Discussing Data Standards and Compatibility: A Day at the FDA
I recently attended and presented at the FDA public meeting held at the White Oak campus in Maryland to discuss solutions for study data exchange standards. The meeting provided an open forum for industry sponsors, technology vendors, standards organizations and the public to discuss the current exchange standards landscape directly with FDA stakeholders. In addition to presentations from key FDA leadership, participants were invited to speak on various topics concerning the exchange of data with FDA. I presented on the use of the CDISC ODM data exchange standard as an integral component to the exchange of data between Medidata’s Clinical Cloud and various clinical systems in use by our customers and partners across the industry. In 2008, Medidata was the first to be certified in all eight CDISC ODM use cases and has continued to stress the importance of using open industry standards to promote and facilitate inter-system data and metadata exchange. Medidata relies on its open, standards-based API's to support over 350 global sites and customers managing millions of study records on a daily basis.
Focusing the scope of the discussion proved challenging. Honing in on the specific set of requirements from the FDA was quite a large task, as the conversation quickly expanded beyond the use and challenges of using the current SAS xpt V5 transport for data submission to a discussion on how the current clinical domain space can be captured and modeled in a standard and uniform technical expression. The conclusion was simply that the current use of SAS xpt files was no longer sufficient to meet the increasing demand and requirements.
Standards across both HL7 (V3 messages) and CDISC were debated, with an obvious greater representation from CDISC supporters and users. In fact, beyond a single FDA pilot project presentation on the use of the HL7 CDA for documents, the CDISC standard community was clearly better represented.
There was immediate consensus on the fact that current two-dimensional views of data and metadata are not sufficient to provide the regulatory review process with the level of information they desire. The FDA is generally receiving more data than ever and the ability to harness the knowledge and meaning of that data is proving more and more difficult. More robust representation of the data, as well as a more robust model or definition of the data is needed. Topics such as ontologies and the semantic web were proposed as means of enhancing the current state of data and metadata.
Medidata has long been a proponent of the model-based approach to defining the clinical domain. In 2004, Medidata was the first to model the study design and protocol design into a robust, UML representation implemented via XML. Medidata was also the first to integrate the concepts of electronic data capture (EDC) and clinical data management (CDM) and provide them through a Software as a Service (SaaS) solution. It was gratifying to see the innovations brought to market by Medidata over the past decade validated against the needs of current regulatory and sponsor processes. The FDA is looking to industry to provide input into their data exchange road map over the next several years and Medidata was grateful to have this opportunity to share our experiences.
Read David’s recent blog post, “Protocol Standards are Statistically Significant: Presenting at a PhUSE Single Day Event on PRM.”