Association of GCRC Statisticians August 7, 2004

1
Association of GCRC StatisticiansAugust 7, 2004
Information Integration -The Next Frontier

• Joyce C. Niland, Ph.D.
• Professor and Chair, Information Sciences
• City of Hope National Medical Center

2
Executive SummaryOctober 2003 GCRC Workshop
3
Outline

• Why This Workshop, and Why Now?
• Overview of Workshop Presentations
• Vision of the GCRC Informatics Cores
• Group Discussion

4
Challenges in Clinical Research

• Efforts to mount clinical research are enormous
• Data rarely can be effectively exchanged,
  integrated
• Lack of standards hampers efficient trial design,
  decision support, analysis

5
National Strategic Initiativesin Support of
Biomedical Research
6
30,000 Foot View National Health Information
Infrastructure (NHII)

• Department of Health and Human Services (HHS)
  initiative to improve effectiveness, efficiency,
  and overall quality of health care
• Stemmed from April 2004 executive order to
  establish interoperable electronic health records
  within 10 years
• Kickoff session July 21-23, 2004 in Wash DC
• DHHS Secretary Tommy Thompson
• Health Information Technology (HIT) Coordinator
  David Brailer, MD, PhD
• Honorable Bill Frist, Rep Nancy Johnson, Newt
  Gingrich
• Heads of FDA, CDC, VA, NIH, DoD, AHRQ, OPM

7
National Health Information Infrastructure (NHII)

• Four main goals of HIT initiative
• National electronic health records
• Interoperable infrastructure to promote patient
  record portability across clinicians and practice
  sites
• Facilitating patient participation and
  personalized care
• Using HIT to improve population health
• Estimated that completed project will save
  130Bin health care costs
• Other tangible benefits include
• Decreased medical errors
• Earlier detection of bioterrorism
• Improved local health care systems
• Increased efficiency in medical research

8
Clinical Research SIG Preamble
We the people . . . . hereby recommend that the
NHII ensure that clinical and translational
research needs are interwoven through the NHII
strategic plan, in order to form a more perfect
Union between clinical care and research
endeavors. Be it known that we also urge our
nations government to articulate the long-term
vision and mission for research within the NHII,
in addition to the short-term actionable
articles set forth herein.
9
The NHII should be concerned with all facets of
research related to human health and well-being,
including but not limited toclinical research,
basic science, population research, environmental
sciences, bioengineering, and informatics.
Clinical Research SIG Preamble
. . . . These truths we hold self-evident.
10
NIH Roadmap
11
Dr. Zerhounis NIH Roadmap

• Speed biomedical research by
• Building on strengths
• Identifying best practices
• Increasing collaboration among researchers

12
Three Arteries of the NIH Roadmap
13
Dr. Zerhounis NIH Roadmap

• Speed biomedical research by
• Building on strengths
• Identifying best practices
• Increasing collaboration among researchers

• Build the National Electronic Clinical Trials
  Research Systems (NECTAR)
• Enhance clinical research through informatics
  tools
• Standardize data, software tools, and network
  infrastructure for seamless data and sample
  sharing
• Maximize connectivity among existing and newly
  created clinical research networks

14
Roadmap Will Require the Full Scope
of the Discipline of Informatics
Research,development, and application of tools
and approaches for measurement, collection,
annotation, analysis, and interpretation of
patient care / clinical research data
lt---BIOMEDICAL INFORMATICS---gt
Basic Science Clinical
Research
Bioinformatics Clinical Research
Informatics
Genotype Phenotype
Measurement, collection, annotation, analysis and
interpretation of biological data, often at the
genome level. (Employs tools / techniques from
Computational Biology and Biostatistics)
15
Contributing to the Global Solution

