How can data standards help “improve medical research and related areas of healthcare,” the goal as stated in the Clinical Data Interchange Standards Consortium (CDISC) mission statement? The answer is simple. Data standards can improve medical research and related areas of healthcare by:
1. Reducing—hopefully eliminating—the time industry loses in the clinical development process typically spent organizing and harmonizing years of clinical research data.
2. Facilitating a standard review environment, both for industry and regulatory agencies.
3. Facilitating the evolution of standardized analysis tools.
4. Facilitating regulatory agency marketing application reviews, Q&A and eventual marketing approvals.
5. Facilitating FDA performance of complicated meta-analysis across multiple products, which in turn will allow the FDA to strengthen its partnership with industry towards achieving its primary mission, to protect and promote public health.
Maybe the answer is not so simple, but industry has begun to realize that the CDISC Study Data Tabulation Model (SDTM) can help industry achieve all those goals, and more. Dr. Steve Ruberg, current chairman of the CDISC Board of Directors, noted in February 2002, “The essential kernel of the whole clinical development process is the data. Thus, without a data-centric approach to developing any e-clinical solution, we are unlikely to be fully successful. The data is the foundation on which we build our entire effort.”
Since that time, the SDTM has quickly become the nomenclature and structure/format foundation of this data-centric approach. While not necessarily the format for collection of all the data, it is the currently recognized format for submitting clinical and preclinical trial data to the FDA to support product marketing application review and approval.
The SDTM originated as the Submission Data Model (SDM), developed by the CDISC Submission Data Standards (SDS) team. The SDS team is comprised of about 30 volunteers from industry and three key FDA observers representing the Offices of Information Management, Statistics and Medical Review. Industry representatives come from quite a cross section of industry, including most major Pharmas, numerous small to mid-level Pharmas, Contract Research Organizations (CROs) and service providers. The SDS team collaborates regularly through bi-weekly teleconferences, quarterly face-to-face meetings and literally thousands of regular email communications, all in an effort to support the CDISC mission, to “. . . improve medical research and related areas of healthcare.”
The SDTM and the Submission Data Standards v3.1 (SDS v3.1)—the SDTM implementation guide for Human Clinical trials—is revolutionary in the industry as this is the first standard ever developed for submitting all data collected in clinical and preclinical trials. While previous versions of the standard were revolutionary in their own right, there was a major shortcoming in their long-term potential. They focused primarily on common safety data—essentially the core data routinely collected in most trials—but did not address all the other data collected in the clinical trials. This resulted in each sponsor having to choose its own modeling approach for the all the remaining data, leading to an unlimited number of possible formats.
While the CDISC SDTM is fairly well known in industry, it is still a fairly new initiative. The preliminary draft version was published in June, 2003 as the CDISC Submission Data Domain Models Version 3.0, better known as SDS v3.0. The official first version was published as the SDTM v1.0 and SDS v3.1 in June, 2004. During that one-year period from June 2003 to June 2004, there were a number of enhancements leading to the final approved version. A joint FDA and industry pilot to test a preliminary version was completed in early 2004, which was followed by two public review/comment periods. Once the production version of the SDTM was published, the watershed moment was when the FDA referenced it in eCTD Guidance in July, 2004 as the standard for submission of clinical trial data.
The FDA is helping to reinforce the strategic value of the SDTM in numerous ways. They recently recognized the SDS team with the Commissioner’s Leveraging/Collaboration Award in June 2005. Additionally, the FDA began announcing its Critical Path agenda earlier this year. Dr. Janet Woodcock, acting Deputy Commissioner for Operations at the FDA, presented the Critical Path agenda to the CDISC board of directors in January 2005. Dr. Randy Levin, Director for Health and Regulatory Data Standards in the Office of the Commissioner and current chair for the FDA data council, first presented the key components of the critical path agenda, relative to CDISC initiatives, at the FDA public meeting about the CDISC SDTM on February 1, 2005. Recently, at the June, 2005 DIA Annual meeting, FDA representatives made numerous presentations recognizing the strategic importance of CDISC initiatives in the critical path agenda.
From all these indications, CDISC’s overall initiatives are clearly a significant component of the FDA’s Critical Path agenda, with specific focus on the SDTM and future mapping of the SDTM into the CDISC Operational Data Model (ODM) for eventual submissions of the SDTM in XML, to replace the current requirement for SAS v5 transport files. This was a clear reinforcement of the strategic benefits the FDA sees in CDISC and the SDTM.
The immediate benefit for sponsors is that they are not required to submit data listings or patient profiles in marketing applications when submitting data in the SDTM, and they may be able to leverage the SDTM further by negotiating additional submission exceptions with their review division on a case-by-case basis. While these opportunities can be a tremendous immediate benefit, they do not even scratch the surface of the bigger picture, the short-term and long-term strategic benefits.
In the short term, many sponsors are starting to look to the SDTM to set the foundation for the future. They are realizing that there are significant benefits to adopting a standards-based vision supported by industry consensus and FDA acceptance, instead of internal practices. Organizations that have multiple standards in place (possibly due to mergers or acquisitions) are looking to the SDTM to be the motivator for change. Organizations that have either no complete standard in place yet, or multiple standards that need to be harmonized, are starting from the premise, “Why shouldn’t we use the SDTM?” And, yes, while there are many skeptics and critics, none have been able to make a compelling argument why the SDTM should not be implemented, and none have been able to provide a better alternative. Criticism alone, without providing substantive alternatives, generally has no influence on such a trend.
