Previous Versions

BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Open access policies of leading medical journals: a cross- sectional study

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2018-028655 review only Article Type: Research

Date Submitted by the 18-Dec-2018 Author:

Complete List of Authors: Ellison, Tim; PharmaGenesis London Koder, Tim; Oxford PharmaGenesis Ltd Schmidt, Laura; Oxford PharmaGenesis Ltd Williams, Amy; PharmaGenesis London Winchester, Chris; Oxford PharmaGenesis Ltd

article processing charges, CC BY, Creative Commons, funding, open Keywords: access, pharmaceutical

http://bmjopen.bmj.com/ on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 29 BMJ Open

1 2 3 Open access policies of leading medical journals: a cross- 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 sectional study 6 7 8 9 Authors: 10 1 11 Tim S Ellison (https://orcid.org/0000-0003-0307-725X), 12 2 13 Tim Koder (https://orcid.org/0000-0001-6152-7365), 14 Laura Schmidt2 (https://orcid.org/0000-0001-6117-781X), 15 16 Amy Williams1 (https://orcid.org/0000-0002-9354-6402), 17 2 18 Christopher C WinchesterFor peer (https://orcid.org/0000-0003-3267-3990) review only 19 20 21 1PharmaGenesis London, London, UK 22 23 2Oxford PharmaGenesis, Oxford, UK 24 25 26 27 28 Correspondence to: 29 30 Dr Tim Ellison 31 Address: PharmaGenesis London, 9 Whitehall, 4th Floor, London, SW1A 2DD 32 33 Email: [email protected] 34 35 Phone: +44 203 675 7065 36 http://bmjopen.bmj.com/ 37 38 39 40 Keywords: article processing charges, CC BY, Creative Commons, funding, open 41 42 access, pharmaceutical, industry, commercial, disclosure 43 44 Word count (excluding abstract, references and figures and tables): 2827 on September 30, 2021 by guest. Protected copyright. 45 Abstract word count: 299 46 47 Figures and tables: 6 (excluding 2 supplemental tables) 48 49 References: 45 50 51 52 53 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 29

1 2 ABSTRACT 3 4 Objectives BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Academic and not-for-profit research funders are increasingly requiring that the 7 research they fund must be published open access, with some insisting on 8 9 publishing with a Creative Commons Attribution (CC BY) licence to allow the 10 11 broadest possible use. We set out to clarify the open access variants provided by 12 13 leading medical journals for research in general and industry-funded research in 14 particular, and record the availability of the CC BY licence for commercially-funded 15 16 research. 17 18 For peer review only 19 20 Methods 21 We identified medical journals with a 2015 impact factor of ≥15.0 on 24 May 2017, 22 23 then excluded from the analysis journals that only publish review articles. Between 24 25 29 June 2017 and 26 July 2017, we collected information about each journal’s open 26 access policies from their websites and/or by email contact. We contacted the 27 28 journals by email again between 6 December 2017 and 2 January 2018 to confirm 29 30 our findings. 31 32 33 Results 34 35 Thirty-five medical journals publishing original research from 13 publishers were 36 http://bmjopen.bmj.com/ 37 included in the analysis. All 35 journals offered some form of open access with 38 39 varying embargo periods of up to 12 months. Of these journals, 21 (60%) provided 40 immediate open access with a CC BY licence under certain circumstances (e.g. to 41 42 specific research funders). Of these 21, 20 only offered a CC BY licence to authors 43 44 funded by non-commercial organizations and one offered this option to funders who on September 30, 2021 by guest. Protected copyright. 45 required it. 46 47 48 49 Conclusions 50 51 Most leading medical journals do not offer to authors reporting commercially funded 52 research an open access licence that allows unrestricted sharing and adaptation of 53 54 the published material. The journals’ policies are therefore not aligned with open 55 56 access declarations and guidelines. Commercial research funders lag behind 57 58 academic funders in the development of mandatory open access policies, and it is 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 29 BMJ Open

1 2 time for them to work with publishers to advance the dissemination of the research 3

4 they fund. BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 29

1 2 3

4 Strengths and limitations of this study BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6  This manuscript includes a systematic analysis of open access policies of 7 8 journals with a high impact factor, including society-owned journals, from 9 10 multiple publishers. 11  The open access policies of all journals analysed were clarified, and 12 13 confirmation of our findings was received by email from 97% of the contacted 14 15 journals. 16 17  Open access policies of the journals and publishers analysed are subject to 18 change, soFor the information peer presented review here may only not be current. 19 20  By selecting journals with a high impact factor, our analysis does not include 21 22 prestigious journals from specialized therapy areas and regional or non- 23 24 English language journals, which may have lower impact factors. 25  Although our study covers only a small number of journals, extending such a 26 27 manual analysis to a greater number of journals without loss of detail and 28 29 verification of all results would be cumbersome and inefficient by relying on 30 31 traditional analysis tools. 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 29 BMJ Open

