Recommendations about publishing and publication procurement practices across the health and social care system

Abstract

<b>Recommendations about publishing and publication procurement practices across the health and social care system</b><br /> <br /> November 2020<br /> <br /> An independent report by Information Power<br /> <br /> <b>Introduction</b><br /> This study was commissioned by the Department of Health and Social Care and Health Education England to inform development of Open Access (OA) strategy and policy in the UK health and social care system. OA ensures that research publications, such as journal articles or books, are freely available online to everyone for access to read and re-use. There are new opportunities to embrace OA approaches that have already been developed, tested, and proven effective for medical research funders, by Plan S [link: <a href="https://www.coalition-s.org/">https://www.coalition-s.org/</a>] for example, and for employers, by members of Universities UK [link: <a href="https://www.universitiesuk.ac.uk/policy-and-analysis/research-policy/open-science/Pages/uuk-open-access-coordination-group.aspx">https://www.universitiesuk.ac.uk/policy-and-analysis/research-policy/open-science/Pages/uuk-open-access-coordination-group.aspx</a>] for example.<br /> <br /> The NHS in England is undertaking more research than ever before. However, investment on access to research outputs, through library subscriptions, is relatively low. The available data suggests that Arm’s Length Bodies collectively spend under &pound;1m per annum on primary journals and a similar amount per annum on journals in aggregated databases. NHS Trusts collectively spend an estimated &pound;4m per annum on journal subscriptions. Taken together, the combined NHS expenditure is less than a single medium-sized UK university which seems starkly out of alignment with the importance of evidence-based care decisions.<br /> <br /> In addition, NIHR spends a significant amount to publish articles that report on its funded research. We estimate that this figure is in excess of &pound;10m per annum, however the precise figure is unknown because costs are included in grants awarded through a range of programmes, and open access expenditure is not explicitly tracked at present. We analysed 121,915 papers published between 2015 -2019 by authors employed in, or funded through, the UK health and social care system. 85,542 or 70% of the papers were published Open Access (OA) which means they are freely available for anyone in the world to read. 34% of the total articles are immediately available (19% of all papers were published as fully Gold OA, and 15% as hybrid Gold OA) and 36% of the total articles are available after a delay period of six months (19% were published and made available as Green OA, and 17% as Bronze OA).<br /> <br /> Publishing costs have been paid upfront&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <ul> <li>Gold – costs typically paid by author from research grants</li> <li>Platinum – costs typically paid by employer, funder, or another sponsor often with a per-article charge</li> <li>Diamond – costs typically paid by employer, funder, or another sponsor without any per-article charge</li> </ul> <br /> Near-final draft has been shared publicly, after a delay of six months: <ul> <li>Green – a version of the article (either the manuscript as originally submitted or as accepted, or the final published version) is shared via an institutional repository or subject repository</li> <li>Bronze – a version of the article is shared via the publisher’s site</li> </ul> This represents good compliance with the NIHR OA mandate, and there is scope for further improvement. Papers with a co-author affiliated with a higher education university are more likely to be available OA (80% vs. 65%). There is some degree of overlap, of course, as some papers include authors from both sectors. This suggests that looking to good practice in the UK university sector offers insight about good ways forward.<br /> <br /> Our qualitative research suggests that there is support for the principles of Open Access by researchers, research managers, and library and information professionals in the health and social care system. Funding for APCs is the primary barrier to researchers publishing their articles OA, however researchers affiliated to, or collaborating with someone affiliated to, universities have less difficulty making their research outputs available OA. Central funding, for example block grants, would drive beneficial change as would centralised and strengthened engagement with publishers, and more education and training. Finally, OA is essential but not enough in itself as there is a clear need for short, actionable summaries of research outputs to inform clinical practice.