This is a talk I gave at my own organisation - National Institute for Health Innovation (NIHI) of the University of Auckland on 6 Aug 2014. Abstract as follows:
In this talk I’ll first cover the topic of clinical registry – an invaluable tool for supporting clinical practice but also gaining momentum in research and quality improvement. NIHI has been very active in this space: we have delivered the prestigious and highly successful National Cardiac Registry (ANZACS-QI) together with VIEW research team and also very recently launched the Gestational Diabetes Registry with Counties Manukau DHB & Diabetes Projects Trust. A few others are in likely to come down the line. This is a huge opportunity for health data driven research and NIHI to position itself as ‘the health data steward’ in the country given our independent status and existing IT infrastructure and “good culture” of working with health data . NIHI’s ‘health informatics’ twist in delivering these projects is how we go about defining ‘information’ – using a scientifically credible and robust methodology: openEHR. This is an international (and now national too) standard to non-ambiguously define health information so that they are easy to understand and also are computable. We build software (even automatically in some cases!) using models created by this formalism. I’ll give basics of openEHR approach and then walk you through how to make sense out of all these. Hopefully you may have an idea about its ‘value proposition’ (as business people call) or Science merit as I like to call it ;)
Health research, clinical registries, electronic health records – how do they (if at all!) come together?
1. Health research, clinical registries,
electronic health records
how do they (if at all!) all come together?
Koray Atalag MD, PhD, FACHI
k.atalag@auckland.ac.nz
Vice Chair HL7 New Zealand
openEHR Localisation Program Leader
Health Information Standards Organisation (HISO) Committee Member
NHITB Sector Architects Group Member
3. Registry defined
An organised system that
uses observational study methods
to collect uniform data(clinical and other)
to evaluate specified outcomes for a population
defined by a particular disease, condition, or exposure,
and that serves a predetermined scientific, clinical or
policy purpose(s).
GliklichR, Dreyer Ne. Registries for Evaluating Patient Outcomes: A User's Guide Prepared by Outcome DEcIDECenter[Outcome Science, Inc.
dbaOutcome] under Contract No. HHSA290200500351TO1). Rockville, MD: Agency for Healthcare Research and Quality, 2007; Publication No.
07-EHC001-
4. Clinical Registries
Register / Registry
Clinical (+quality) / disease / patient / incidence / screening etc.
Repository of individuals with certain conditions/characteristics
Ease of access to important info
Track clinical processes & (risk adjusted) outcomes
Longitudinal history of correspondences & interventions
Prompt / feedback to participants and providers
Data linkages & Reporting
Supporting clinical practice
◦ Screening, risk prediction, intervention/recall, safety monitoring
Clinical quality improvement
◦ Organisations, clinicians, policy makers
Research & education
5. Why do we need them?
Because we don’t have the mighty EHR!
Registries are a ‘quick fix’ to some ‘can’t wait’ type
problems / for ‘quick wins’; capturing
◦ observations, diagnoses, procedures, clinical processes and
most importantly outcomes
Provide an infrastructure on which intervention studies
can be established with relative ease.
Who get’s a registry?
◦ Those with funding of course!
Clinical significance / popularity (eg. CVD, diabetes)
Well established network/specialised (e.g. Spina Bifida)
national/intl policies (MoH / WHO – cancer etc.)
leadership / persistence / charisma / luck (GDM?)
6. Around the world & NZ
A lot of them!
Overarching principles / regulations /
minimal standards
Shared resources (hosted by dedicated
organisations / infrastructure)
A growing number of them
All go own ways – (under privacy rules)
Hosted/curated by source groups with limited
technical/data management resources
7. Typical Uses
Incidence/prevalence of diseases/conditions in
populations & monitor trends/survival rates over time
safety & quality of products and treatments
clinical and/or cost effectiveness of treatment
(including drugs, devices and procedures) across a
population
provide denominator & vehicle for interventional
studies
and sometimes decision support too!
8. Electronic Health Records (EHR)
All directly recorded or derived information
about an individual within healthcare context
in electronic form
It is called many names – EHR, EMR, PHR, CPR,
EPR, CBPR, AMR, EHCR, ...
Different perceptions: function, purpose,
disease, place etc.
Ref: Ed Hammond
9. What does the EHR Contain?
DATA
Person-centred
Comprehensive
Longitudinal
Organized
High data integrity
Timely
Structured
Semantically coherent
Shareable
Trustable and accountable
Secure and private
Ref: Ed Hammond
10. What does the EHR Provide?
Information
for
Direct patient care
Effective decision support
Prevention of medical errors
Improved quality of care
Better clinical communication
Enabling shared care
Evidence based care
Cost effective care
Workflow management
Bio-surveillance
Research
Epidemiology
Billing/reimbursement/health policy/planning
Ref: Ed Hammond
13. Barriers to EHR Adoption
Large initial investment, unfit funding models
Poor user acceptance (workload?)
