PubMed: Release data voluntarily

As public access proponents and private-sector officials fight over influence in disseminating government-funded health information, the National Institutes of Health is trying to strike a balance with a new public access policy for the department's database.

Beginning May 2, PubMed Central, a free digital archive of published, full-text, NIH-funded research, will require scientists to voluntarily release manuscripts as soon as possible and within 12 months of scientific journal publication. The announcement was made Feb. 3, following months of deliberations.

The new public access policy applies to all recipients of research grants, cooperative agreements, contracts and National Research Service Award fellowships. It calls for scientists who conduct research with NIH funds to submit an electronic version of their final manuscripts, including modifications from the publishing peer review process, upon acceptance for publication. Then, authors can choose a time frame for public release.

Currently, articles on PubMed Central are available for a fee, by subscription to pricey journals or at medical libraries. Before this week's announcement, scientific publishers argued that the original plan — posting results six months after publication in a scientific journal — would cut into journal sales.

PubMed Central, launched in February 2000, is a permanent and searchable electronic repository. In the past, it was a voluntary archiving service for entire journals, with both non-NIH-funded research and taxpayer-funded research. Now, it is a voluntary service for grantees.

NIH estimates that 60,000 to 65,000 papers published in 2003 stemmed from NIH-supported research. The agency funds researchers with about $30 million annually in direct costs for publication expenses, including page fees, color charges and reprints.

The new policy for the service contradicts an earlier proposal that angered some publishers. In a draft policy for public comment issued in September, NIH officials said all researchers who receive public funding must make their results free to the public via PubMed Central no later than six months after the work is published in a professional journal. The agency received more than 6,000 public comments about the proposal.

Members of the Association of American Publishers (AAP) argued that "there was not justification in trying to establish this one-size-fits-all window for publication when all operate under a different business model," Allen Adler, AAP's vice president for legal and governmental affairs, said last month. Publishers might have had little opportunity to recoup costs because many people would forgo subscriptions, knowing the information would be available online within six months, critics of the September proposal said.

The policy announced this week shows that NIH Director Elias Zerhouni has addressed some of the publishers' concerns, particularly regarding flexibility in the submission and online posting of articles, Adler said. However, a number of their questions remain unanswered, including the impact on the business models of various for-profit and nonprofit publishers, the effect on the peer review process and the possibility of authors paying for their own articles, he said.

Free-access advocates also scorned the new policy for putting the researchers in an awkward position. Scholarly Publishing and Academic Resources Coalition Director Rick Johnson doubts the policy will lead to greater NIH grantee participation.

"We think NIH certainly grasped the importance and rationale for taxpayer access to NIH-funded research, [but] their execution on that left something to be desired," he said. "We're ready to stand back and look at the coming results, to see if 1. Are most scientists depositing research? and 2. Is most research available soon after publication? The unfortunate thing is that scientists have been put in the middle here. We believe that NIH had the responsibility and the authority to call upon all NIH grantees to deposit their research articles."

Patrice McDermott, deputy director of the American Library Association's Office of Government Relations, expressed similar disappointment over easy access to NIH-funded research. "We think it's a move back from what NIH initially agreed to...it will delay no-fee access to this taxpayer-funded research and it will only be voluntary," she said. "They're going to ask the researcher to do it, not even say the researcher should or shall."

NIH officials say the new public access policy will not harm scientific publishing. Only a portion of scientific journal articles are NIH-funded, so scientists and libraries will still need journal subscriptions to read many articles. Publishers said they would discuss the issue further, during a review process. NIH will establish a public access advisory working group, representing patient advocacy, scientific, library and publishing communities.

"In developing this policy, we made a concerted effort to balance the importance of this archive to NIH's public health mission with the need to provide flexibility for authors, their institutions and publishers in those cases where immediate release is not possible," Zerhouni said in a press release. "NIH recognizes the importance of preserving quality peer review and the viability of a diversity of publishing models. Nevertheless, we expect that only in limited cases will authors deem it necessary to select the longest delay period."