When rumors increase, and when there is an abundance of noise and clamor, believe the second report. Alexander Pope, 1688–1744

In recent weeks, there have been a number of rumors about the Clinical Data Interchange Standards Consortium CDISC standards and their continued use by the Food and Drug Administration (FDA). These whispers suggest that the Study Data Tabulation Model (SDTM) standard is being replaced and that the Operational Data Model (ODM) should no longer be used. The rumors have caused confusion at best and fear of standards abandonment at worst. This is a second report—an attempt to clarify what’s happened and how the industry should move forward.

CDISC, to state the obvious, is a global standards organization. While the FDA plays a vital role in the drug development industry in the U.S., there are other global stakeholders that use CDISC standards. CDISC’s influence extends as far east as Japan, Australia and China—and is well established in Europe.

An International Agenda

Just as biopharmaceutical communities are expanding globally, CDISC is being asked to meet with regulatory authorities in China and with academic representatives in Singapore, India and Brazil. CDISC has Liaison A status with the International Standards Organization Technical Committee 215 and works closely on global standards harmonization with CEN, ISO and HL7 as a member of the Joint Initiative Council.

The FDA, naturally, receives and processes vast amounts of information. This information is currently submitted in a variety of formats: SAS transport files (XPT), HL7 V3 messages and Excel files. The FDA has stated that, going forward, HL7 V3 messages are the preferred electronic exchange format for health care information. HL7 V3 messages are already in use to carry content to the Janus data warehouse, the structured product label (SPL), the integrated case safety report (ICSR) and regulated product submission (RPS) messages.

Currently, the FDA eCTD guidance specifies the use of SAS transport files for the transport of clinical observations to the agency. This method has a major limitation. The flat file format of SAS transport files does not inherently capture relationships between study data or between study data and study design. Adding these relationships post-facto is invariably incomplete, inefficient and inconsistently handled.

Long Live SDTM

As a result, as FDA has stated since 2004, it would like to move away from the SAS transport file format towards a more robust exchange format for clinical observations that inherently relate the observations both with each other (such as the HL7 ICSR) and with planned observations at the point of data collection. This will ensure such data can be reliably and consistently conveyed to FDA information systems now and in the future. FDA recognizes that currently these important relationships are not often captured (or are captured inconsistently) at the point of data collection. As electronic health record systems (EHRs) come into more widespread use, the opportunity to capture such relationships automatically at the point of data generation and collection will increase.

So why the sudden flurry of concerns? Why the muddying of the waters?

The FDA has stated that SDTM will be used for the content for electronic submission of data tabulations. For the next four to five years, SAS transport files will be the mechanism to transport the SDTM data to the agency. At that point, there will be a transition in the way that the SDTM is transported to the FDA from SAS transport to HL7 V3 messages. SDTM content will not disappear. It may well evolve during this time and be improved, but it will not disappear.

By aligning the SDTM content with the HL7 transport technology and the HL7 Reference Information Model (RIM), CDISC and the FDA will improve the SDTM standard and provide industry and regulators with a more robust content standard.

Indeed, SDTM is so well received by FDA reviewers that they have become familiar with its structure and have written a two-year contract to support internal training. CDISC will ensure FDA personnel have adequate training in the use of SDTM, ADaM and advanced CDISC topics; the training will be delivered using face-to-face as well as computer-based virtual learning methods. This contract is a testament to the SDTM standard.

Future ODM Role

The work on developing the HL7 V3 messages to transport the SDTM has only just started, and there are many CDISC team members participating in the process to ensure an optimal transition. The content of SDTM will evolve but will remain stable and consistent, allowing the industry to remain focused on what is important—the science.

It is true the FDA canceled an ODM pilot project investigating the use of ODM for the submission of eCRFs to support the regulatory review process. Given the transition to HL7 V3 messages, this decision makes sense at FDA and as a global scientific policy.

But the CDISC ODM is not on the sidelines or, as rumored, “dead.” Far from it. FDA, for starters, is not opposed to the industry using ODM for other purposes. The CDISC ODM standard has many valuable uses at the front end of the medical research process. It is the best and currently the only means of exchanging case report form (CRF) data and its associated metadata, in particular the recently published CDASH standard (an FDA Critical Path Initiative).

Looking Ahead

Tools which employ ODM can be certified. To date, two of the largest vendors (Medidata and Phase Forward) have achieved certification. Others are hoping to follow, as are a growing set of sponsors that are demanding such certification from their vendors. When the industry is looking for a method of exchanging data, archiving data, keeping an audit trail secure while ensuring compliance to the relevant regulations, ODM is the only standard method of doing this. Period. ODM is based on robust XML technology and is a future-proof standard no matter what happens in the EHR landscape.

The take-home message is fairly simple. CDISC wants to see the separation of content from transport of the data. CDISC wants to ensure that industry investment in data standards is protected. How to achieve this? By developing the Biomedical Research Integrated Domain Group (BRIDG) model. It will harmonize all of the existing standards, uniting them as one. BRIDG represents the clinical research domain in the context of the HL7 Reference Information Model (RIM); it was initiated at the recommendation of HL7 experts.

Moving Forward

CDISC had the foresight to form a collaborative relationship with HL7 in 2001. Since then, it has approached clinical research standards development as a global process. For the CDISC global community, BRIDG is vital to ensure that any particular data point can flow seamlessly through the entire medical research process and not just be submitted to a regulatory authority.

In the forthcoming months, an updated CDISC technical roadmap will be available with the aim to continue to strengthen and build global integrated standards that will support clinical research and health care, as well as regulatory authorities worldwide. Maintaining CDISC’s collaborative teamwork will enable the CDISC technical strategy to ensure consistency across the CDISC standards, stability of the CDISC standards and their evolution in ways that accommodate both industry and government.

Rebecca Kush is CEO of CDISC. Frank Newby is COO of CDISC. David Iberson-Hurst is VP of technical strategy at CDISC. And Amanda J de Montjoie is communications specialist at CDISC.