NIH Public-Access Policy

Frequently Asked Questions

I stopped updating this page in April 2006. It does not describe the newer, mandatory version of the policy approved by Congress in December 2007, released by the NIH in January 2008, and implemented in April 2008.  —Peter Suber.

What is this policy?

The general policy is to provide free online access to full-text, peer-reviewed journal articles arising from taxpayer-funded research. The National Institutes of Health (NIH) asks every scientist who receives an NIH research grant, and who publishes the results in a peer-reviewed journal, to deposit a digital copy of the article in PubMed Central (PMC), the online digital library maintained by the NIH. PMC will then provide free online access to its copy some time after the article is published in a journal, the length of the delay to be determined by the author.

The NIH is the largest funder of medical research in the world, and the largest funder of non-classified research in the U.S. federal government. Its budget for fiscal year 2005 is $28 billion.

The NIH maintains its own page on the policy, including its own FAQ.

The policy took effect on May 2, 2005.

Some procedural history

On July 14, 2004, the U.S. House Appropriations Committee adopted a set of recommendations for the 2005 federal budget. One key recommendation instructed the National Institutes of Health (NIH) to develop a policy requiring free online access to articles based on NIH-funded research no later than six months after their publication in peer-reviewed journals.

On September 3, 2004, the NIH released its draft policy, Enhanced Public Access to NIH Research Information, for a 60 day period of public comment ending on November 2. In this draft, the NIH reduced the House's suggested "requirement" to a "request" but retained the six-month embargo period.

On September 8, 2004, Rep. Ernest Istook (R-OK) and Rep. Ralph Regula (R-OH) engaged in a "colloquy" on the House floor (Congressional Record, September 8, p. H6833), clarifying some of the legislative intent for the record.

On September 9, 2004, the House of Representatives adopted the NIH recommendation in the House appropriations report by an overwhelming bipartisan vote of 388-13. The Senate Appropriations Committee decided to remain silent on the House recommendation in its own appropriations bill.

On September 17, 2004, the NIH published its draft policy same text in the Federal Register, for another 60 day period of public comment ending on November 16. This was the same text released on September 3. On September 24, the NIH merged the two comment periods so that both would end on November 16, 2004.

On November 20, 2004, the House-Senate conference committee reconciling the two appropriations bills reaffirmed a version of the House recommendation, and the resulting appropriations bill was approved by both houses of Congress. President Bush signed it on December 8. Here is the language on the policy from the conference commimttee (pp. 104-105):

The conferees are aware of the draft NIH policy on increasing public access to NIH-funded research. Under this policy, NIH would request investigators to voluntarily submit electronically the final, peer reviewed author's copy of their scientific manuscripts; six months after the publisher's date of publication, NIH would make this copy publicly available through PubMed Central. The policy is intended to help ensure the permanent preservation of NIH-funded research and make it more readily accessible to scientists, physicians, and the public. The conferees note the comment period for the draft policy ended November 16th; NIH is directed to give full and fair consideration to all comments before publishing its final policy. The conferees request NIH to provide the estimated costs of implementing this policy each year in its annual Justification of Estimates to the House and Senate Appropriations Committees. In addition, the conferees direct NIH to continue to work with the publishers of scientific journals to maintain the integrity of the peer review system.

On February 3, 2005, the NIH released the final version of the policy, to take effect May 2, 2005. Its official title is Policy on Enhancing Public Access to Archived Publications Resulting from NIH-Funded Research. This version modifies the permissible delay from six months after publication to "as soon as possible (and within twelve months)" after publication, at the author's discretion. The document containing the policy also includes a lengthy section carefully answering many of the objections raised during the public comment period.

On April 29, 2005, the NIH released some instructions on implementing the policy and launched a web site on the manuscript submission process.

On May 2, 2005, the policy went into effect and the NIH began accepting submissions.

On November 15, 2005, the Public Access Working Group, appointed by the agency to advise it on implementing and improving the policy, recommended (1) that the request for public access be strengthened to a requirement and (2) that the permissible delay be shortened from 12 months to 6 months. The working group was responding to NIH data showing that fewer than 5% of NIH grantees were complying with the request for public access.

In early February 2006, the NIH sent a progress report to Congress (dated January 2006). Among other things it reported that the rate of compliance with its request for public-access was below 4%; that handling existing submissions under the policy cost the agency $1 million/year; and that handling submissions under a 100% compliance rate would cost the agency $3.5 million/year.

On February 8, 2006, the NLM Board of Regents endorsed the November 2005 recommendations of the Public Access Working Group (see above) in a letter to NIH Director Elias Zerhouni.

Are grantees required to participate?

No. The policy requests participation and does not require it. The NIH emphasizes that compliance is voluntary and that there are no penalties for non-compliance.