• GCRCs represent a substantial investment in
  biomedical research and data
• Major opportunity and potential to contribute to
  the NIH Roadmap and NECTAR
• Informatics Cores operating independently
• Often inventing their own informatics toolsto
  meet these challenges (of necessity)
• General lack of communication /consensus on
  informatics approaches
• Rationale for two-day workshopat the NCRR last
  October

16
Organizing Committee

• Composed of NCRR, AGITP, BIRN, and FIRST
  representatives
• Dr. Elaine Collier
• Dr. Bob DiLaura
• Dr. Randy Gollub
• Dr. David Ikle
• Dr. Joyce Niland
• Dr. Bob Rice
• Dr. Hemant Shah
• Dr. Doug Stahl

17
Substantial Interest from GCRC Cores

• Evidenced by over 100 attendees from 54GCRC
  institutions
• Excellent abstract submissions and active
  participation in two-day meeting
• Objective Identify approaches to make GCRC
  Informatics Core most effective contributor to
  biomedical research through IT approaches

18
Day One Agenda

• Welcome Dr. Judith Vaitukaitis Dr.
  Anthony Hayward
• National Initiatives BIRN caCORE
  FIRST GCRC Survey
• Local GCRC Initiatives Clinical Research
  Management Tools for Research
  Collaboration
• The Future A Standards Perspective

19
Day Two Agenda

• Moderated Panel I Vocabulary Data Standards
• Moderated Panel II Open Source Development
• Poster Sessions
• Breakout Sessions Role of the GCRCs in
  Developing and Adapting Informatics Tools
• Wrap-up Discussion

20
Informatics InitiativesPresented
21
Survey of GCRC Informatics Cores (B. DiLaura)

• Excellent response rate (95)
• Some key findings
• Many one-off solutions in place
• Mixed levels and types of technologies
• Lack of current cross-GCRC interactions
• Many unmet needs scheduling, SAE reporting, etc.

22
Local GCRC Initiatives

• Presentations and posters on GCRC Informatics
  Cores roles in
• Clinical research management tools
• Web-services suite of technologies
• Barriers to data sharing and integration
• Web-based applications to integrate
  heterogeneous data sources
• Facilitating research through the AGITP

23
BIRN Biomedical Informatics Research
Network (R. Gollub)

• Testbed for a biomedical knowledge infrastructure
• Creation and support of federated bioscience
  databases
• Data integration and interoperable analysis/data
  mining tools
• Scalable and extensible
• Driven by research needs pull, not technology
  push

24
FIRST Fully Integrated Research Standards
Technology (J. Niland)

• 3-year GCRC supplement grant conducted at COH
• Input from USC and CHLA
• Collaboration from Los Alamos National
  Laboratory (Dr. David
  Forslund)
• Goal Develop a fully integrated scalable
  information management and decision support
  environment
• Creation of a unified informationarchitecture
  with standard
• Data Model
• Vocabularies
• Data Interfaces
• Informatics Training Program

25
Functionality to be Supported in FIRST
Results
Research Idea
Protocol Tracking
Outcome Measures
Eligibility Screening
Protocol Monitoring
Patient Tracking
26
Supporting Biomedical Research at City of Hope
Division of Information Sciences (136 Members)
Protocol Design
Project Mgmt
Data Analysis
Systems Biology
Data Collection
Applications
Study Management
Architecture
Database Modeling
27
Incorporating Standards

• Standardized Unified Modeling Language (UML)
  models for clinical research functions
• Use of standardized vocabularies, and semantic
  analysis to map among vocabularies
• SNOMED
• LOINC
• NCI CDEs
• Data interexchange standards
• eXtensible Markup Language (XML)

28
Incorporating Standards

• Health Level 7 (HL7) standard for clinical data
  interchange Reference Information Model (RIM)
• V3.0 includes standardized class and block
  diagrams to represent healthcare information
• HL7 Regulated Clinical Research Information Model
  (RCRIM)
• Evolving a standardized model for research
• City of Hope participating as HL7 institutional
  member
• Also participating in CDISC modeling

