Transcript Health Information Technology

Interoperability,
Health Information Exchanges
and
Clinical Data Research
Networks
Shyam Visweswaran, MD, PhD
Associate Professor
Biomedical Informatics
Clinical and Translational Science Institute
12 May 2016
“Medicine used to be simple,
ineffective and relatively safe.
Now it is complex, effective and
potentially dangerous.”
Sir Cyril Chantler
Former Dean Guy’s, King and St. Thomas’s
Medical and Dental School
Lancet 1999
“Current practice depends upon
the clinical decision making
capacity and reliability of
autonomous individual
practitioners for classes of
problems that routinely exceed
the bounds of unaided human
cognition.”
Daniel R. Masys, MD
October 15, 2001 IOM Annual Meeting
Health information technology
is a critical enabler to broad
transformation in health care
with the ultimate goal of
obtaining high quality health
care.
Premise of the
Health Information Technology Economic
and Clinical Health (HITECH) Act
Health Information Technology (HIT)
uses:
• In the clinical enterprise to
– Integrate services and reduce fragmentation of care
– Improve patient safety and reduce errors
– Enhance communication between providers and
patients
• In the research enterprise to
– Leverage EHR for clinical and translational research
– Enhance quality improvement and outcomes research
Health Information Technology for
Networks
• Health Information Exchange (HIE)
– Verb: HIE is the mobilization of health care information
electronically across organizations within a hospital system,
region or nation
– Noun: HIE refers to the organization or entity that facilitates the
exchange
• Clinical Data Research Network (CDRN)
– CDRN is a network to share standardized health information
that is collected during the routine course of patient care for
research
• Interoperability
– Is the ability to share electronic data generated by one system
to be accessed and (re-)used in a meaningful way by another
system (whether or not the latter system is based on different
technologies)
Interoperability
• Interoperability describes the extent to which
systems and devices can exchange data, and
interpret that shared data
• Interoperability in computerized healthcare
information systems lags far behind other
domains such as finance
• Without interoperability, data and information
such as patient records can't easily be shared
across and sometimes within enterprises
• Interoperability is critical for building networks
like HIEs and CDRNs
Interoperability Challenges
• Health information technologies, like EHRs, where they
have already been deployed, may not have been designed
to support interoperability
• A lot of computerized clinical data are stored in legacy
systems in proprietary formats which are difficult for other
systems to access, re-represent and transfer for (re)use
• Many healthcare information standards to support
interoperability are only just now being developed
• Where healthcare information standards do exist they may
also compete, making interoperability more difficult to
achieve
• Implementation of interoperable health information
systems may require a high degree of technical expertise
not readily available to small organizations in particular
Three Levels of Health Information
Technology Interoperability
1.
Foundational interoperability
– allows data exchange from one information technology system to be
received by another
– does not require the ability for the receiving information technology
system to interpret the data
2.
Structural interoperability
– is an intermediate level that defines the structure or format of data
exchange (e.g., the message format standards)
– structural interoperability defines the syntax of the data exchange
3.
Semantic interoperability
– provides interoperability at the highest level, which is the ability of
two or more systems or elements to exchange information and to use
the information that has been exchanged
– this level of interoperability defines the semantics in addition to the
syntax of the data exchange
– this level of interoperability is required to improve quality, safety,
efficiency, and efficacy of healthcare delivery
ONC
• The Office of the National Coordinator for Health
Information Technology (ONC):
– is a division of the U.S. Department of Health and
Human Services
– was established in 2009, with the enactment of the
HITECH Act
– Is charged with formulating the federal government’s
health information technology (health IT) strategy and
coordinating federal health IT policies, standards,
programs, and investments
ONC Activities
• Policy Development and Coordination
– Health IT Policy
– Privacy and Security
– Health IT Safety and Usability
• Standards, Interoperability, and Certification
–
–
–
–
Standards Development and Harmonization
Health Information Exchange
Certification and Accreditation of EHR systems
Federal Health Architecture
• Adoption and Meaningful Use of Health IT
– Provider Adoption Support
– Consumer eHealth
– Planning, Evaluation, and Monitoring
Health Information Exchanges
(HIEs)
Benefits of HIEs
• Comprehensive electronic patient information
when and where needed
• Allow physicians to have complete and current
information upon which to base clinical
decisions
• Physicians and patients would receive
reminders about most recent clinical
guidelines and research results
What will HIE enable?