The July 14 House language instructed the NIH to craft a policy requiring participation. But the NIH's September 3 draft published for public comment reduced the requirement to a request.

Who is being asked to deposit articles in PubMed Central?

The request is directed to the grantee, not the grantee's publisher. However, there's nothing in the language to prevent services from springing up to help grantees in this regard. Grantees could be helped by their publishers, their employers, PMC itself, or third-party entrepreneurs. Grantees will not have to tag or mark-up their texts, or prepare them in any special way, before deposit.

What version of an article should grantees deposit?

The version accepted by a journal's peer-review process. Preprints will not suffice, but there are at least two kinds of postprint. If the peer-review process introduced one layer revisions and copy editing introduced another, then grantees are only asked to deposit the peer-reviewed version. However, publishers may replace that version with the final, copy-edited, published version if they wish.

When should deposit take place?

Upon "acceptance" of the article by a peer-reviewed journal. Hence the policy does not apply to grantees who do not submit their work to a journal (for example, because it contains patentable discoveries). The same goes for authors who write books instead of journal articles, or whose work is rejected by journals. Moreover, even when the policy does apply to a grantee, it only requests deposit in PMC and does not require it.

When will PMC release its free online edition of an article?

The timing of public release is decided by the grantee when the article is deposited. The NIH "strongly encourages" grantees to authorize public release "as soon as possible" after publication, and at least within 12 months of publication. Publishers may authorize earlier public release than the grantee, if they wish.

The July 14 House language would have required immediate online access if the NIH paid any part of the costs of the article's publication. But the September 3 NIH draft dropped that provision. Both the July 14 House language and the September 3 NIH draft provided that public release for depositied articles would take place within six months after publication.

Why PubMed Central?

PMC is maintained by the NIH; it already houses a very large body of medical literature; it has benefited from years of infrastructure refinements; it is committed to free online access, long-term preservation, and interoperability. In addition, it will help the NIH manage its portfolio of funded research projects and insert links from research articles to various NIH-supported databases, such as GenBank and PubChem.

Nothing in the policy prevents grantees from depositing the same articles in other repositories as well as PMC.

Why is it called "public access" rather than "open access"?

Open access literature is free of charge and also free of most copyright and licensing restrictions. (This is true even of the PMC's own definition of "open access".) But while the articles on deposit in PMC will be free of charge, they will usually be under copyright and nothing in the policy requires copyright holders to permit more than what U.S. copyright law considers fair use.

Is this a step toward "government-controlled science"?

No. PMC will not perform peer review. It will only host or disseminate articles that have been peer-reviewed by independent (non-government) journals. Nothing in the policy tells scientists what to study, what to conclude, or what to write. Nothing in it tells journals what to publish or how to conduct peer review.

Nor will PMC be the only source for any of these articles. The policy only applies to articles that have been published elsewhere, so fears of "government censorship" through control of PMC are entirely misplaced.

Will this policy cause journals to lose subscribers?

The policy makes several large concessions to publishers designed to preserve subscriber incentives and publisher revenue. Beyond these concessions there are several background reasons to think that subscriptions will not suffer.
  1. Deposit in PMC is voluntary. Many authors will not participate.
  2. PMC will not release the free online edition of an article until some time after publication. --up to 12 months and perhaps even beyond. Only the publisher can authorize release earlier than the author did. This gives libraries a reason to maintain their subscriptions. Here's how the Association of College and Research Libraries put it in its comment on the NIH policy: "We wish to emphasize, above all, that academic libraries will not cancel journal subscriptions as a result of this plan....Even if libraries wished to consider the availability of NIH-funded articles when making journal cancellation decisions, they would have no reasonable way of determining what articles in specific journals would become openly accessible after the embargo period."
  3. PMC will distribute the version approved by a journal's peer-review process but not necessarily the version polished and revised by the journal's copy editors. This makes the journal the exclusive distributor of the final, official edition --forever, if that is its choice. PMC will only distribute the copy-edited, published version with the publisher's consent.
  4. Most issues of most journals publish more than NIH-funded research. This means that PMC will only provide free online access to a subset of the articles in a given issue of a given journal, again preserving library incentives to maintain their subscriptions. (The NIH estimates that articles based on NIH-funded research comprise only about 10% of the annual volume of biomedical journal articles.)
  5. Most journals publish more than peer-reviewed research articles, including letters, editorials, opinion pieces, review articles, book reviews, news, and conference information. NIH is not asking authors to deposit any of these in PMC.
  6. Some journals report that delayed open access increases subscriptions. For example, the American Society for Cell Biology reports that subscriptions and submissions to Molecular Biology of the Cell both grew steadily after it adopted the policy to provide open access to its articles with just a two-month delay. The increase seems to be a result of the heightened visibility, citation impact, and usage of the journal's articles.
  7. In physics, where nearly 100% of new articles are freely available from birth in an open-access archive (arXiv), subscription-based journals have continued to thrive. Moreover, two leading publishers of physics journals, the American Physical Society and the Institute of Physics, are so comfortable with arXiv that they have launched their own mirrors of it. (See the APS mirror and news of the IOP's intention to launch one of its own.)
  8. A study commissioned by the ALPSP (March 2006) found that high journal prices far surpassed OA archiving as a cause of journal cancellations.
  9. The NIH policy essentially calls for the free online archiving of refereed articles or postprints after a ccertain delay. Elsevier, the world's largest publisher of scholarly journals (more than 1,800+) already permits its authors to deposit their postprints in free online archives, and with no delay at all. Moreover, Elsevier allows free online archiving of the final version of the text, after both peer review and copy editing. Many other journals are adopting similar policies, especially since Elsevier announced its policy in June 2004. For example, SAGE Publications adopted the same policy in October 2004. If free online archiving of postprints harmed a journal's subscriber base, then the number of subscription-based journals permitting it would not be increasing.