Privacy concerns
Lack of solid evidence?
Fear & reluctance for the unknown
Political/societal ignorance
Medico-legal issues
Risky business (for vendors/purchasers)
Lack of common information / processes
....Interoperability
14.
15. Types of Interoperability
Technical Interoperability: systems can send and receive data successfully.
(ISO: Functional/Data Interoperability)
Semantic Interoperability: information sent and received between
systems is unaltered in its meaning. It is understood in exactly the same
way by both the sender and receiver.
Process Interoperability: the degree to which
the integrity of workflow processes can be
maintained between systems.
(This includes maintaining/conveying
information such as user roles between systems)
(HL7 Inc.)
16. If the Banks Can Do It,
Why Can’t Health?
Clinical data is wicked:
◦ Size (breadth, depth) and complexity
◦ >300,000 concepts, 1.4m relationships in SNOMED
◦ Variability of practice
◦ Diversity in concepts and language
◦ Conflicting evidence
◦ Longevity
◦ Links to others (e.g. family)
◦ Peculiarities in privacy and security
◦ Medico-legal issues
It IS critical…
17. Can Clinicians Agree on Single
Definitions of Concepts?
“What is a heart attack?”
- 5 clinicians: ~2-3 answers – probably more!
“What is an issue vs. problem vs. diagnosis?”
- No consensus for conceptual definition for years!
BUT
There is generally agreement on the structure and attributes of information
to be captured
Problem/Diagnosis name
Status
Date of initial onset
Age at initial onset
Severity
Clinical description
Date clinically recognised
Anatomical location
Aetiology
Occurrences
Exacerbations
Related problems
Date of Resolution
Age at resolution
Diagnostic criteria
Acknowledgement: Sam Heard
19. Why bother?
(with a standard structured Medication model)
“If you think about the seemingly simple concept of
communicating the timing of a medication, it readily
becomes apparent that it is more complex than most
expect…”
“Most systems can cater for recording ‘1 tablet 3 times a
day after meals’, but not many of the rest of the
following examples, ...yet these represent the way
clinicians need to prescribe for patients...”
Dr. Sam Heard
25. Open source specs & software for representing
health information and person-centric records
◦ Based on 18+ years of international implementation experience
including Good European Health Record Project
◦ Superset of ISO/CEN 13606 EHR standard
Not-for-profit organisation - established in 2001
www.openEHR.org
Extensively used in research
Separation of clinical
and technical worlds
Big international community
26. Logical building blocks of EHR
Compositions
EHR
Folders
Sections
Clusters
Elements
Data values
Entries
27. Patterns in Health Information
Actions
Published
evidence base
Personal
knowledge
Evaluation
Observations
Subject
Instructions
Investigator’s
agents
(e.g. Nurses,
technicians, other
physicians or
automated devices)
Clinician measurable or
observable
clinically interpreted
findings
order or initiation of a
workflow process
Recording data
for each activity
Administrative Entry
Acknowledgement: openEHR
33. What about secondary use?
Interoperability for clinical information
systems – great
◦ But what about population health & research?
Research data also sits in silos – mostly C
Drives or even worse in memory sticks!
Difficult to reuse beyond specific research
purpose – clinical context usually lost
No rigour in handling and sharing of data
34.
35.
36.
37.
38.
39. Exploiting the Content Model for Secondary Use
Atalag K. Using a single content model for eHealth interoperability and secondary use. Stud Health Technol Inform. 2013;193:282–96
Single Content Model
CDA
FHIR
HL7 v2/3
EHR Extract
UML
XSD/XMI
PDF
Mindmap
PAYLOAD
System A
Data Source A
Map
To
Content
Model
System B
Data Source B
Native openEHR Repository
Secondary Use
Map
To
Content
Model
Automated Transforms
No Mapping
41. Gestational Diabetes Registry
Development
in CMDHB
Dr. Koray Atalag MD, PhD, FACHI (National Institute for Health Innovation)
Dr. Carl Eagleton MBChB, FRACP (Counties Manukau District Health Board)
Karen Pickering (Diabetes Projects Trust)
42. Aims
100% successful screening of women for type 2
diabetes (T2DM) within 3 months after a pregnancy
with GDM
Annual screening of all women for new onset T2DM
Early warning to healthcare providers (GPs,
Maori/Pacific Health, others) about GDM history in
subsequent pregnancies
43. Gestational Diabetes Mellitus (GDM)
GDM is characterised by glucose intolerance with
onset or first recognition during pregnancy & is
identified by an oral glucose tolerance test (OGTT)
A repeat OGTT performed 6 weeks post-partum
checks for resolution
◦ If normal, an annual fasting glucose or glycosylated
haemoglobin (HbA1c) screening test is recommended for
T2DM, according to New Zealand (NZ) guidelines.