• Complete medical record always available
• Test results and imaging results always available
(eliminating repeat tests and imaging)
• Decision support always available: guidelines and
research results
• Real-time aggregation to detect patterns (e.g. bio
surveillance)
• Quality and payment information derived from
record of care – not separate reporting systems
• Patients have access to their own records
Challenges
•
•
•
•
•
•
lack of interoperability
lack of defined standards
privacy
security
stakeholder cooperation
high costs
Elements of HIE
• EHR systems
– in-patient
– out-patient
– community practices
• Ancillary health care systems
–
–
–
–
–
Pharmacy
Laboratory
Imaging
Physical therapy
Home health
Elements of HIE
• Communication and networking systems
– Information moves with patient
– Integrated information from all types of providers
• Standard medical terminologies
– Diagnoses & procedures
– Medications
– Laboratory test results
• Decision support
• Privacy and security
• Policy development and coordination
Common HIE Technical Architecture
Models
• Centralized Model
– patient data is collected from local sources but stored in a
central repository
– a request for patient data obtains data from the central
repository
• Decentralized or Federated Model
– patient data resides locally in participating organizations
– provides a framework for data sharing
• Hybrid Model
– a cross between a centralized and decentralized
architecture
– often provides an interface engine through which
organizations in the HIE communicate
Pennsylvania HIE
• Pennsylvania eHealth Partnership Authority was
created in July 2012
• The Authority will develop, implement, and
maintain:
– State-level patient identity management services
– State-level patient consent management services
(opt-out, opt-back-in) reflecting patient choice
– State-level directory of healthcare providers
– Certification programs for entities engaging in HIE
– Education and awareness materials for consumers and
providers about HIE
ClinicalConnect
• ClinicalConnect is western Pennsylvania’s first HIE
• Formed in 2009 as a non-profit organization
• Push/pull model: participating organizations
pushing continuity of care documents (CCDs),
which then get aggregated in ClinicalConnect
• Physicians view the information from
ClinicalConnect directly via their EHR system
• A button within each of the health system’s EHRs
launches a query to the HIE to provide a
summary view of the patient’s record
ClinicalConnect Members
ClinicalConnect View of Patient
• Physician logs into EHR
and selects patient
• Physician clicks a link to
access HIE data
• Patient data is displayed
from HIE participants in
an integrated view
Summary View
PCORnet & PaTH Network
Funded by PCORI
What is PCORI?
• The Patient-Centered Outcomes Research
Institute (PCORI) was established by the 2010
Patient Protection and Affordable Care Act
• PCORI is a non-governmental non-profit institute
• PCORI is charged with examining the "relative
health outcomes, clinical effectiveness, and
appropriateness" of different medical treatments
by evaluating existing studies and conducting its
own studies
What is PCORnet?
