You tell ‘em: FDA Meeting on Study Data Exchange Standards
Written by Joel Finkle, Senior Strategist, Regulatory Services Delivery
Out with the bad air, in with the good
The SAS XPORT format that has been in use since 1999 (before the eCTD era) is long in the tooth, doesn’t include metadata to show data relationships (which is why we have a complex define.xml file to describe it separately), and in fact the date stamp on the XPORT file isn’t even Y2K compliant! (the data within the file is, though)
So FDA is holding a meeting on November 5, 2012 for discussion on study exchange standards, and what could be used to replace them. There’s limited capacity (300) so registration is required. Comments can also be entered on the Regulations.gov website: http://www.regulations.gov/#!documentDetail;D=FDA-2012-N-0780-0001 or sent in written form to
Division of Dockets Management (HFA–305)
Food and Drug Administration
5630 Fishers Lane, rm.1061
Rockville, MD 20852.
Be sure to list Docket No. FDA–2012–N–0780.
Send registration information (including name, title, firm name, country of citizenship, address, telephone and fax number, and email address) to
Fatima Elnigoumi
Center for Drug Evaluation and Research
10903 New Hampshire Ave.
Bldg. 51, rm. 1195
Silver Spring, MD 20993,
301–796- 4863
email:CDERDataStandards@hhs.fda.gov.
What they want to know
There are ten questions on the docket, which can be found here: http://www.gpo.gov/fdsys/pkg/FR-2012-08-14/html/2012-19748.htm
FDA wants to know what’s causing problems with current study data management (not just exchange of data), what tools and data standards are currently used, whether HL7, CDISC or open standards would be preferred in the future state, and what sort of timeframe should be used.
It’s great to see this level of openness from FDA. I remain skeptical as to whether we’ll escape the Big Two of HL7 and CDISC for standards, or whether those standards would be changed much. Coming from a long history in information technology,
What’s wrong with the old way? And get off my lawn!
I still find XML-based data standards horribly verbose – we could never have handled the extra text space needed on the first systems I’d worked on – but the world of data exchange has certainly upgraded since those days. Self-describing data formats such as XML help keep data reliable (nothing like a comma in a comment field to mess up your CSV files, right?) and validatable.
FDA’s asking whether HL7 or CDISC formats should be used is something of a “magician’s choice”: they’re not quite the same now, but projects within both organizations exist to unify them, and they’re approaching that rapidly (keeping the verbosity of the HL7 standards, unfortunately). The taxonomies and vocabularies are pretty much there today. It’s pretty likely that something built on this “BRIDG” model will be what is submitted in the future – but at least the general public has the opportunity to give some input..