CDISC & the World of Health Information Technology

The Clinical Data Interchange Standards Consortium (CDISC), a nonprofit organization based in Austin, TX, is on a mission. Its goal is to improve medical research and related health care through the development and support of platform-independent data standards that enable the interoperability of information systems. Since 2001, that pursuit has also included enabling the use of Electronic Health Records for clinical research.

To find out what CDISC is doing to achieve these goals and to learn more about the organization's capabilities, models, and relationships, ACT went straight to the source. Following is a discussion with four members of CDISC, with input from an executive in the software solution & services industry.

Q: CDISC is involved in a number of areas. Can you explain what those are?

A: CDISC has spent the past nine years developing and establishing standards that facilitate the acquisition/collection, exchange, reporting/submission, and archive of clinical research data along with the underlying standard terminology.

Some of the key areas of attention for CDISC are maintaining and enhancing the standards as well as providing education to support the many various users of these standards. In addition, CDISC has been developing complementary standards, such as a standard protocol representation, and also leveraging the standard for reporting case report tabulation data to define data collection standards that will benefit investigators, project leaders, and clinical research associates.

Another area that CDISC has been focusing on since 2001 is the harmonization of clinical research standards with the relevant standards from health care. This is also intended to benefit investigators by enabling the use of electronic health records (EHR) for clinical research, including regulated clinical trials. A formal relationship with Health Level Seven (HL7), the Single Source proof of concept, and a demonstration for Healthcare Information and Management Society (HIMSS) 2007 are examples of projects in the area CDISC calls "Healthcare Link."

A document developed by a CDISC team, upon encouragement from FDA, explains in-depth how electronic source data such as EHR information can be used in regulated clinical research today, in the context of existing regulations; Version 1.0 of the eSource Data Interchange (eSDI) document was posted in November 2006 on the CDISC Web site (http://www.cdisc.org/eSDI/forms/eSDI.pdf).—Rebecca Kush, Founder and President, CDISC

Q: Can you explain your relationship with the U.S. Food and Drug Administration?

A: CDISC has been working with FDA since 1998. However, the relationship between the two organizations must be one that maintains respective independence. For example, CDISC brings to FDA recommendations for standards. FDA takes these recommendations and determines their path in terms of adoption and guidance or regulation for the industry. FDA representatives are often liaisons and/or observers to CDISC teams, but are not team members and cannot join CDISC.

CDISC supports the FDA education and use of the standards as requested, and there have been a number of pilot projects done in collaboration. One such example is the recent Study Data Tabulation Model-Analysis Data Model (SDTM-ADaM) pilot, where at least a dozen FDA reviewers participated in a mock submission review created by CDISC teams. The pilot's outcome was to understand and facilitate the way submissions are developed and reviewed using the CDISC standards. An ODM pilot is to be announced early this year. CDISC and FDA continue to work together so that the standards can be effectively integrated into the FDA review process.—Rebecca Kush and Frank Newby, Chief Operating Officer, CDISC

Q: Have you formed any other relationships with the EU Guidance Group, the International Harmonization Committee, or EMEA?

A: Although CDISC has a close relationship with FDA (and FDA is the only regulatory authority to require that data be included in the submissions for marketing approval of a new product), CDISC also works with other regulatory agencies and organizations throughout the world.

CDISC has very active Coordinating Committees in Europe (E3C) and Japan (J3C). The European Agency for the Evaluation of Medicinal Products (EMEA) has been interested in the CDISC Protocol Representation standard (since certain fields are common with their EudraCT, the European Clinical Trials Database) and also in the eSDI document, since it analyzes the regulations relevant to the European Union and the recommendations within it can make it easier for field auditors to audit electronic trials conducted using electronic data capture (EDC) or eDiary data collection technologies.

The Japanese government funded a study in 2005–2006 to determine means to achieve clinical research digitalization. The report, written by a team led by a professor from the University of Tokyo, recommended the use of CDISC standards, particularly at hospital settings where clinical research studies are being done. These agencies are, therefore, interested in CDISC standards for the data acquisition and exchange value, not just the reporting/submission aspects.

The International Conference on Harmonisation (ICH) is a unique project that brings together the regulatory authorities of Europe, Japan, and the United States and experts from the pharmaceutical industry in the three regions to discuss scientific and technical aspects of product registration. CDISC cannot be a part of the ICH because ICH membership is limited to regulatory authorities and pharmaceutical manufacturers associations. However, a recent ICH decision was made to work with external standards groups such as HL7. Since CDISC brought the clinical research domain expertise to HL7 and established that relationship through an HL7 Technical Committee (RCRIM: Regulated Clinical Research Information Management) formed with FDA/CDISC/HL7 leadership, this collaboration will, by default, include the CDISC initiatives.—Rebecca Kush

Q: How is CDISC furthering HIT initiatives in the United States as well as globally?

