?
• How do we hold to the course?
• Who is the trusted authority?
• Who provides the leadership?
• Who controls the data?
• Who owns the data?

26
Issues

• Take what we can get or try to achieve the best
  system we can conceive
• Transition sets vs structured data
• Narrative versus coded data

27
Recommendation 1

• Push toward each and all institutions moving
  toward interoperability within their own
  institutions
• Terminology
• Data elements
• Integration and aggregation
• Create a registry and recognition of institutions
  who accomplish this goal

28
Recommendation 2

• Create a basic EHR that we will provide free to
  any provider or clinic
• Primer coat
• Minimum system (demographics, lab results,
  medications, allergies, problem lists,
  encounters)
• Adequate decision support
• Simple, low maintenance, little education to use
  required
• Contract not open source
• Push for 100 adoption by providers in 3 years
• Provide specifications and funmctionality set for
  vendors

29
Recommendation 3

• Create an integrated terminology using inputs
  from SNOMED CT, LOINC, ICD, MEDRA, RxNorm, VA NDF
  RT, MEDCIN, NCI, IEEE, ICPC
• Major market for provider institution to switch
  to integrated terminology
• Create process for update of terminology
• Do real time distribution

30
Recommendation 4

• Identify who can influence movement to adoption
  of EHR
• All medical specialty organizations
• AAMC, AMIA, eHI, HIMSS, AMA, AHA, others
• Hospital groups
• Professional groups
• Make adoption of EHR a primary goal for any and
  all groups with influence
• Educate and promote within consumer groups push
  for EHR
• Create financial incentives

31
Recommendation 5

• Established trusted agency to be responsible for
  regional centers and to set up infrastructure.
• Probably 200-300 centers in US with creation cost
  of about 20M per region and an infrastructure
  cost of 2B (6B) and operational costs of around
  4M per regional center 800M plus operational
  costs of 200M for infrastructure for a total of
  1B per year.
• Model process after Highway Interstate System
• Provide vendor specifications to participate

32
Recommendation 6

• CMS and other payers provide all health care
  plans, including rules for authorization,
  approvals for reimbursement, etc. in computer
  readable and computer understandable form.
• Work toward real time adjudication of changes and
  real time reimbursement. Minimize human
  resources required to deal with reimbursement

33
Recommendation 7

• Make a recommendation on universal health care
  identifier
• Algorithm based on several data elements
• Assigned unique identifier
• Voluntary unique identifier