<br /> <br /> The following recommendations focus on opportunities to increase value for money to the taxpayer by eliminating elements of double payment in this system (for subscriptions and OA publication), by raising standards through better access to research information, and by ensuring publicly funded research information is opened for wider economic and social gain.<br /> <br /> <b>Recommendations to project sponsors</b><br /> <br /> (1) Develop a shared OA strategy and common policies and principles<br /> <br /> A shared OA strategy across the health and social care system is needed to advance the quality and speed of research, and to enable equitable access to knowledge. By working together across the DHSC, NIHR, and Arm’s Length Bodies it will be possible to best leverage the sector’s scale, align and change practice, avoid duplication of effort, and obtain best value for money. The NIHR Open Access policy and the Concordat for Maximising Digital Knowledge are good vehicles for this, and the recommendations in this report can form the basis of a shared approach.<br /> <br /> Common policy and high-level principles would enable organisations to focus on collaborative action to drive immediate open access for health and social care system research outputs. NIHR’s direction of travel (i.e., immediate open access with no embargos, no barriers to re-use and dissemination, publications freely discoverable, and reasonable costs covered) resonates with stakeholders with whom we engaged.<br /> <br /> A coordination group across the DHSC, NIHR, and Arm’s Length Bodies is needed to align goals and behaviour, to promote mutual understanding, to cut through complexities and obstacles, and to cultivate buy-in and consensus over time.<br /> <br /> (2) Invest in financial, publication, and compliance tracking<br /> <br /> In order to monitor the impact of your strategy and policies, more attention is needed to build a database that will enable you to understand how information is created, accessed, paid for, and used in the health and social care system. More attention to data gathering can also help you to set policies, negotiate with publishers, and make the case to DHSC for funding.<br /> <br /> We recommend that you establish reliable methods for monitoring research publications going forward: <ul> <li>invest in databases such as Dimensions, Scopus, or Web of Science, and/or by implementing CRIS systems so your researchers or librarians can track research outputs. There are costs in terms of both time and money to each approach.</li> <li>ask Dimensions to set up GRID codes for the remaining organisations, to facilitate future tracking; this can be done relatively simply and at no cost.</li> <li>ensure RORs [link: <a href="https://ror.org/">https://ror.org/</a>] are created for each organisation in the health and social care system and consider working with database providers such as Dimensions, Scopus, and Web of Science (and other similar service providers) to give them the identifiers they would need to improve their matching algorithms.</li> <li>ensure researchers in the health and social care system have ORCID [link: <a href="https://orcid.org/">https://orcid.org/</a>] researcher identification numbers.</li> <li>improve data in ResearchFish by encouraging researchers to supply the DOI of the final published version of articles and to ensure correct employer and funder affiliation data is included in their articles.</li> <li>review what instructions are given to these researchers regarding the correct acknowledgment of employer and funder/s in papers, making sure that the text is specific about how NIHR and the Arm’s Length Bodies should be cited.</li> <li>seek information from other funders for insight into best practice in encouraging grant recipients to include correct employer and funder acknowledgments in papers.</li> <li>track all expenditure made with any publisher, both for subscriptions and for OA. The negotiation of more cost-effective agreements with publishers also requires reliable data about the amounts paid for APCs and of all articles written by affiliated researchers.</li> </ul> <br /> To maximize compliance with your open access policies, incentives and sanctions will be needed. Enabling only papers immediately available open access to be entered for hiring and promotion decisions would be a powerful incentive. Ineligibility for further NIHR funding would be a powerful sanction.<br /> <br /> Ensure that all Arm’s Length Bodies and NHS Trusts are accountable for ensuring the research of their staff members is available open access immediately upon publication. Interviews with researchers suggested that they receive little or no organisational support to make their outputs open access unless they are in the privileged position of having a joint appointment to a higher education institution.<br /> <br /> Rather than have OA funding follow the grant and be administered by individual researchers, we recommend you consider ways of channelling OA funding via these employers, for example by using block grants. While there is a different sort of bureaucracy involved with these, by transmitting money to organisations in this way you would create an environment with increased accountability which will drive up compliance with your policies.