The US government should make all research it funds open access

Many federal policy changes are well known before they are announced. Hints in speeches, leaks and early access to reporters at major outlets set the stage for eventual confirmation. But on Thursday, the White House Office of Science and Technology Policy (OSTP) released a big one that seemed to catch everyone off guard. Starting in 2026, any scientific publication that receives federal funding must be made openly available on the day it is published.

The move could further shake up the scholarly publishing industry, which has already adopted preprint archives, similar mandates from other funding agencies, and greatly expanded access to publications during the pandemic.

There was a change Alondra Nelson announced, acting head of OSTP (a permanent director is pending Senate confirmation). The official policy is set out in the accompanying memorandum.

Embarrassing story

The US government is likely to be the world’s largest funder of scientific research. For medical research, the US National Institutes of Health spends more than the rest of the top 20 organizations combined. However, for decades the scientific publishing system was set up so that the government (much less the people it represents) did not necessarily have access to the research it funded. Instead, access was based on a paid subscription to the journals in which the work was published.

Two trends have undermined this limitation. One is the rise of open access journals, which charge researchers upfront fees and then provide anyone with Internet access until the research is finally published. The second is the “preprint” trend. Preprint servers created by researchers in physics and astronomy provide access to manuscripts submitted by publishers for review. Their use in the life sciences has greatly expanded during the COVID-19 pandemic.

Some people in scholarly publishing worried that these trends would undermine the finances of the entire publishing industry, while others hoped to encourage them to open up all scholarly publishing. That tension has played out in the halls of Congress, where the legislation is competing mandate or block open access to federal research. A truce of sorts was reached during the Obama administration. For federally funded research, publishers had two options: either make the publication open access from the start or have subscription-only access for a year before opening things up. Government-funded repositories were opened to host copies of documents that were not publicly available on the publisher’s website.

During this time there was a large increase in the number of open access journals, and many subscription journals allowed authors to pay a fee for immediate access to published articles. Most subscription journals have also offered COVID-related papers in open access at no extra charge. OSTP appears to have decided that these adjustments have prepared the industry to survive even higher levels of access.

Wide open

OSTP also claims that the benefits to society will be great. “When research is widely available to other researchers and the public, it can save lives, give policymakers the tools to make important decisions, and achieve fairer outcomes in all sectors of society,” Nelson said in a statement. The policy paper elaborates on this, arguing that the COVID-19 pandemic has shown that the benefits of immediate public access (as opposed to just an annual subscription) are even greater.

Under the new policy, any agency with more than $100 million in grant funding has 180 days to submit a revised policy to OSTP; smaller agencies have a year to do so. These plans should provide for the placement of any publications resulting from this funding in public repositories on the day they appear in the scientific journal. They can still appear in subscription-only magazines, but the copy must be publicly available. Separately, any data used in the publication must also be placed in a public repository.

By the end of 2024, agencies must have plans to ensure that information about everyone involved in the publications is also available in the repositories. All data and documentation related to publications must have a digital identifier (such as DOI), related to them. By the end of 2025, all these policies must be implemented.

To some extent, this is consistent with existing NIH policy that allows authors to retain pre-publication versions of their articles; publishers can still add value through formatting, integrated graphics and movies, and convenient cross-references to other studies, thus (potentially) justifying the subscription price. Charging a fee for publishing documents in the open will also remain an option. So it won’t kill the scientific publishing industry right away.

The most striking change is the immediate availability of a parallel copy on the day of publication. But in the long run, having a system to identify and access the underlying data may be more important. Most publishers now require the sharing of relevant data, but there is often no formal way to enforce this requirement and it is often ignored. Imposing this funding condition can obviously change that.

As now, compliance with the new requirements will be key to the impact of this program. By the time an article is peer-reviewed and published, many of the contributors have moved on to other projects. There are no obvious rewards for taking the time to commit their earlier work to the repository, so there must be some sort of enforcement mechanism to encourage adoption.