A: CDISC has engaged the HIT community in three distinct ways, represented by three acronyms: BRIDG, RCRIM, and RFD.

BRIDG (Biomedical Research Integrated Domain Group). At the most abstract level, the BRIDG project—jointly governed by HL7, FDA, CDISC, and NCI—formally models the clinical trial space in a way that will harmonize with the HL7 Reference Information Model. The BRIDG project will enable software development (such as the various caBIG projects) and message development (such as the various RCRIM standards) to interoperate with other information systems both within the clinical trial space and across the divide between clinical research and health care.

RCRIM. This technical committee within HL7 focuses on messaging standards that involve both clinical research and health care. RCRIM adopted BRIDG as its domain model, and all new work from RCRIM will conform to this model. With joint leadership from HL7, CDISC, and FDA, RCRIM is well positioned to create standards that will enable interoperability between clinical research and health care.

RFD. Retrieve Form for Data-capture (RFD) is an IHE profile, the development of which was initiated by CDISC, and allows an electronic health system to host a guest data capture form from an external system. In the clinical trial space, RFD allows a case report form or a drug safety data capture form to be filled out by investigative site personnel working inside of an EHR session. RFD interoperability will be demonstrated this month at the HIMSS Interoperability Showcase. Participants include Pfizer, Eli Lilly, Novartis, Genzyme, Phase Forward, Outcome, SAS, SAIC, Sentrx/Relsys, SEC Associates, Assero/IPL, and Digital Infuzion from the clinical research side, and Allscripts, Cerner, Siemens, IBM, and Accenture from the health care IT side.—Landen Bain, CDISC Healthcare Liaison

Q: What is the BRIDG model and what does it have to do with clinical research and health care?

A: The BRIDG Model is a domain analysis model (DAM) representing protocol-driven biomedical/clinical research. A DAM is a set of requirements that are submitted for design and implementation procedures and processes. It is a structured way to describe and document the information requirements of a particular area of interest (or domain). The BRIDG Model was developed to provide an overarching model that could readily be understood by domain experts and provide the basis for harmonization among standards within the clinical research domain and between biomedical/clinical research and health care.

Three important streams of development have been brought together into this collaborative framework:

• CDISC—In 2003, CDISC started constructing a Domain Analysis Model to support harmonization of their models as well as harmonization with HL7.

• NCI—In 2004, the NCI's caBIG initiative joined the CDISC BRIDG effort to construct a structured protocol representation and to achieve interoperability among clinical trials research in cancer.

• HL7—In early 2005, the BRIDG model was adopted by the HL7 RCRIM Technical Committee as the RCRIM Domain Analysis Model, and is being implemented at NCI.

The model emerged from an unprecedented collaborative effort among clinical trial experts from CDISC, the NIH/NCI, the FDA, HL7, and other volunteers. The modeling effort is based on the HL7 Development Framework (HDF).

The BRIDG model serves to bridge standards, as well as organizations, and various communities, including academic research institutions and pharmaceutical product development organizations and related service and technology providers. It is also bridging the gap between clinical research and health care. For example, the FDA is currently using BRIDG for CDISC HL7 message development.

The model, which can be found at www.bridgproject.org, is represented in the Unified Modeling Language (UML), using a software tool called Enterprise Architect. A software package named SubVersion is used for model version control.—Julie Evans, Director, Technical Services, CDISC

Q: How can the CDISC standards help physicians who are involved in clinical research?

A: CDISC is positioned to greatly help physicians who participate in trials as investigators. Current practice demands that physician-investigators and their staffs interact with multiple data capture systems, often to absurd levels. RFD will permit physicians to use one system, their own EHR, to gather data for all clinical trials.

Another relevant CDISC-related project is the CDASH Initiative. CDASH, which stands for Clinical Data Acquisition Standards Harmonization, will enable the use of standards for data collection (e.g., Case Report Forms) across research projects and sponsors, which will most certainly help physicians by eliminating the variability in how data is collected.—Landen Bain

Q: Aren't CDISC and HL7 mutually exclusive standards?

A: CDISC and HL7 are separate Standards Developing Organizations, but CDISC and HL7 work together, along with the FDA, through the HL7 RCRIM Technical Committee. One of the results of this collaboration is the availability of the CDISC LAB standard in HL7 V3 format as well as ASCII, SAS transport, and CDISC LAB XML. Through RCRIM, CDISC and others provide the clinical research domain expertise to HL7 so that the clinical research standards are harmonized with the HL7 standards.