<br /> <br /> (3) Invest in improved access to research information<br /> <br /> We recommend you explore ways to increase investment to ensure that researchers and clinicians can both access to the best scientific information from around the world and publish their research articles open access. The public contributor workshops carried out by NIHR as part of their OA policy review identified that even if patients and the public do not access primary research themselves, they expect that clinicians and care practitioners do and are making decisions based on the best scientific information.<br /> <br /> The NHS in England is undertaking more research than ever before. However, investment on access to research outputs, through library subscriptions, is relatively low. The available data suggests that Arm’s Length Bodies collectively spend under &pound;1m per annum on primary journals and a similar amount per annum on journals in aggregated databases. NHS Trusts collectively spend an estimated &pound;4m per annum on journal subscriptions. Taken together, the combined NHS expenditure is less than a single medium-sized UK university which seems starkly out of alignment with the importance of evidence-based care decisions.<br /> <br /> Based on pilots already done [link: <a href="https://community.jisc.ac.uk/system/files/515/NHS%20%28Finch%29%20Pilot%20outcomes%20Nov%202016%20and%20Cochrane%20website%20sharing.pdf">https://community.jisc.ac.uk/system/files/515/NHS%20%28Finch%29%20Pilot%20outcomes%20Nov%202016%20and%20Cochrane%20website%20sharing.pdf</a>], we estimate that providing subscription access to the scope of scientific journals available in UK universities would cost an additional of &pound;1-2m / year. This investment could be targeted in various ways, but one approach to consider is targeting on the journals in which researchers linked to DHSC Arm’s Length Bodies and NHS Trusts publish, but to which there is no access. We also encourage continued migration to central discovery platforms and services.<br /> <br /> NIHR spends a significant but unknown amount per annum on APCs for journal articles. In 2019, 27,416 articles were published by researchers linked to DHSC Arm’s Length Bodies and NHS Trusts. While the good news is that 70% are OA, 30% remain behind publishers’ paywalls. Providing financial support to ensure all research outputs are published gold OA could therefore cost an additional of &pound;17.7m [unfunded articles 8256 x average APC &pound;2147 = &pound;17,725,632].&nbsp; This is clearly a significant additional expenditure, but by implementing the other recommendations in this paper it is possible to significantly reduce this figure.<br /> <br /> (4) Centralise and strengthen your negotiations with publishers, ideally in partnership with UK universities<br /> <br /> Currently the journal content available via HEE is acquired through public sector procurement processes. Whilst this helps ensure best value in markets where there is a choice of supplier, it not ideal in a market where publisher’s journal content is unique and cannot be obtained via other providers. Neither are public sector procurement process designed for the detailed discussion and negotiations needed to develop innovative OA journal agreements which cover both accesses to journal content, and publication in those same journals.&nbsp; We encourage you to centralise and strengthen your engagement with publishers.<br /> <br /> We recommend that you implement a system-wide policy to avoid any form of Non-Disclosure Agreement with publishers. We were rather concerned that one of the Arm’s Length Bodies had done so and were therefore unable to share their expenditure information with any other Arm’s Length Body.<br /> <br /> We recommend you explore a collaborative partnership with Jisc Collections so that your negotiating strength and power is coupled with those of UK universities. This leverage is your best option to control costs and secure better publishing agreements from a range of relevant publishers but particularly the largest. Between them Elsevier, Springer Nature, and Wiley publish 54% of NIHR-funded papers.<br /> <br /> Jisc is expert in negotiating cost-effective journal agreements, but does not have deep sectoral knowledge. We therefore recommend the creation of a strong steering group chaired by HEE to guide them.&nbsp; A potential investment could be the appointment or secondment of a member of staff at Jisc to strengthen their understanding of the health sector and medical publishers.<br /> <br /> Where there are publishers who are important for the health and social care system, but not the HEI part of it, then consider strengthening your own negotiations by partnering with the Royal Colleges and strengthening the knowledge and skills of the NICE negotiating team.