1 2 3

4 INTRODUCTION BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Hundreds of billions of US dollars are invested in medical research by governments, 7 8 charities and commercial organizations each year, with the aim of extending and 9 improving human lives.1 Publication plays an important role in the dissemination of 10 11 scientific innovation.2 3 However, translation of medical research into clinical practice 12 13 is slow; one study has suggested that it takes an average of 17 years for research 14 15 evidence to reach 50% adoption in clinical practice, with the longest delays occurring 16 after successful publication of clinical trial results.2 3 17 18 For peer review only 19 20 Open access publishing has the potential to improve innovation and speed up its 21 22 adoption. Complete access to research literature encourages viewing of more 23 articles than partial access,4 5 and open access articles appear to be downloaded 24 25 more often and receive more citations than subscription articles, indicating a greater 26 6-9 27 academic impact. There is also evidence suggesting that open access articles 28 have a broader societal impact based on Altmetric data that measure the amount of 29 30 attention publications receive in the news media and social communication 31 32 channels.9-11 Depending on the restrictiveness of its licensing, open access can 33 34 facilitate public and commercial reuse of research results, which is beneficial for 35 collaboration, education and innovation.9 Furthermore, enabling access to the full 36 http://bmjopen.bmj.com/ 37 text of research articles increases transparency, which benefits the public by helping 38 39 both doctors and patients to find complete and current evidence to inform treatment 40 41 decisions, and by preventing potentially harmful decisions being made based on the 42 abstracts of paywalled articles.9 12-14 It is worth noting that the publishing model used 43 44 by a journal (i.e. open access or subscription) has no impact on the quality of articles on September 30, 2021 by guest. Protected copyright. 45 15 16 46 published. 47 48 49 ‘Open access’ is a broad term that encompasses a range of definitions, from ‘free-to- 50 51 read’ (full text available to read on demand, without charge to the reader) to ‘free-to- 52 53 read and reuse’ (with the additional ability to reuse text, tables and figures in different 54 formats). When a journal offers open access, it has wide leeway in the choice of 55 56 policy or policies it will apply, using one of the Creative Commons licences that allow 57 17 58 reuse under specific terms, or offering free-to-read access without a licence. 59 60

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 29

1 2 The Open Access Scholarly Publishers Association and the Budapest Open Access 3

4 Initiative recommend the Creative Commons Attribution (CC BY) licence because it BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 allows sharing and adaptation of published materials for any purposes (both 6 7 commercial and non-commercial), subject only to attribution of the original source.17- 8 9 19 Common alternatives to the CC BY licence include CC BY Non-Commercial 10 11 (CC BY-NC), which restricts commercial reuse; CC BY No Derivatives (CC BY-ND), 12 which restricts adaptation; and CC BY-NC-ND, which restricts both (table S1).17 20 13 14 15 16 Open access with a Creative Commons licence is typically facilitated by article 17 processing charges. Following payment of such a charge by the research author, 18 For peer review only 19 institution or funder, open access articles with a Creative Commons licence are 20 21 usually made available on the journal’s website at the time of publication in the 22 23 publisher’s typeset format (Version of Record). Open access articles that do not 24 include a Creative Commons licence at the time of publication typically involve an 25 26 embargo period before the published articles are freely accessible and may allow 27 28 access only to the accepted manuscript (a version that has not been edited and 29 30 typeset by the journal), which is made available on the author’s institutional website 31 or on PubMed Central/Europe PubMed Central without a requirement for payment. 32 33 34 35 There has been an increasing trend towards open access publishing over the last 36 8 http://bmjopen.bmj.com/ 37 20 years, and almost 50% of articles were published open access in 2015. Many 38 academic and not-for-profit research funders now require the research they fund to 39 40 be published open access.9 21-26 Prominently, the Wellcome Trust and the Bill & 41 42 Melinda Gates Foundation insist on publishing with a CC BY licence to allow the 43 broadest possible use.21 23 Commercial research funders, which fund approximately 44 on September 30, 2021 by guest. Protected copyright. 45 half of all medical research,1 27 28 have been more hesitant to require open access 46 47 publishing but now commonly pay for open access when the option is available.24 In 48 49 January 2018, Shire became the first commercial research funder to require all 50 research manuscripts it funds to be published open access.29 30 At present, no 51 52 commercial funder requires open access publishing with a CC BY licence. 53 54 55 We set out to clarify the open access variants provided by leading medical journals 56 57 for research, in general and industry-funded research in particular, and establish the 58 59 availability of the CC BY licence for industry-funded research. 60