Many sponsors converting legacy data over to the SDTM for submission to the FDA consider this an incremental step to a much more grand vision, where all data that they exchange with regulatory agencies, partners, CROs and service providers is standardized and exchanged seamlessly. They are moving away from the “sand-castle” approach of preparing submissions, where practices used for one submission are eventually washed away by the waves; they are moving to a repeatable standardized process that supports both the FDA reviewers and internal sponsor reviews. Sponsors are looking to capitalize on the FDA review and eventual mining vision by implementing similar approaches internally, fed by the standardized SDTM structures.
Sponsors who are already implementing the SDTM are realizing additional benefits with their developmental partners that are also implementing the SDTM. Terminology, approaches to data retrieval, and approaches to data analysis are all becoming more standardized between partners because of the common foundation. This facilitates better communication and exchange/sharing of processes/procedures that can be applied to multiple projects with little or no modification. Analysis and reporting routines can be developed far ahead of actual data availability, since the SDTM provides the specification for the eventual data format. All of this, of course, results in time and cost savings, although specific research results are not yet available to quantify the savings.
For its part, the FDA has been developing the JANUS warehouse as the repository to store all the submitted data. It will provide a stable foundation for FDA standard review tools. Much of the review functionality the FDA envisions as necessary to help reviewers expedite their reviews has already been built into the standard review tools. The tools have been developed through Cooperative Research And Development Agreements (CRADAs) between the agency and various vendors. These projects focused on functionality first. Tools were later converted to support the production version of the SDTM once it was published.
Efforts are well underway to harmonize the SDTM with other CDISC and Health Level 7 (HL7) initiatives. The Trial Design models in the SDTM are being harmonized with the HL7 Protocol Standardization Group. The CDISC define.xml standard is now available as a replacement for the traditional define.pdf, allowing much greater flexibility in the metadata describing the data content. The CDISC Terminology team has been redesigned to work across all CDISC teams to develop standardized, controlled terminology across all the models. The CDISC Analysis Data Modeling (ADaM) team has developed a number of Analysis level standards, using the SDTM as the foundation. A sub-team has formed with representatives from the ADaM and SDS teams to jointly propose a best-practice approach to producing and submitting Analysis Datasets and/or Analysis logic, and to propose standard Analysis models harmonized with the SDTM. A sub-team has also formed with representatives from the SDS and ODM teams to map the SDTM into ODM. And the list goes on and on. What does it all mean? CDISC and HL7 teams are working closely together towards a common endpoint, standardization and harmonization of all regulated clinical and preclinical research data.
All of these efforts are now allowing industry to clearly see the potential long-term benefits. Once the SDTM is mainstream for sponsors supporting the submission of marketing applications, achieving additional benefits is simplified because the data is already standardized. The FDA’s acquisition of a standard toolset for reviews would support data being submitted earlier in the process, potentially to the Investigational New Drug (IND) application at the time the Clinical Study Reports (CSRs) are submitted. This would facilitate FDA review of individual CSRs well ahead of the submission of actual marketing application(s). It also supports the agency’s ability to develop cumulative pooled analysis as the clinical development pro-cess progresses. Additionally, this would allow sponsors to minimize their own typically extensive end-stage effort of preparing and submitting all the individual CSR databases at the time of the marketing application. Sponsors can now focus more time on the pivotal trials necessary to support the primary endpoints and the Clinical Summary of Safety (CSS) and Clinical Summary of Efficacy (CSE) and/or the traditional Integrated Sum-mary of Safety (ISS) and Integrated Summary of Efficacy (ISE).
Sponsors who take this approach will simultaneously be supporting internal efforts to develop their own standard warehouse and reporting tools, helping to achieve similar efficiencies in their own clinical development program. Once data for completed studies is routinely being submitted to the IND with the CSR, additional efficiencies will become apparent. Would the FDA be interested in developing a standard review toolset for IND annual reports based on subsets of SDTM data? Would industry be interested in this approach and/or the same toolsets? Could this simplify the process for everyone? These are all questions that will most likely come into clearer focus as industry gains experience with these standard data models.
Most importantly, however, is the long-term strategic value of helping the FDA help industry develop safer more efficacious products. Achieving the FDA vision of mining the most robust database of clinical trial data anywhere in the world has benefits that we cannot even quantify at this time. How much better will industry be—and more importantly, how much better will healthcare be for patients—when the FDA achieves this goal of strengthening its partnership with industry? What is the potential individual and corporate value to the FDA helping industry identify safety concerns prospectively? Only time will tell, but all indicators are extremely positive that we are on the right track to success.
Tom Guinter is vice president, Clinical Data Strategies at Octagon Research Solutions, Inc., a Development Partnering Organization that delivers regulatory, clinical, process and IT expertise to drive innovative advancements in the drug development process. He can be reached at tguinter@octagonresearch.com.