UK Research and Innovation (better known as UKRI) directs the UK government’s research and innovation funding across all disciplines, to the tune of £8b annually, making it a force to be reckoned with in the UK and beyond. UKRI’s open access (OA) policy aligns with cOAlition S, in order to improve consistency and reduce the burden of complying with the somewhat staggering 1,155 current OA policies and mandates worldwide (according to ROARMAP). The current policy applies to UKRI-funded peer-reviewed research articles (including reviews and conference papers), as well as monographs, book chapters, and edited collections. There are two routes to compliant open access: through immediate OA of the version of record (journals – within 12 months for books); or, if publishing in a subscription journal, by immediately depositing the Author’s Accepted Manuscript (or Version of Record, where the publisher permits) in an institutional or subject repository on publication.

Many of the technical requirements in the policy can only be achieved or evidenced through metadata such as licensing information, persistent identifiers for items and contributors, links to funding grants, and relationships between versions of articles. However, for a variety of reasons, the current metadata landscape does not support consistent implementation of the technical requirements set out in UKRI’s OA policy, making it difficult for individuals or organizations to fully comply (and show that they have done so). This is especially the case for organizations with fewer resources, such as smaller publishers and less research-intensive institutions.

In order to help establish where the pain points are and what can be done to alleviate them, UKRI asked MoreBrains (of which I am a co-founder) to analyze the policy text (especially the technical requirements), identify the metadata required by the two OA routes, establish the common metadata standards and frameworks (or ‘schema’) used in the publishing community (for route 1) and the repository community (for route 2), and establish which of the metadata required by the policy were present or missing from each. Where metadata was available, we assessed the availability of information and gauged the scale and extent of gaps in metadata coming from publishers and repositories, by analyzing how often the relevant metadata was present for UK research in the Crossref Registry (for journal publications) and the CORE database (for repositories).

Of the 16 UKRI requirements that involve metadata – ranging from data access statements (DAS) and licensing information to persistent identifiers (PIDs) for various entities – we found only two for which metadata fields are available and well-described with adequate adoption. In the vast majority of cases, either metadata fields are available but are poorly used or misused, or there are no metadata fields and no mechanism to monitor levels of compliance.

Something needed to be done – and that something was a series of community consultations. These took place over several months in late 2023 and involved 145 individuals representing more than 40 research organizations from around the world – publishers, repositories, infrastructure and service providers, and more. Through a mix of round table discussions, focus groups, and workshops, we examined the challenges, identified potential existing or future solutions, and developed a set of practical steps to enable UKRI-funded authors to meet the metadata and technical requirements of UKRI’s OA policy – and, importantly, to make it easy for them to show that they have done so.

Specifically, we recommended that:

• UKRI should lead by example and start registering Grant IDs for its own research grants
• Relevant stakeholders should collaborate to conduct targeted outreach and communication programs
• Existing technical resources should be used by publishers, publishing technology vendors, and repository communities
• More research should be conducted into the barriers to compliance faced by the publishing and repository communities
• Specific plugins for repositories should be developed, for example, to help them with PID integration

Most of these actions will involve contributions from a range of stakeholder groups, with UKRI playing its part, for example by making direct interventions, convening and facilitating discussions, conducting additional investigations, and making recommendations for updates to global standards and services. These recommendations will take some time to implement, and most will require a lot of community input, engagement, and effort, but we did identify some quick wins. For example, this PID matrix (developed as part of the work we undertook for the Canadian PID Advisory Committee) provides a set of criteria for evaluating persistent identifiers that could easily be adapted for use by other countries. Likewise, a short project to identify the characteristics of a suitable repository would be a quick and effective step forward, which could be used by a wide range of stakeholders.

At a September webinar to discuss the report we developed, speakers Catriona McCallum (Wiley and OASPA), Melissa Harrison (European Bioinformatics Institute), and Agustina Martinez-Garcia (UK Council of Open Research and Repositories aka UKCORE) shared their perspectives on these recommendations.

Catriona started by reiterating the critical importance of community action and the fact that the UKRI policy relies on every participant in the research communication process creating and sharing consistent, reliable, standardized metadata. She noted that it’s not just publishers and repositories that need to play their part; funders also have an important part to play by collecting metadata right at the start of the research life cycle. And she applauded the report’s focus on how all stakeholders can collaborate to address the challenges around metadata. At the same time, Catriona was very cognizant of how much harder it is for smaller organizations and those in less well resourced communities to implement the changes that will be needed. Lastly, she expressed a hope that UKRI will take the lead on the additional research that is required, as well as taking up the two recommended follow-on projects about PID selection and characteristics of a suitable repository.

Melissa focused on a couple of specific aspects of the report. First, she noted the importance of referring to PIDs in general throughout, rather than Digital Object Identifiers (DOIs), since most databases in her space use accession numbers rather than DOIs. For example, the European Nuclear Tide Archive has over 3.8 billion accession numbers compared with Crossref’s (at the time) 164.24 million registered DOIs. So, while DOIs are the best solution for many publishers, as with all things, there is no one size fits all solution. Melissa next raised the issue of using a CC BY license. While recognizing that there is a lot of concern around reuse of research, she stressed that open infrastructure and the reuse of content has clear benefits for society. Finally, Melissa talked about the importance of standards in terms of ensuring a consistent approach globally. She focused specifically on the Journal Article Tag Suite (JATS); its working groups, which focus on elements such as Grant DOIs and open access metrics, help publishers to semantically tag their outputs as closely as possible to each other, meaning that a larger corpus of content can be reused.

Agustina was speaking from a UKCoRR perspective, and she notes that they might be able to take the lead on some of the activities specifically involving repository management. However, like Catriona, she also flagged that many of their institutions are facing strained financial circumstances, making additional work challenging. Agustina flagged the importance of outreach and education on PIDs and the need for appropriate guidance on how they should be implemented, as well as the proposed institutional engagement at senior level around the importance of investing in PIDs. She suggested that it would be helpful to extend this to include the importance of investing in repository infrastructure more generally. Agustina noted that the heterogeneity of approaches to metadata standards and metadata creation in repositories in general is challenging. She felt that UKRI could help by recommending and endorsing adoption of a specific set of standards and by supporting institutional repositories with the implementation of these standards.

All the speakers – as well as UKRI itself – are clear that the value of this work extends beyond UKRI and the publishers and repositories that are publishing its funded research. A better research information landscape will also benefit the research endeavor more widely, for example, by improving transparency, increasing operational efficiency in publishing companies, and reducing bureaucratic burden in universities. Importantly, many of the learnings from this deep dive into the challenges and opportunities for improvement around the technical implementation of UKRI’s open access policy are equally applicable to other OA policies around the world; they will help deliver these benefits to researchers and research organizations everywhere.