Interoperability in Clinical Trials

1
Interoperability in Clinical Trials

• Peter Covitz, Ph.D.
• Chief Operating OfficerNational Cancer Institute
  Center for Biomedical Informatics and Information
  Technology

caBIG Clinical Trials Management Systems
Workspace Meeting May 30, 2007
2
What is the Clinical Trials Management Systems
Workspace Trying to Achieve?

• Facilitate the planning and instantiation of
  clinical trials, (and their monitoring of trials
  once they are instantiated)
• Facilitate the conduct of clinical trials
• Facilitate the reporting and sharing of clinical
  trial data to existing/new destinations,
• Achieve interoperability
• Increase the ability of systems to access and use
  the data and functionality of other systems
• Facilitate the integration of new sources and
  destinations of data

3
Interoperability Relations
Applications within an organization (or virtual
org)
Organization and its sponsors and regulators
Organization and its peers in the research
community
4
How to achieve interoperability

• Standards-based vocabulary, data elements, data
  models to facilitate information exchange
• Use accepted international standards
• Work toward convergence of standards and models
• Common infrastructure, widely distributed
• Permits cancer community to focus on scientific
  innovation, not making software
• Collection of interoperable applications and
  services
• Component-based and developed to common standards
• Data Sharing
• Cancer information can be made widely available
  to diverse communities

5
Model-driven approach

• caBIG uses models to specify information system
  contents and as a basis for harmonizing across
  systems
• BRIDG is the model that forms the foundation for
  CTMS systems
• Consistent with both HL7 and caCORE methodology
• Early versions being leveraged in CTMS
  development and interoperability projects
• Production release of BRIDG 1.0, June 15 2007

6
Platform for interoperability caGrid

• caCORE
• Terminology Services
• Metadata registry
• Model-driven architecture

• Globus-based data and analytical service grid
• Strongly typed data objects
• Registered data models
• Grid-wide object identifiers
• XML transport format
• Federated security and identity management
• Rapid service creation tools

• Mobius
• Global Model Exchange (GME) XML infrastructure

• Globus
• Standards-based toolkit for grid computing

7
How does this all fit together?

• CTMS Interoperability (CTMSi) demonstration
  project
• Selected inter-application data transmission use
  cases from clinical trials experts
• Highlighted need for BOTH interoperable
  programming AND messaging interfaces
• Assembled caBIG CTMS applications, caGrid
  infrastructure, and message rou
• Created live, working multi-software component
  deployment that satisfies use cases

8
CTMSi Current
Demo Prototype
BRIDG
9
CTMSi Proposed
Proposed Target Architecture
Any caBIG CT Data Mgmt Product
API
MSG
API
MSG
API
MSG
API
MSG
API
MSG
Security Advertising Discovery Invocation
LOCAL
MSG
caXchange/ESB
10
Local ? National Grid
CTOM
Grid Services
NATIONAL
Credentialed caGrid User
11
(No Transcript)
12
caBIG Data Sharing Happening Now The
Inter-Prostate Biomarker Study (IPBS)

• 11 Prostate SPORE Sites Exchange Samples, Test
  Biomarkers
• Use caTissue Core/Clinical Annotation Engine to
  Track
• Retrospective Arm Samples in hand, limited to
  data previously collected
• Prospective Arm New accruals, sites jointly
  decide data to collect
• Prostate SPORE sites jointly decided to use
  caTISSUE as medium for sharing

13
TCGA Data Sharing(graphic Martin Ferguson, Ph.D)

1. Clinical data are collected / generated at
   contributing site. Sent to BCR.
2. BCR re-formats and harmonizes data into caBIG
   compliant CDEs. Sent to DCC.
3. DCC integrates clinical with genomic data. Makes
   available to research community.

