caBIG Clinical Tools

1
caBIG Clinical Tools

• John SpeakmanAssociate Director, Clinical Trials
  Products and Programs
• Center for Bioinformatics and Information
  Technology
• National Cancer Institute

Joint NCRI Informatics caBIG Conference London,
12th July, 2007
2
Nixons War on Cancer (1972) Clinton-era
doubling of NIH budget over five years
Explosion of Knowledge about Molecular Basis of
Cancer
3
Translating Molecular Understanding to Patient
Outcomes how are we doing?
Information
Time
(from Robert C. Bast, Jr., MD)
4
FDA (Innovation or Stagnation, aka Critical Path,
Woodcock 2004)

• Explosion of knowledge from biological research,
  continued growth in research spending
• Slowdown in innovative therapies reaching
  patients
• as measured by the number of new submissions to
  FDA
• The drug development process has not kept up
  with basic scientific innovation
• 8 of drugs entering Phase I get to the market
• For drugs entering human clinical trials for the
  first time between 1989 and 2002 the average
  cost per new drug was 868 million in a range
  from 500 million to more than 2,000 million
  
• Health Affairs 25, no. 2 (April 2006)

5
Slowest transfer by far is not between domains
but within Clinical Trials
Clinical Trials
6
Clinical Information Integration Challenges
7
Ideal Clinical Research IT Infrastructure an
end-to-end pipeline
Clinical Systems
External Reporting
Clinical Trials
8
Inside the Pipe
External Reporting
Clinical Systems
Clinical Trials
TranslationService
etc.
HL7-v3, Janus
HL7-v3, Janus
HL7- v2.x,other
Labs, Pharmacy, etc.
HL7- v3
Lifecycle Management
ClinicalResearchInformation Exchange
HL7 trans-actionaldatabase

Adverse Events
FDA
Participant Registry
Pharma
EDC
NCI
Clinical Data Mgmt
other
Electronic MedicalRecord
ResearchDataWarehouse
De-identification Services
9
Industry Approach Electronic Data Capture (EDC),
aka Remote Data Entry (RDE)

• 2005 82 of industry-sponsored clinical trial
  data is captured using paper Case Report Forms
  CDISC
• EDC 1 Laptop computer
• Preconfigured with software, delivered to site
• Sponsor preference 1 (simplifies technical
  support)
• Sites last preference (limited real estate for
  all those laptops)
• EDC 2 Thin client software
• Loaded onto existing computers at site, not local
  data store
• Sites prefer, but central IT issue with software
  installation
• EDC 3 Zero footprint browser-based software
• Access CRFs online using a web browser sites 1
  preference
• NCI-sponsored trials have employed 2/3 for many
  years
• Because were stingy

10
Site-based Clinical Trials Database Systems at
Large Cancer Centers

• Designed to provide institutional accountability
• All trials irrespective of sponsor
• Send data direct to NCI
• Could send data direct to industrial sponsors
• Requires lengthy per-trial negotiations
• Usually agreement is not reached
• Otherwise (i.e., usually) data must be abstracted
  from the site system and entered into sponsor
  forms by hand
• Rekeying is the norm
• New learning curve for each new sponsor data
  entry system
• Site-based clinical trials database systems are
  not point-of care systems they are clinical
  trials databases
• Someone has already had to abstract data from the
  chart

11
EMR as Front End to Structured Data Sources

• Most of whats in an EMR is scanned-in scribbles
  and dictated clinic notes
• May also contain copies of the real structured
  source data
• Lab, Admission/Discharge/Transfer (ADT), OT,
  Pharmacy
• Even if it does, may be preferable to go to
  source (faster)
• Result direct pairwise interfaces between
  clinical source systems and clinical trials
  systems
• Significant human resources required to build and
  maintain
• Majority of clinical sources do not produce
  standard data
• Current HL7 standard, version 2.x, allows local
  extensions
• HL7 version 3 eliminates this problem, but not
  yet in widespread use, at least in USA

12
Starting to Lower the Barriers caXchange

• caBIG Data hub
• Transforms non-standard clinical source data into
  standard HL7 Version 3 data for delivery to a
  clinical trials database
• Centers will only have to specify metadata
  (rules for local data)
• Doesnt eliminate the barrier, just lowers it
• Currently limited to laboratory data
• Lab data was born electronic, i.e., never keyed
• Lab data constitutes majority of all clinical
  trial data points
• However caXchange is extensible to any input data
  format
• Or output format, for that matter
• Goal eliminate need for proliferation of
  expensive pairwise custom interfaces

13
Reinventing the Electronic Medical Record

• Change in the fundamental approach to the EMR
• data in it must be structured and interoperable,
  but
• ability of a clinical care professional to
  quickly grasp the narrative of a patient must
  not be diminished
• given advances in visualization techniques,
  should be possible to enhance it
• Being pursued as part of the US Nationwide Health
  Informatics Network (NHIN)
• caBIG and others are active to ensure that the
  needs of clinical research stay high on the NHIN
  agenda
• Pay-for-Performance is a development we can
  harness
• Way station Pay for Data

