The latest in a recent stream of data accessibility statements from funding agencies and the scientific community comes from the National Institutes of Health, which has updated its policy for sharing research data.
The new policy, available at http://grants.nih.gov/grants/guide/ notice-files/NOT-OD-03-032.html, updates only slightly a draft policy that had been in effect since March 2002. The new version now includes the provision that applicants seeking $500,000 or more in direct costs in any year of a project are expected “to include a data-sharing plan in their application stating how they will share the data or, if they cannot share the data, why not.”
The agency has also revised its definition of “timely release and sharing” of data to be no later than a dataset’s acceptance for publication.
“Scientists working in many different areas are already sharing their data through a variety of mechanisms,” Belinda Seto, acting deputy director for extramural research at NIH, explained via e-mail. “However, some disciplines are less familiar with this process and associated processes. If federal funds are used to support research, including data collection, it is reasonable to expect that the final research data be made available to others, if possible, to allow for the replication of research findings, and to allow analyses to be conducted by other researchers, thus maximizing the returns on federal research dollars.”
NIH isn’t the only organization that felt the need to issue a friendly reminder about this topic recently. The revised statement follows quickly on the heels of NHGRI’s release of a draft document for an extended version of the “Bermuda principles,” which govern the release of genomic data (available at www.genome.gov/page.cfm?pageID=10506537).
The scientific community has also publicly weighed in on the issue of data release in recent months: A report from the National Academies of Science called for a set of ground rules that scientists, publishers, and funding agencies can follow when determining when and how to make published research results public [BioInform 02-17-03]; and at the February meeting of the American Academy for the Advancement of Science, 32 journal authors and scientists released a collective statement endorsing a self-monitoring policy for the publication of research results with potential applications in bioterrorism [BioInform 02-24-03].
The storm of policy statements highlights the central role that large data sets are playing in biological research — as well as the challenges the scientific community faces as it attempts to walk the fine line between scientific openness, public safety, and commercial opportunity.
The NIH policy comes very close to covering all bases: While encouraging the free release of data, it does permit organizations to charge “a reasonable fee” to cover preparation and distribution costs. Researchers can also request funds as part of their application to help cover the cost of developing user-friendly forms of the dataset.
In addition, a researcher can provide several valid reasons for not sharing his or her research data, including reasons of national security or in circumstances where a private co-funder for a project has requested restrictions on data availability.
In the case of Small Business Innovation Research grants, where commercial value is a condition of the award, recipients will be entitled to keep their data proprietary for four years. If anything, the new policy should help small database development companies, Seto said: “We would expect existing small businesses whose business is data archiving to be approached with more regularity by researchers seeking assistance with archiving their data and associated user support.”