29
Clinical Data Interchange Standards Consortium
(CDISC)

• An open, multi-disciplinary, non-profit
  organization
• Began in 1997
• Mission lead the development of global,
  vendor-neutral, platform independent standards to
  improve data quality and speed product
  development
• Development of industry standards to support
  electronic acquisition, exchange, submission, and
  archiving of clinical trials data

30
Where CDISC Fits in the World of Standards
International Conference on Harmonization (ICH)
US FDA
Japan MHW
EU EMEA
Pharmaceutical Industry EU USA
Japan EFPIA PhRMA JPMA
REGULATORY AUTHORITIES
Health Care Providers Pharmacies - NCPDP
CDISC
Clinical Trials
Standards HL7, XML
Dictionaries MedDRA, LOINC
Models NCI, OMG
31
Standards Perspective from the Experts

• Health Level 7- V3.0 (Dr. Gunther Schadow)
• Reference Information Model (RIM)
• Entities (people, animals, devices) in their
  roles
• Participate in acts (clinical trial)
• Utilizes UML standards
• Block diagrams to save screen real estate
• A grammar to construct medical information
• Emerging RCRIM Regulated Clinical Research
  Information Model
• CDISC to submit their model to RCRIM for ballot

32
HL7 RIM Class Diagram
33
FIRST Block DiagramProtocol Registration
34
A Standards Perspective

• Distributed, Service-Based Architecture (Dr.
  Forslund)
• Interoperable components (plug and play)
• Standardized components that separate
  specification from implementation through
  interfaces
• Service-oriented architecture approach
• Define fundamental capabilities around an object
  model (UML)
• Data models are fully extensible and independent
  of any database representation
• Use of common services would allow GCRCs to work
  together

35
Significance of CORBA Technology

• Advantages through flexibility and reusability of
  CORBA services
• Minimizes additional development
• E.g. built-in security, audit trail
• Places emphasis on knowledge representation,rathe
  r than database construction
• Database created automatically via XML Data Type
  Definition (DTD)

36
CORBA Graphical Representation
37
A Standards Perspective

• Standard Terminologies (Dr. James Cimino)
• Standards support merging of data, reuse of
  information
• Several established standards (LOINC, SNOMED-CT)
• Many have shortcomings (CPT, ICD-9)
• Emerging standards show great promise (Rx-NORM)
• Unified Medical Language System (UMLS)
• Not created as a terminology
• Subsumes other terminologies, maps among them

38
Moderated Panels

• Vocabulary Data Standards (Cimino, Covitz)
• Experiences and problems in implementing
  vocabulary standards
• Motivations for utilizing, successes, removing
  impediments
• Open Source Development (Forslund, Ho, Schadow)
• Development of national healthcare initiatives
• Potential for collaborative development within
  GCRCs

39
NCIS caCORE Infrastructure for Clinical Research
Scientific Applications
40
Biomedical Information Tsunami

• Overwhelming data volume
• Multitude of sources

41
NCI caBIG

• Goal A virtual web of interconnected data,
  individuals, and organizations redefines how
  research is conducted, care is provided, and
  patients/participants interact with the
  biomedical research enterprise
• Cited by DHHS Secretary Tommy Thompson as
  potential precursor to NIH NECTAR

42
caBIG Desiderata

• Open access, open source
• Derived from common information models
• Standards for data exchange formats
• Data and metadata following ISO 11179
• Consume appropriate public, open access standards
  when available
• (Note These same principles followed in FIRST)

43
Timeline

• Phase I (summer 2003)
• Meetings to discuss caBIG initiative
• NCI-designated Cancer Center needs/capability/inte
  rest presented at cooperative development
  meetings
• Extensions to NCICB infrastructure
• Phase II (fall 2003)
• Selected pilot participants
• Established working groups
• Established communications portal
• Phase III (spring 2004)
• Establishment of master contracts
• Assessment of pilot statements of work

