The FDA recently organized a meeting to discuss clinical data exchange standards between the FDA and the pharmaceutical industry. There were speakers from the FDA, pharma, and technology vendors. The good news is the FDA is receiving data in Study Data Tabulation Model (SDTM) format, although currently they only receive 35% of the data this way.
While the FDA is pleased that sponsors are using the CDISC SDTM standard, the main message at this event was that the FDA would like more insight into the origin of the data. They want access to audit trails and information regarding the data cleaning that was done. In addition, the FDA would like to find an alternative to the limited SAS Transport file that is currently being used to transfer data.
The FDA is in much better shape than they were five years ago when it comes to clinical data exchange. This is due in part to the 2009 creation of the FDA Computational Science Center (CSC). The CSC's mission is "To support Center for Drug Evaluation and Research (CDER) in continually improving the optimal drug evaluation and review process for the entire drug lifecycle." They do this by providing the following:
- Creation of an integrated review environment
- Bio-informatics/business intelligence platform
- Easy access to multiple useful review tools
- Data management support
- Review tool development support
In other words, they are the group supporting clinical data exchange standards to the FDA! Click here for an update on what the FDA Computational Science Center is doing. Click here for a list of standards recommended by the FDA.
What Goes On Behind the Scenes of Clinical Data Submission
According to Dr. Chuck Cooper of the CSC, the FDA knows there is a data cleaning, review, and coding process that goes on prior to the submission of the data. The FDA thinks it is important to have a view into these data cleaning activities; they are interested in understanding any changes to the data or any additions to the data such as during the medical coding process.
Currently, clinical trial sponsors adhering to CDISC standards submit a file called define.xml that describes all of the data and metadata being submitted. Dr. Cooper says the define.xml file is the first file a reviewer looks at to understand what is being received in the submission. One challenge the FDA is seeing is that these define.xml documents are not always reflective of the data being submitted.
The Issue with SAS Transports: Too Much Wasted Space
The FDA's Douglas L. Warfield, PhD, Division of Data Management Services & Solutions, spent some time talking about problems seen at the FDA by reviewers when it comes to dealing with submission datasets. Currently sponsors are required to submit their data in a SAS transport format which assures the file can be read by any version of the SAS statistical software.
According to Dr. Warfield, there is a lot of "wasted space" in these large files and this wasted space is "too costly." Like any good statistician, he presented us with a variety of statistics, charts and graphs showing the size of the same files in different formats, proving that the SAS transport files are larger than other formats. He wants to explore options for reducing the file size of these large files. This discussion took me back many years to when disk space was very expensive.
Bill Gibson from SAS talked about new SAS V5 transport file extensions that became available in October 2012. These extensions bridge the gap between the old SAS V5 transport files and SAS ODM. Some of the old limitations are gone, such as field names that previously were 8 characters can now be 32 characters. SAS Labels, previously limited to 40 characters can now be 256. Many people in the room including the FDA were unaware that this new option existed from SAS.
Overall, there was much talk about replacing these SAS Transport files with an agnostic type of transport file that also contains the metadata such as the audit trail. There was discussion around CDISC ODM and HL7 v3. There appeared to be overall agreement that CDISC SDTM is here to stay.
For me, it is always amazing to understand the vastly different issues facing the FDA and the pharmaceutical industry when it comes to clinical data exchange. While the FDA is struggling with old hardware, disk space issues, and a lack of training, the high-tech pharma companies want the FDA to provide guidelines and clinical data exchange standards for them to jump into semantic ontology-based systems. This is like asking the postal service to provide standards for instant messaging.
This is why companies like BioClinica are so important. They act as the middle ground between the FDA's mission of ensuring safety, efficacy, and security of drugs and the pharmaceutical industry's role in discovering and efficiently testing new treatments. By providing guidance to sponsors on realistic best practice data collection methods and clinical data exchange standards, BioClinica can help the industry achieve its goals while meeting the demands of the FDA.