Will this policy restrict the freedom of authors to publish in the journals of their choice?

  1. Participation is voluntary. Authors who dislike the policy needn't deposit their articles in PMC.
  2. The policy encourages and facilitates postprint archiving, without mandating it. But all open-access (OA) journals and about 80% of surveyed non-OA journals already permit postprint archiving. So NIH-funded authors will be able to choose among all OA journals and 80% of the rest. Indeed, there are indications that this latter number may grow. The number of journals that permit postprint archiving has been rising steadily since Reed-Elsevier decided to do so in June 2004. For example, SAGE Publications adopted the same policy in October 2004. Moreover, a February 2004 study showed that an overwhelming majority of authors surveyed would "willingly" abide by a free online archiving requirement from their funding agency.
  3. Finally, not a single journal has announced that it will refuse to accept work based on NIH-funded research. If this pattern holds, then 100% of journals will be available to NIH-funded authors.

Will the cost of enlarging PMC to handle the new submissions divert funds from the NIH research mission?

The NIH estimates that cost of this program will be $2 to $4 million/year, a tiny fraction (about 0.01%) of the agency's $28 billion annual budget.

The amount is even a small fraction of the $30 million/year the agency spends on page charges and other subsidies to subscription-based journals.

Some critics of the policy argue more generally that any NIH funds spent on public access could be better spent finding cures for diseases. This is a false dilemma that overlooks or denies the significant role of public access in accelerating research for cures. This role is made clear, for example, in the endorsement of the NIH policy from the Genetic Alliance, a coalition of 600+ advocacy, research and health care organizations whose overriding concern is to accelerate research on cures. Also see the many disease advocacy organizations that have endorsed the policy and joined the Alliance for Taxpayer Access. The disease groups that oppose the policy tend to be those that publish subscription-based journals.

Will this policy force subscription-based journals to become open-access journals?

No. The policy requests deposit in a free online archive, not publication in a free online journal. Free online access through PMC will have consequences for journals, but it will not require journals to convert to open-access business models. (For background, read more about the distinction between open-access archives and open-access journals.)

The policy does not regulate publishing or tell publishers what they must do. It regulates grant recipients, not publishers. It adds one more condition to the grant contract between a funding agency and its grantees.

If the worry is that the policy will cause subscription-based journals to lose subscribers, forcing them to convert to open access in order to survive, then see the reasons to think that the policy will not have this effect, above.

Will this policy affect peer review?

No. The policy only applies to articles accepted at peer-reviewed journals, and the peer review will be conducted by those journals as they see fit.

Will this policy affect copyright?

No. NIH-funded research is currently copyrightable and will remain copyrightable under this policy. Initially the copyright belongs to the author. If the author submits an article based on NIH-funded research to a journal that requests transfer of copyright, then the author is free to transfer copyright and the journal is free to acquire it.

Why not just "let the market work"?

There are three problems with this argument.