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 29 BMJ Open

1 2 3

4 METHODS BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Using Journal Selector (Sylogent, Newtown, PA, USA), we identified medical 7 8 journals with a 2015 impact factor of at least 15.0 (accurate on 24 May 2017). To 9 focus on journals publishing original medical research, we excluded journals that 10 11 only publish review articles. We collected information on the open access variants 12 13 provided by the included journals from their websites and by email contact when 14 15 information was missing or unclear, making up to three attempts between 16 29 June 2017 and 26 July 2017. 17 18 For peer review only 19 20 For each journal, we recorded the following information: 21 22  for immediate open access, whether a CC BY licence or other Creative 23 Commons licence was provided 24 25  for delayed open access, the length of embargo period for open access 26 27  for both immediate and delayed open access, which version of the article 28 29 would be available (published Version of Record or accepted). 30 31 32 For journals that provided a CC BY licence, we additionally collected information on: 33 34  the requirements for obtaining a CC BY licence (e.g. dependence on funding 35 36 source) http://bmjopen.bmj.com/ 37  article processing charges. 38 39 40 41 Between 6 December 2017 and 2 January 2018, we emailed the journals’ editorial 42 43 offices requesting confirmation of our findings (table S2). Once open access variants 44 were recorded, we categorized the most open variant provided by each included on September 30, 2021 by guest. Protected copyright. 45 46 journal using our own classification, as shown in table 1. 47 48 49 50 51 52 53 54 55 56 57 58 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 29

1 2 3

4 RESULTS BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Included journals 7 8 Fifty-three journals listed in the Journal Selector database had a 2015 impact factor 9 of at least 15.0 (figure 1). After 16 review journals and two non-medical journals were 10 11 excluded, 35 journals from 13 publishers were included in this analysis. Of the 15 12 13 journals that were contacted to clarify information that was missing or unclear, 14 14 15 replied with clarification. Once all information was collected and tabulated, we 16 received confirmation of our findings from 34 (97%) of the 35 journals. 17 18 For peer review only 19 20 Open access variants offered 21 22 Proportions of journals in each category of the most open variant of open access are 23 shown in figure 2A. Immediate open access with a Creative Commons licence was 24 25 provided by 21 (60%) of the 35 journals analysed. The types of Creative Commons 26 27 licence available from these 21 journals under different circumstances were: CC BY 28 from 21 journals (100%); CC BY-NC from 4 journals (19% of all journals offering CC 29 30 BY); and CC BY-NC-ND from 18 journals (86% of all journals offering CC BY). 31 32 33 34 When the 35 analysed journals were categorized by impact factor, immediate open 35 access with a CC BY or other Creative Commons licence was provided by 10 (66%) 36 http://bmjopen.bmj.com/ 37 of the 15 journals with an impact factor between 15.0 and 19.9, and 3 (30%) of the 38 39 10 journals with an impact factor over 30.0 (figure 2B). 40 41 42 All 14 journals, from six publishers, that did not provide open access with a Creative 43 44 Commons licence provided access to different versions of the article either on September 30, 2021 by guest. Protected copyright. 45 46 immediately, after a 6-month embargo period or after a 12-month embargo period 47 under different circumstances (table 2). 48 49 50 51 52 53 54 55 56 57 58 59 60

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 29 BMJ Open

1 2 3

4 The cost of open access with a CC BY licence BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Of the 21 journals that offered a CC BY licence, 19 (90%) disclosed article 7 processing charges on their websites. Across these journals, charges ranged from 8 9 USD $3000 to $5000; the most common article processing charge was $5000 (in 10 11 13 [62%] of journals; figure 3). Details of the fees charged by the remaining 12 two journals (10%) were not available from their websites because the details were 13 14 only provided when the article was accepted. 15 16 17 18 Relationship betweenFor funding peer source review and the availability only of open access 19 20 variants 21 Table 3 shows the open access policies of the journals offering open access with a 22 23 CC BY licence. Of the 21 journals listed, 20 journals allowed open access with a 24 25 CC BY licence for research funded by specific non-commercial organizations, and 26 only The BMJ offered it to organizations who required it, regardless of the nature of 27 28 the funding source. 29 30 31 32 DISCUSSION 33 34 Here, we present a systematic analysis of open access policies of journals with a 35 36 high impact factor, including society-owned journals, from multiple publishers. We http://bmjopen.bmj.com/ 37 met our objective to clarify the open access policies of all journals analysed and 38 39 received confirmation of our findings by email from 97% of the contacted journals. 40 41 We found that all leading medical journals in this study provided some form of open 42 43 access, but there was little consistency across their policies. Over half of the 44 included journals provided a CC BY licence; however, with the exception of one on September 30, 2021 by guest. Protected copyright. 45 46 journal, this option was only available only to authors funded by non-commercial 47 48 organizations. One journal (The BMJ) allowed authors to obtain a CC BY licence 49 when the work was supported by funders who required its use. Therefore, if 50 51 pharmaceutical companies had a policy that required open access with a CC BY 52 53 licence, the The BMJ would be suitable, and other journals might be inclined to 54 55 change their policy. 56 57 58 Limitations of this study are that we investigated journals listed in the Journal 59 60 Selector database with an impact factor of at least 15.0, and that, because impact