• PCORnet is the National Patient-Centered
Clinical Research Network
• A national network for conducting
comparative effectiveness research (CER)
• Enable a range of observational and
experimental CER by establishing a resource of
clinical data gathered in healthcare settings,
and by patient groups
Components of PCORnet
• Clinical Data Research Networks (CDRNs) are networks
of healthcare systems, such as hospitals and health
plans, and collect health information during the
routine course of patient care
• Patient-Powered Research Networks (PPRNs) are
networks operated by groups of patients and their
partners, and are focused on a particular condition or
population and whose members are interested in
sharing health information for research
• The Coordinating Center, led by Harvard Pilgrim Health
Care Institute and Duke Clinical Research Institute,
provides technical and logistical support to the
networks
13 CDRNs + 20 PRRNs + 1 CC
CDRN Disease Cohorts
CDRN
Common Cohort
Rare Cohort
ADVANCE
Diabetes
Co-infection with HIV and hepatitis C
virus
CAPriCORN
Anemia; Asthma
Sickle cell disease; Recurrent C. Difficile
colitis
Great Plains
Collaborative
Breast Cancer
Amyotrophic Lateral Sclerosis
Louisiana Clinical Data
Research Network
Diabetes
Sickle Cell Disease, Rare Cancers
NYC-CDRN
Diabetes
Cystic fibrosis
Mid-South CDRN
Coronary Heart Disease
Sickle Cell Disease
PEDSNet
Inflammatory bowel
disease
Hypoplastic left heart syndrome
PORTAL
Colorectal Cancer
Severe Congenital Heart Disease
pSCANNER
Congestive Heart Failure
Kawasaki Disease
PaTH
Atrial Fibrillation
Idiopathic Pulmonary Fibrosis
SCIHLS
Osteoarthritis
Pulmonary arterial hypertension
PCORnet Common Cohort
• Common Cohort: All CDRNs contribute to a
common cohort
• Currently, ~85 million patients across all
CDRNs
• Temple University
• UPMC/Pitt
• Penn State University
• University of Utah
• Johns Hopkins University
Goals of PaTH
• Link EHR data across five diverse healthcare
systems so that it can be used to answer
questions to improve health and health care,
while ensuring patient privacy
• Engage patients and clinicians to identify critical
research questions for improving health care and
to participate in research
• Develop convenient survey methods to let
patients share their perspectives on health topics
with their healthcare teams
Moving Data from EHRs to i2b2
ETL: Extract-Transform-Load
Cerner
PSU
i2b2
PSU
Heterogeneous
Homogenous
i2b2
Epic
Temple
Extract
Transform
Site specific
Little data
modification
Uniform across PaTH
Common Data
Elements
Load
Uniform across PaTH
Same i2b2 schema
Temple
Epic
i2b2
JHU
JHU
Cerner
Epic
UPMC
i2b2
UPMC
Use Case: ADAPTABLE Trial
• May 2015: First major demonstration of PCORnet for a
trial
• ADAPATABLE Trial: Aspirin Dosing: A Patient-centric
Trial Assessing Benefits and Long-Term Effectiveness
• Objective: To compare the effectiveness and safety of
two doses of aspirin (81 mg and 325 mg) in high-risk
patients with coronary artery disease
• Primary composite outcome —death, hospitalization
for nonfatal myocardial infarction, or stroke
• Primary safety end point — major bleeding
complications
Use Case: ADAPTABLE Trial
• Pragmatic clinical trial
– Explanatory trials generally measure efficacy - the benefit a
treatment produces under ideal conditions, often using carefully
defined subjects
– Pragmatic trials measure effectiveness-the benefit the
treatment produces in routine clinical practice
• Embeds the trial within usual care
• Recruits a diverse patient population with minimal
eligibility criteria
• Relies on electronic data collection with reduced need for
costly primary data collection
• Will recruit 20,000 patients with heart disease
• Expected to cost less than $1000 per participant, an
amount far below that of a typical trial of this scope
ADAPTABLE Protocol
Screening of CDRN EHR data with
computable phenotype
Electronic outreach to potential patients
with trial introduction and link to
ADAPTABLE web portal
•
•
Web-Based, Electronic Informed Consent
Initial patient contact via web portal  text and video consent options