CDISC has an Associate Charter Agreement with HL7 that was originally signed in 2001 and renewed in 2004. The renewed agreement includes reciprocal organizational memberships and reciprocal advisory board positions, in addition to educational and financial opportunities and benefits for organizations that become members of both organizations. In addition, the renewed Charter included the formation of an Outreach Committee for Clinical Research (OCCR). This committee, which is comprised of members from FDA, CDISC, HL7, NIH, and other groups involved in clinical research, was appointed by the Board of HL7 to strategically represent the domain of clinical research within the context of HL7 and to reach out to the clinical research community to encourage harmonized standards development.—Julie Evans

Q: CDISC has developed and is developing a number of models for operational data, laboratory, study data tabulation, and analysis data models—can you explain the reason behind the models and the future of these models?

A: The CDISC production data standards were developed to address a variety of use cases for clinical data interchange. Since the requirements and constraints varied for each of these use cases, it was necessary to develop different standards for each. Specifically:

• The CDISC LAB model was intended to address transfer of bulk laboratory results from a central laboratory to a sponsor system, together with administrative information that is necessary to manage lab data. The LAB model can be implemented in either ASCII, SAS, XML or HL7 V3 messaging formats.

• The CDISC Operational Data Model (ODM) was intended for transfer and archive of study metadata and data for a clinical data management system (including Part 11–compliant audit trail) in a single package. The hierarchical structures and relationships of the metadata and data could not be effectively handled in a flat file structure (such as ASCII or SAS Transport format), so the standard was written in an extensible XML format.

• The CDISC Study Data Tabulation Model (SDTM), as represented in the SDTM Implementation Guide for Clinical Trials, was intended to be used to submit tabulation data to regulatory authorities in a generalizable format that would work with both legacy and prospective studies. Because FDA designated that the SAS V5 Transport format be used for all data transfers to the agency, the SDTM was designed to fit within the constraints of that format in a manner familiar to data managers and statisticians.

• The CDISC Standard for Exchange of Nonclinical Data (SEND) was adapted from the SDTM conceptual model to handle data from nonclinical animal studies, and was subject to the same format constraints as the SDTM for trials.

• CDISC Analysis Dataset Models (ADaM) were intended to define standards for submitting analysis datasets associated with a study in SAS Transport format.

More recently, CDISC developed the CRT Data Description Specification (define.xml), which uses the ODM standard to represent metadata associated with data submissions to the FDA. FDA's acceptance of this standard and willingness to accept future submissions in XML has made it possible for CDISC to more effectively integrate its existing production standards into a single, comprehensive specification not limited by some of the prior constraints.

The CDISC Technical Roadmap has thus been defined to harmonize these existing standards—along with the forthcoming CDISC standard for structured protocol representation—into a unified, comprehensive standard that will address the full lifecycle of clinical data flow from protocol design through data collection, analysis, submission, and archive, while retaining backward compatibility and traceability.—Wayne Kubick, Senior Vice President & Chief Quality Officer, Lincoln Technologies

Q: Where does HIT come into play with CDISC? What does CDISC have to do with HIT or EHR?

A: CDISC is the standards link to health care (Electronic Health Records) for the biopharmaceutical industry.

CDISC is involved at a national level to ensure that clinical research has a place in U.S. Health care IT initiatives.

• CDISC is actively participating in the ONCHIT-initiated Health Information Technology Standards Panel (HITSP) technical group and holds a seat on the board of HITSP. Specifically, CDISC is working to ensure that the information needs of the regulated research community are represented in HIT and other EHR initiatives.

• CDISC has international projects to support use cases and scenarios for using EHR in trials and pharmacovigilance.

• CDISC now has an approved integration profile, through the international organization Integrating the Healthcare Enterprise (IHE) and HIMSS. This integration profile (called RFD) is a vendor-neutral and platform-independent standard to provide a link for various applications into an EHR; it will currently accommodate four use cases, including EDC trials, biosurveillance, pharmacovigilance, and trial registry.

• CDISC has authored a document to encourage the use of eSource (including EHR) in clinical research in the context of existing regulations in the United States and Europe. Three of the five scenarios offered in this eSource Data Interchange document involve EHRs used for clinical research.

Of course, we invite PhRMA to sit down with CDISC to compare agendas regarding EHRs and to explore ways to support each other's objectives.—Rebecca Kush