  1. Scientific research and its dissemination are permeated by government spending and government policies. Hence, this is not a market in any ordinary or classical sense at all. For example, in the U.S. most funded scientific research is funded by the taxpayers, most scientists work at public institutions and are paid by taxpayers, and most subscriptions to subscription-based journals are purchased by public institutions and paid for by taxpayers. If journal publishers really meant "keep government money and policymaking out of this sector", most would go bankrupt immediately.
  2. Almost every neutral observer believes that the journal publishing system is dysfunctional and unsustainable. These include major research universities, like Harvard, Stanford, Duke, and Cornell, that can no longer afford to subscribe to the range of journals they need, financial analysts like PNB Paribas and Credit Suisse First Boston that have analyzed the industry for investors, and government panels like the UK House of Commons Science and Technology Committee that have investigated the industry for universities and taxpayers. Insofar as it's a market, it has failed.
  3. If the objection is that government should take no actions that reduce profits for private-sector businesses, then it is easily answered. The purpose of this policy is to accelerate research, the pace of discovery, and the translation of bench science into concrete benefits for the public. It is to share important research about disease and health with researchers who can build on it, physicians who can apply it, and taxpayers who have paid for it. It is to make NIH-funded research as useful as it can possibly be and to take full advantage of the rapid growth in NIH research spending. It is to take advantage of the internet to increase the return on the taxpayer investment in research. It's about advancing the public interest, not a private interest.

Is this policy radical?


  1. The policy is a compromise between the private interests of publishers and the public interest in full, immediate open access to the final form of taxpayer-funded research. There are four large concessions to publishers in this policy, each one a step away from the public interest: (1) requesting rather than requiring grantees to provide free online access to their work, (1) delaying free online access for up to 12 months (or even more) after publication, (2) releasing the edition approved by peer review but not necessarily the edition refined by copy editing, and (4) retreating from full open access to mere no-fee access that limits users to "fair use" under copyright law.
  2. The NIH has been providing free online access to full-text articles through PMC since February 2000. The National Library of Medicine (NLM), the division of NIH containing PMC, has provided long-term access and preservation for medical literature since 1876, under the authority of Congress (42 U.S.C. 286.b.1). The new policy takes the justified, incremental step of adding the articles that result from the NIH's own funded research projects to the NIH's online library of medical research.
  3. The United States has already committed itself to the principle of open access by signing the World Summit on the Information Society (WSIS) Declaration of Principles and Plan of Action (December 12, 2003) and the Economic Co-operation and Development (OECD) Declaration on Access to Research Data From Public Funding (January 30 2004).

How does this policy differ from the Sabo bill?

Rep. Martin Sabo (D-MN) introduced the Public Access to Science Act to Congress in June 2003. The NIH public-access policy differs from the Sabo bill in three critical ways.

  1. The Sabo bill would have revised U.S. copyright law to deprive federally-funded research of copyright. The new policy does not affect copyright law. NIH-funded research was copyrightable before this policy was adopted and it remains copyrightable now.
  2. The Sabo bill would have put a certain body of literature into the public domain without creating free online access to it. The new policy provides free online access itself (without putting any literature into the public domain).
  3. The Sabo bill would have applied to classified military research, revenue-producing books, and patentable discoveries --not by intent but because the language was not crafted to exclude them. The new policy does not apply to classified research (because the NIH doesn't fund any) or to books (because the language is limited to journal articles). It might apply to research that results in patentable discoveries, but only when the grantees themselves choose to publish their research in journals and choose to participate in the NIH public-access program.

Now that the policy has been adopted, what should open-access proponents do next?

  1. Let's strengthen the policy. First, let's return to the intent of the House of Representatives and make the request a requirement. Taxpayer access to publicly-funded medical research should be guaranteed, not discretionary. Second, reduce the permissible delay after publicationn to six months at the most. Third, rely on the NIH's government-purpose license (45 CFR 74.36) to disseminate NIH-funded research rather than copyright-holder consent; the latter simply invites publishers to block implementation of the policy.
  2. Ask Congress to monitor compliance. What percentage of NIH grantees deposit their work in PMC within 12 months? Of those that do, what is the average authorized delay after publication?
  3. Persuade other funding agencies --inside and outside the U.S.-- to adopt the best parts of the NIH policy but not all its concessions to publishers. Do provide free online access to publicly-funded research, but don't make it discretionary and don't delay the public release more than six months.
  4. Encourage NIH-funded researchers to deposit all their eligible publications in PMC and to authorize public release immediately upon publication. Self-archiving in repositories other than PMC can serve most of the same purposes. SPARC has put together a good page on the benefits for researchers in complying with the NIH request.
  5. Encourage biomedical journals not to oppose the NIH request to authors, and not to force authors to choose between conflicting requests from their funder and their publisher. If you belong to a scientific society that publishes journals in biomedicine, or if you edit or referee for journals in biomedicine, then let them know that this is how you feel.

Major endorsements (chronological order)

Also see the NIH's own FAQ on the policy, Prue Adler's FAQ on the policy, and the links and materials collected by the Alliance for Taxpayer Access and SPARC.

Last revised April 13, 2006.

For more background, see my overview, newsletter, and blog on open access.

Peter Suber
Open Access Project Director, Public Knowledge
Research Professor of Philosophy, Earlham College
Senior Researcher, SPARC

Creative Commons License
This work is licensed under a Creative Commons Attribution 2.5 License.