<br /> <br /> (5) Pivot to Read &amp; Publish agreements with publishers<br /> <br /> In order to maximise value for the public purse, we recommend you bring together your subscription spend with publishers with your expenditure for OA publishing and seek agreements with publishers that support both reading and publishing. This means that affiliated authors can publish OA without paying an APC. PHE has already done some experimentation in securing such agreements with publishers.<br /> <br /> We recommend that you seek OA agreements with small and medium sized publishers as well as large publishers. There is quality content produced by Society and other publishers, and it is essential for cost constraint that there is good competition between publishers for authors.<br /> <br /> In advance of any negotiations, we recommend that any publisher with whom you have an agreement is required to complete and return a data collection template.<br /> &nbsp;<br /> In the absence of such agreements, we recommend that you do not provide funding for OA publications in a publisher’s hybrid titles. This is because your existing subscription expenditure will give you full access to the content in these titles, and additional APC payments do not lead to a transition to full OA, merely boost publisher profits and surplus.<br /> <br /> (5) Retain your copyright and publish under open licences<br /> <br /> A very powerful way to increase your negotiation power with publishers is to encourage or require employees in the health and social care system to retain necessary copyrights. This is already a requirement for Crown Servants. We recommend that as a condition of NIHR funding, or employment with an Arm’s Length Body, researchers should be required to retain sufficient intellectual property rights to comply with their funder and employer OA requirements.<br /> <br /> Attention also needs to be paid to education around, and compliance tracking of, the open licences attached to OA articles published by researchers in the health and social care system. We strongly recommend use of Creative Commons (CC-BY) and Open Government (OGL) licences across the health and social care system, especially as rights need to be retained in order to enable immediate green OA (see next recommendation).<br /> <br /> DHSC, its Arm’s Length Bodies, and NHS Trusts all publish reports and other materials on their websites. The copyright status of these publications is often unclear, and we would encourage you to use a CC-BY or OGL license wherever possible on these publications.<br /> <br /> (7) Immediate green OA for articles not published gold OA<br /> <br /> As the proportion of articles published OA in the health and social care system grows, you need to ensure that any articles published under the subscription model also become immediately available. Including green OA options – specifically the immediate self-archiving of peer reviewed accepted manuscripts in one health-oriented repository such as Europe PMC – should be an essential part of your strategy. Including this requirement in your agreement with publishers will enable you to maximise OA outputs while constraining costs.<br /> <br /> Understandably publishers have been unwilling to agree to immediate green OA, and there is too often little incentive for them to do so. The immediate availability of accepted manuscripts under a CC-BY licence is perceived by publishers as likely to undermine the value of their subscription sales in all other parts of the world. Every new article published OA and brought out in front of their paywall not only erodes the value of subscription sales to other parts of the world, but can also fuel rival online services underpinned by so-called black OA (i.e., content piracy) [link: Gold, green, and black open access by Bo‐Christer Bj&ouml;rk <a href="https://onlinelibrary.wiley.com/doi/full/10.1002/leap.1096">https://onlinelibrary.wiley.com/doi/full/10.1002/leap.1096</a>].<br /> <br /> The debate around whether there is risk to publishers of short embargo green OA is as heated as it is long standing. We take a pragmatic view: only the test of time will prove whether or not there is a real risk to subscription revenue from short embargo green OA, however publisher perceptions of risk mean few are willing to incorporate short embargo green OA into their transformative (and other OA) agreements. It is in their interest to control costs and therefore pragmatic for the health and social care system to find a way forward.&nbsp;<br /> &nbsp;<br /> Publishers incur new costs to provide new OA services (e.g. adding author and funder metadata, informing authors about terms of agreements, encouraging authors to share articles providing new reports and metadata, etc). Where publishers expect an uplift in price in exchange for uncapped article numbers in a transformative agreement because the level of publishing output is high, and are willing to provide these new OA services, then we recommend modest financial incentives in otherwise cost-neutral agreements to incentivise the inclusion of immediate green OA. This would help the health and social care system secure more affordable transformative agreements and enable researchers to retain a wide choice in where they publish while delivering both full OA and maximising value for money.