44
The graph below shows the frequency of needs
identified by Cancer Centers duringthe
Cooperative Development Meetings to select the
final participants for the caBIG pilot.
Leading priority
45
NCI caBIG Award Announcement (December 19, 2003)
City of Hope selected as Developer sitefor
caBIGclinical trial management systems(with
Universityof Pittsburgh)
46
WorkshopBreakout Sessions
47

• Created 4 Teams
• Mission -
• Key themes, issues or priorities?
• Significant challenges/barriers to progress?
• Changes required or helpful to get started?
• (include human, IT, other resources)
• Where do we go from here - recommendations?
• Support
• Assigned each Team a Facilitator
• Team picked a Scribe and Spokesperson

48
Wrapping It UpSummary ofTeam Deliberations

• GCRC Informatics Meeting
• October 30, 2003

49
Vision for Future GCRC Collaboration

• Remarkable consensus among participants to work
  cohesively and cooperatively in future
• Transform GCRCs into powerful facilitatorsof
  biomedical research
• Vision kindled excitement and energyacross all
  four Working Groups
• Integration theme was embraced by all four
  Breakout Groups

50
(No Transcript)
51
Unanimous Sentiment of the Groups

• Increased sharing and collaboration
• Among GCRCs
• With other clinical research networks, regulatory
  agencies, IRBs, and laboratories
• Shared resources could include
• Data
• Software and source code
• Knowledge and expertise
• Adoption of collaborative approaches to solving
  common informatics needs a priority

52
Vision for a Changing Culture

• Increased input from the various stakeholders
• Regarding common and site-specific needs
• GCRC-wide discussion on shared guidelines
• Policies for data sharing
• Concerns regarding intellectual property
• Explore shared utilization of resources and
  infrastructure
• Open source software movement may provide
  inspiration tools for GCRC-wide collaboration

53
Specific Issues Raised

• Tension between role of GCRC Informatics Core as
  service-provider versus informatics research
• Leads to unmet expectations and frustration
• Remove silos and isolation of individual GCRCs
• Sharing of expertise, knowledge and current
  activities
• Security and privacy must be addressed
• Comply with local, state, and federal regulations

54
Sample Initial Beneficial Project

• UML-based documentation of workflow and process
  requirements across GCRCs
• Numerous benefits
• Define common areas for building of informatics
  tools
• Help new GCRCs get started more quickly
• Allow new or diverse personnel (and
  investigators)to understand overall operations
  and supporting infrastructure of GCRCs

55
Standards Will Be Critical!

• Global consensus that standards are vital
  component for collaboration and advances
• Necessary to speed scientific discoveries,
  promote information shareing, and allow data
  integration
• Recognized the value of common vocabularies and
  data models across GCRCs
• Adoption of current and emerging standards also
  would facilitate goal of interactions with other
  research communities

56
Think globally,
act locally
57
Where Do We Begin???

• Initial collaborative GCRC projects should
• Have high priority
• Be relatively simple, fast and inexpensive
• Notion of Low Hanging Fruit
• GCRCs should articulate discrete projects
• Utilize forum of the national meetings
• Embrace open source software and component-based
  development approaches

58
Recommendations for Going Forward

• Training of all GCRC personnel
• Informatics Managers
• Investigators
• Mentoring program for ongoing development
• Link more advanced centers and staff withthose
  less experienced
• Additional workshops and conferences
• Address specific identified needs

59
Follow-up to October 2003 Workshop

• Share information gleaned and ideas generated
  with informatics personnel at April 2004 meeting
• Done!
• Continued dialogue with all interested partiesto
  help refine the vision and the first steps
• Hopefully ongoing progress at this meeting!
• Specific speaker presentations can be found at
• http//www.infosci.coh.org/gcrcworkshop/presentati
  on.htm

60
Traveling the Highway to Enhanced GCRC
Collaborations
61
Questions Discussion