44. Opportunities & Motivation for the Registry
Long term consequences can be prevented by regular
screening for early detection of T2DM or high CVD risk
◦ CMDHB found 20% of women with a history of GDM were not
follow-up tested in a 4 year period; (37% for 2 year period)
◦ Sending out reminders improve adherence / better compliance
with screening recommendations
Risk of developing T2DM can be substantially reduced
by early identification of women at high risk + targeted
lifestyle & pharmacological interventions
Registry can also be used to drive clinical quality
improvement and enhance patient safety
◦ by identifying variations in processes and clinical outcomes.
45. Main Considerations
Privacy / Confidentiality
◦ Privacy Act 1993 and Health Information Privacy Code 1994 (“HIPC”)
◦ Recent changes to offshore hosting
◦ Connected Health secure network
Security / Recovery / Availability
◦ Univ. of Auckland’s secure IT infrastructure
IT standards & components
◦ W3C, Microsoft Net, SQL Server, Angular JS
◦ HISO Interoperability Reference Architecture
◦ openEHR
Existing systems
◦ CMDHB: Maternity CIS & others
◦ Regional/National: MoH datamart? VDR, PMS, Shared Care etc.
46. GDM Registry Pathway
Entry
• Referral from primary care with a diagnosis of GDM
Education
• Attendance at Group Session
• Registry information supplied
Consent
• Attendance at DiP Clinic
• Consent obtained and entry into the registry
Postpartum
• 6 week OGTT request or 3 month HbA1c
• GP & Patient advised of results
Annual
• Annual HbA1c with copy to primary care
• GP & Patient advised of results
Next time
• Positive pregnancy test detected in Testsafe
• Requesting healthcare provider advised of Diabetes history by the Registry
RegistryDirected
47. Golden principle: Minimal data entry, Maximal reuse!
Technical Development
Used an international (and HISO) standard:
◦ Consistent dataset
◦ Interoperability / integration
◦ Manage change over time
Used a Web-based data set development tool to
review & finalise
Automatically converted dataset into “software
code” [domain objects]
Built on NIHI’s clinical data repository framework
51. The Registry System
Three main parts (exc. System admin)
◦ Demographics
◦ Clinical view + entry
◦ Intervention
Role based access: clinical and/or admin
Entry status:
◦ Temporary: record still being populated (import/data
entry), not included in actions
◦ Active: records visible to all & complete
◦ Inactive: only admin can see, suspended/opt-out
Enter a new participant Activate
Enter / Update clinical data
Interventions preview
54. EHR Providing a Canonical Representation
so we know what kind of info goes into which bucket!
Demographics
ClinicalEncounter
VitalSigns
Medications
Diagnoses
DiagnosticTests
Interventions
FamilyHistory
PastHistory
PhysicalExam
Genetics
LifeStyle
etc.etc.etc.
Subject A
Subject B
Person-Centric Record Organisation
NZ Address
Ethicity1,2.
Whanau
USAddress
State
Next of kin
GP visit
Flu-like
PHO enrolm.
Hospital adm.
Diabetes
Priv insurance
BP 130/90
HR 90
T: 38.5 C
BP 120/70
(24 hour avg)
HR 70
T: 37 C
Rx A
Dispense
Administer
Rx B
Dispense
Administer
Dx 1
Dx 2
etc.
Diabetes Dx
-Type
-Severity
-Course etc.
Routine Blood
Urine
X-Ray
Specific blood test
Urine culture
Genomic assay
Retinography
Rx
Fluid Tx
Insuline inj
Infection Tx
Psychologic
N/A
Pedigree
N/A
Chronic
Routine
Detailed
Foot and
eyes
N/A N/A
DNA
Seq.
Assays
Low
sugar
Exercise
Shared Archetypes
Each finding usually depends on other – clinical context matters!
55. Bottom line
We may not have EHR now....but
by using openEHR to represent our clinical information
we are leveraging some of the benefits of EHR today,
including
◦ Expressivity, clinical context, meta-data support
◦ Interoperability
◦ Semantic querying (easy + fast)
◦ Tooling and international content
◦ Standards compliance
and future-proofing registry data!