<br /> <br /> (8) Open sharing platforms<br /> <br /> Shared infrastructure is required to support the cost-effective dissemination of open outputs by your researchers. A collective approach should be part of your strategy to ensure access to research outputs is not fragmented across different organisations and services, and to avoid duplication of costs.<br /> <br /> In developing the common strategy for this infrastructure is important not to reinvent wheels but to explore the utilisation of existing open platforms.&nbsp; NIHR is one of the funders of Europe PubMed Central which is an open-access repository containing millions of biomedical research papers and has potential to serve as a shared repository service for all your affiliated authors.<br /> <br /> Longer term, you may want to consider not only a repository for access to research, but an open shared platform for publishing NHS health and care research. The AMRC (Association of Medical Research Charities) and Wellcome Trust both maintain open research platforms on which all their funded researchers can publish OA any results they think are worth sharing at an extremely low price. Publication is fast, there is transparent peer review and editorial guidance on making all source data openly available. AMRC Open Research publishes other research outputs, for example posters, slides, and documents, reporting any basic scientific, translational, applied, and clinical research studies: we heard at the round table that these types of research output are as important to the health and care sector as scholarly research. Both the AMRC and Wellcome Trust open platforms use technology provided by F1000 Research Limited.<br /> <br /> A shared open platform such as F1000 research could be helpful in decreasing the costs of OA publishing. The list price of publishing a research article on such a platform is only $1350 per article as opposed to nearly $3000 per article in journals.&nbsp;&nbsp;&nbsp;<br /> <br /> (9) More OA education and training<br /> <br /> Our interviews revealed widespread support for the principles of open access, but fragmented understanding of all the flavours of open access and constraints to supporting its delivery in practice. The specific requirements vary by stakeholder group and include: <ul> <li>Libraries – open access models, how to support researchers to publish open access, how to work with publishers to maximise open access outputs and constrain costs for access to research, supporting researchers to share via Europe PMC</li> <li>Researchers – open access models, how to obtain funding to support open access publishing, how to identify and avoid predatory journals, how to correctly use unique identifiers for funders/grants/employers/co-authors</li> <li>Research managers – why a research culture is important to the health and social care section, how to define the impact of research undertaken, why it is important that research outputs are available to all, what open access is, open access models, how to ensure research is immediately available open access</li> </ul> We recommend you develop education and training materials and courses that can be used across the health and social care system to minimise confusion and to provide consistent information and guidance.<br /> <br /> (10) Require overviews of research in plain English, and provide training and support to enable compliance<br /> <br /> Your shared strategy and common policies and principles must consider ’actionable knowledge', as well as open access. Research findings must inform practitioner decisions and practice. Plain English summaries of research outputs will aid busy practitioners who do not have time to read research articles, as well as patients and members of the public<br /> There are already some examples of good practice in the sector: <ul> <li>the NIHR Centre for Engagement and Dissemination is producing accessible and audience-targeted evidence resources and sharing them across a broad stakeholder community</li> <li>The NIHR provides guidance to researchers on how to write plain English summaries.</li> <li>There are also commercially available services and tools which could assist in this effort.&nbsp; For example, Kudos [link: <a href="https://info.growkudos.com/">https://info.growkudos.com/</a>], which is a free to researchers, allows them to describe their work in plain English, and to enrich their descriptions by adding related resources like blog posts and videos. Scholarcy [link: <a href="https://www.scholarcy.com/">https://www.scholarcy.com/</a>] is a charged-for service which uses Artificial Intelligence to create a summary flashcards of research articles.</li> </ul>