Developing a common consent form with selected local adaptations
Focused questions to confirm patient comprehension for informed
consent and eligibility for randomization after consent
Randomization and Aspirin dose assignment
ACT Network
Funded by NCATS/NIH
ACT Network Vision
• Create a federated research data network for
the Accrual of patients for Clinical Trials
• Stage 1: Build network for identifying patients
using EHR data in real time – Cohort Discovery
• Stage 2: Contact and enroll identified patients
into clinical trials – Recruitment
• Stage 3: Patients and practitioners identify
clinical trials – Participation
Stage 1: Cohort Discovery
• Time frame: 1 year (September 2014 – August
2015)
• Number of CTSA sites: 13 in Wave 1 + 8 in
Wave 2 = 21 sites
• Goal: Create a i2b2/SHRINE network that
allows investigators to query EHR data on a
limited set of data domains across all sites in
real time
21 ACT Sites
11
19
7
4
10
13
17
12
2
20
8
3
1
5
1514
16
6
9
21
18
Legend
Red: Wave 1
Blue: Wave 2
ACT Work Structure
Executive Committee
4 PIs (Steven Reis, Lee Nadler,
Robert Toto, Gary Firestein)
plus Work Group Leads
Central PM
Vince D’Itri
(Aspen Advisors)
Governance
Regulatory
Technology
Dipti Ranganathan
Robert Toto
Karen Allen
Doug McFadden
Nick Anderson
Work Groups
Data
Harmonization
Shyam Visweswaran
Michael J. Becich
Work Groups
• Governance Work Group:
– Developed Governance Structure Document
– Terms of Data-Access Agreement
• Regulatory Work Group:
– Developed Regulatory Guidance Document
– Developed IRB Application
Technology Work Group
• Technology Work Group:
– Helped install i2b2 / SHRINE software at all sites
– Installed SHRINE hub at Harvard
– Helped troubleshoot network connectivity
• Data Harmonization Work Group:
– Selected data domains and identified data
elements for the ACT Data Model
– Selected terminologies and classification systems
ACT Data Model
• Demographics
– birth date, sex, Hispanic
status, race, vital status,
death date
• Diagnoses
– diagnosis (ICD-9), diagnosis
date, diagnosis source
(admit, discharge),
diagnosis priority (primary,
secondary)
• Procedures
– procedure (ICD-9),
procedure date
• Visit Details
– admit date, discharge date,
visit type (inpatient,
ambulatory, ED)
• Medications
– medication (RxNorm + NDFRT), order date, order type
(inpatient, ambulatory)
• Laboratory Test Results
– lab test (LOINC + LOINC parts
hierarchy), specimen date,
result location (lab, point of
care), result, units
Medications
NDF-RT classification
RxNorm IN (Ingredient)
RxNorm SCDF
(Ingredient plus dose form)
RxNorm SCD and SBD
(Ingredient plus strength
and dose form)
Laboratory Tests
LOINC parts
hierarchy (LP code)
LOINC code
Use Case: Rheumatoid Arthritis
• Goal: To identify patients with moderate to severe
rheumatoid arthritis (RA) and an inadequate response to
methotrexate.
• Inclusion Criteria:
–
–
–
–
–
–
Diagnosis: Rheumatoid arthritis (ICD-9 714.0)
Duration of disease: <2 years
Active disease: CRP>1.2x ULN or ESR>30 mm/hr
Age: Between 18 and 75 years
Sex: No criteria
Medications:
• Methotrexate >3 months at >7.5 mg/week
• And currently either on Prednisone dose ≤10 mg/day or not on
prednisone
• And no current biologic (etanercept, golimumab, adalumimab,
anakinra, infliximab, certulizumab, , rituximab) or JAK inhibitor
(tofacitinib)
Rheumatoid Arthritis
• Inclusion Criteria contd.:
– Laboratory:
• Hgb >10 g/dl, ALT and AST <ULN
• And T bili <ULN, Creatinine <ULN
• Exclusion Criteria:
–
–
–
–
–
–
Active tuberculosis
Hepatitis B
Hepatitis C
HIV
Pregnancy
Enrolled in another clinical trial
Summary & Plans
• Summary:
– Number of patients in ACT network = ~40 M
– Number of patients satisfying RA query = 7,223 from
16/22 sites
– Time for RA query = 124 seconds
• Plans:
–
–
–
–
Data characterization
Open the network incrementally to investigators
Expand ACT to include all 62 CTSA sites
Add more data domains
ACT Demo
• ACT i2b2: http://dbmi-ncatsprod01.dbmi.pitt.edu/webclient/
• ACT SHRINE: http://dbmi-ncatsprod01.dbmi.pitt.edu:6060/shrine-webclient/
Questions?
Thank you