• Research Community

• Genome Sequencing and Characterization Centers

• Clinical contributing site

• Biospecimen Core Resource (BCR)

• Data Coordinating Center (DCC)

14
Enterprise Deployment

• Rolling out caBIG

15
caBIG Enterprise Bundles

• An expedited path to the use of caBIG tools
• An evolution, not a revolution
• Three bundles are on the drawing board
• CTMS Applications bundle
• Clin. Trials apps on Local caGrid Within the
  organization
• caGrid Applications bundle
• Apps from all caBIG domains, on National caGrid
  Across the research community
• DSIC bundle
• Bundle of policies, best practices and SOPs,
  agreement templates

16
New caBIG Program Components

• caBIG Enterprise Support Network
• Service Providers
• Knowledge Centers
• Program Offices
• Enterprise Adopters
• Leslie Derr will present tomorrow afternoon

17
caBIG Data Sharing and Intellectual Capital
(DSIC) Workspace

• DSIC Mission
• Facilitate data sharing between and among caBIG
  participants by addressing legal, regulatory,
  policy, proprietary, and contractual barriers to
  data exchange
• Afternoon breakouts DSIC will work with us to
• Identify usefully shareable data delivered by
  CTMS applications and projects
• Identify potential non-technical issues that may
  present obstacles to the sharing of this data
• Goal DSIC will leave this meeting with a list of
  work required to resolve these issues

18
Q A
19
Enterprise InteroperabilityOutcomes Research
Sample Questions

• Response
• Do men with early stage prostate cancer suffer
  complications after treatment with Agent x has
  ended?
• Disease Recurrence
• What is the average time to disease progression
  (in days) for all patients on breast cancer
  trials? What is the most common reason for
  progression?
• Survival
• Is a high BMI associated with lower survival
  among breast cancer patients? Is there a
  correlation between BMI and disease stage?

20
Enterprise Interoperability

• Outcomes Research
• Data mining to identify underlying trends and
  effects
• Trial cohort estimation
• How many trial participants (in our institution,
  cooperative group, nationwide) have these
  characteristics?
• Is a proposed trial likely to accrue enough
  patients?
• Mining of (e.g.,) adverse events for safety
  monitoring
• Cross-protocol analyses
• Identifying safety issues early
• Enterprise data sharing is the point of caBIG

21
Enterprise Interoperability The Clinical Trials
Object Model (CTOM)

• Reference Implementation (UML) of BRIDG
• Set of Java APIs (using caCORE SDK) to access and
  serve data in a standardized format from multiple
  clinical trial data source systems
• Piloted with CTMS (Theradex) data

CTOM APIs
CTOM
C3D
Other supporting CT data
CTMS Files/views

• Other supporting CT data
• Eligibility criteria questions
• Site data
• Investigator data
• Additional Study data

CDMS 1
CDMS 2
22
Potential Future NCI ClinicalServices
Architecture
Site CTMS Apps
Community Sites
caAdapter
NCI Enterprise Clinical Services
HL7 V3/BRIDG

NCI Application Databases
NCI Applications
Exposed Web Services
Internal Web Services
NCI Legacy Divisional Apps
Protocol Services
AE Services
People/Org Services
Patient Registration Services
CRF Services
CTEP/Theradex Reporting Services
ESB
ESB
Grid Services
NCI Enterprise QC
CTMS
CTOM
Operational
Summary
caAdapter
NCI
SDTM
Janus
FDA
23
SYNTACTIC
caBIG Compatibility Guidelines
24
Special Interest Groups and Projects

• Reporting/
• Sharing
• Clinical Trials Database
• Routine Data Exchange
• Clinical Trials Object Model
• Janus (FDA Repository)
• Adverse Event Reporting and Collection

• Planning/
• Monitoring
• Investigator and Site Credential Repository
• Study Initiation Tool
• Protocol Lifecycle Tracking
• FIREBIRD
• DCP/DESK

• Conduct
• Standardized Case Report Forms
• Cancer Central Clinical Database (C3D)
• Participant Registry
• Laboratory Interface
• Financial/Billing
• Study Calendar
• Subject Prescreening
• Vendor Systems

• Interoperability
• System Interoperability and Harmonization
• Structured Protocol Representation (BRIDG)
• Clinical Trials Interoperability Project