14
Paths Through the Pipe
Abstraction required No abstractiion required
Site (clinical operations)
Site (research operations)
Sponsor
15
What are we doing in the meantime?
caBIG Program Management and Contracting
Strategic Planning
Integrative Cancer Research
In Vivo Imaging
Tissue Banks Pathology Tools
Clinical Trials Mgmt Systems
Training
Data Sharing Intellectual Capital
Vocabularies and Common Data Elements
Architecture
16
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

17
Special Interest Groups and Projects

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

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

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

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

18
Library of Standard Case Report Forms based on
common data elements

• Admission end-to-end pipeline is not imminent
• Data must be entered de novo into sponsor
  database via CRF (paper or electronic, standard
  or non-standard)
• Initiative of NCI Clinical Trials Working Group
  (2005)
• Standard case report forms would reduce the
  need for investigative sites to manage a wide
  array of different forms and data entry
  processes
• Standardized data capture facilitates cross-trial
  analysis
• Maximize capture of critically important data
• Simpler regulatory review
• One of multiple current initiatives
• e.g., CDASH Industry and FDA
• Yes, we are talking to each other

19
caBIGs Special Sauce Modularity Implies
Interoperability

• Building for the next ten years Need to
  interface rapidly with new data sources and
  destinations
• Only a set of interoperable modules is agile
  enough to handle the
• speed of science
• Anyone can build a
• module that plugs in if
• they build to published
• caBIG standards
• Modularity implies
• interoperability

20
caBIG Clinical Trials Suite

• The caBIG Clinical Trials Suite brings together
  a range of interoperable tools supporting the
  clinical trials enterprise.
• Functions include
• Patient Study Calendar (PSC)
• Cancer Central Clinical Patient Registry (C3PR)
• Adverse Event Reporting (caAERS)
• Laboratory Data Integration (caXchange and Lab
  Viewer)
• and integration with commercial clinical trials
  data collection tools

Clinical Trials Suite
21
Clinical Trials Suite Participant Registry (C3PR)

• Tracks subject registrations to clinical trials
• Verifies registration criteria (study open,
  participant eligible, consent received)
• Stratifies subject into a stratum group,
  randomizes to the trial
• Tracks participants across sites (handles
  multi-site trials)
• Manages study personnel
• Reporting (federal/local requirements, supplies
  NCI Summary 3/4 data)

22
Clinical Trials Suite Adverse Event Reporting
(caAERS)

• Adverse Event Tracking and classification using
  accepted standards
• Data import / export AE data in common/required
  formats
• Automated, rules-based assessment of seriousness
  and reporting requirements (sponsor-level,
  institution-level and protocol-level rules)
• Reporting, including generation of CTEP, DCP, and
  FDA compliant reports
• Ability to submit electronically to CTEP AdEERS
  system

23
Clinical Trials Suite Patient Study Calendar
(PSC)

• Automatically generated study-template-based
  patient calendar
• Accurate, versioned representation of study
  parameter table
• Prospective forecasting of patient visit
  information
• Management of study participant schedules
  (schedule, reschedule, cancel)
• Retrospective outcomes review and reporting of
  calendar activities
• Consent / reconsent notification and tracking

24
Clinical Trials Suite Clinical Source Data
Integration (caXchange)

• Enables automatic transfer of clinical data from
  point-of-care systems in medical centers, e.g.,
  clinical chemistry lab systems
• Translation of multiple source data formats into
  standards-compliant data for use in clinical
  trials the Rosetta Stone of biomedical data
• Incorporates caAdapter mapping and translation
  tool to enable translation of any non-standard
  source and destination format

25
Clinical Trials Suite caXchange Lab Viewer

• Allows viewing of clinical lab data imported from
  clinical chemistry and other lab systems (e.g.,
  data extracted by caXchange)
• Search by MRN and date range
• Labs can be selected for loading into clinical
  trials databases and/or adverse event reporting
  systems
• Automatically flags lab values that may indicate
  toxicities

26
Building the Clinical Trials Suite

• Show that these modules can reliably exchange
  information and functionality
• without a huge harmonization overhead
• Complex clinical workflows dictate that the
  modules must interoperate at both a syntactic and
  semantic level
• Subject matter experts
• define and prioritize use cases
• Analysts
• describe use cases as workflow items with data
  descriptions
• Developers
• translate this into working software
• Architects
• define overall governance and rules of the road

27
Interoperability Use Casesidentified by Domain
Experts

• Enroll Subjects
• Load Lab Data in Clinical Data Management Systems
  (CDMS)
• AE Triggered Schedule Change
• Re-Consent Subjects
• Identify Potential Dose Change
• Quantify CRA Workload
• Study Specific AE/SAE Summary
• Cross Study AE/SAE Summary
• Extract Data for Interim Analysis

28
CTMSi Architecture Modular SOA for
NextGeneration Clinical and Translational
Research
BRIDG
29
Take-away Its not rocket science

• Most of it isnt even computer science
• Other industries established standards-based data
  infrastructures years ago
• But it is hard
• Medical data is complex consensus on semantics
  is a struggle to achieve

30
Thank you! Q A
john.speakman_at_nih.gov