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 29

1 2 3 factors and the open access policies of journals and publishers are subject to 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 change, the information may not be current. Furthermore, by selecting journals with 6 7 a high impact factor, our analysis does not include prestigious journals from 8 specialized therapy areas and regional or non-English language journals, which may 9 10 have lower impact factors. Although our study covers only a small number of 11 12 journals, extending such a manual analysis to a greater number of journals without 13 14 loss of detail and verification of all results would be cumbersome and inefficient by 15 relying on traditional analysis tools. If more extensive mining of journal (meta)data 16 17 becomes feasible, however, this study could be repeated for a bigger cohort of 18 For peer review only 19 journals. 20 21 22 To our knowledge, this is the first report showing that the availability of open access 23 24 options depends on the source of funding. Limitations on the availability of the 25 26 CC BY licence depending on the research funder are not in line with statements 27 such as the Budapest Declaration,18 the Berlin Declaration31 and the Bethesda 28 29 Statement,32 which aim to provide end users with immediate access to research 30 31 articles and to give them the opportunity to reuse material without restrictions. 32 33 Furthermore, placing restrictions on access to medical research owing to its source 34 of funding is not in line with the key principles of human research ethics laid out in 35 36 33

the Declaration of Helsinki. http://bmjopen.bmj.com/ 37 38 39 40 Good Publication Practice 3 (GPP3) guidelines state that authors should take 41 responsibility for the way research findings are published.34 In line with these 42 43 recommendations, pharmaceutical companies can and, we believe, should advise 44 on September 30, 2021 by guest. Protected copyright. 45 authors to reach a consensus on which journal to publish with, to avoid predatory 46 journals, and to adhere to sponsor guidelines and regulations. In the authors’ 47 48 experience, some pharmaceutical companies already have internal guidelines 49 50 recommending open access publishing, and one (Shire) now requires it. 51 52 53 Our research shows that one-third of the journals with a high impact factor do not 54 55 offer immediate access to the published version of a manuscript upon publication, 56 57 even though the open access policies of many funders with respect to embargo 58 periods echo the recommendations set out by open access declarations worldwide.18 59 60 21-23 26 31 32 35 Of note, Horizon 2020, which is supported by the European Research

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 29 BMJ Open

1 2 3 Council, requires its beneficiaries to make publications open access no later than 6 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 months after the official publication date and to make every effort to allow for 6 7 maximum reuse of the materials, whether that be copying, distributing, searching, 8 linking, crawling, mining or some other use.36 37 Furthermore, cOAlition S, a group of 9 10 national research funders with the support of the European Commission and the 11 12 European Council, has committed to Plan S, the key principle of which is that 13 14 scientific publications on research funded by participating national and European 15 funders must be published open access by 2020.35 Under the terms of Plan S, 16 17 authors must retain copyright of their publication with no restrictions, and all 18 For peer review only 19 publications must be published under an immediate open licence (preferably CC BY) 20 35 36 21 that fulfils the requirements defined by the Berlin Declaration. 22 23 24 Policies vary between publishers but also across journals at the same publisher, and 25 26 this is also the case for journals not included in this analysis, as shown, for example, 27 by Taylor & Francis in their table of the policies of all their journals.38 Differences in 28 29 policy have many underlying factors, including the choices of the journals’ academic 30 31 editorial boards and societies. A potential disincentive to publishers offering CC BY 32 33 licences to the pharmaceutical industry is the revenue generated from copyright fees 34 and reprints. Permission to reproduce copyrighted materials can cost hundreds or 35 36

even thousands of dollars; for example, the permission fee requested for reuse of a http://bmjopen.bmj.com/ 37 38 single table containing 40 words in the journal American Family Physician was 39 39 40 $4400. Reprints can cost significantly more than permissions charges; for example, 41 reprint sales from a single clinical trial can total $1 million or more, with a large profit 42 43 margin.40 44 on September 30, 2021 by guest. Protected copyright. 45 46 Research by Lundh et al.41 aimed to quantify reprint revenues as a proportion of 47 48 journal income. Of the six journals investigated, the two European journals, The BMJ 49 50 and The Lancet, disclosed the information requested. The editors of the US journals 51 52 Archives of Internal Medicine, Annals of Internal Medicine, JAMA and the New 53 England Journal of Medicine did not provide the data. For The BMJ, reprint revenues 54 55 constituted 3% of its overall income; The Lancet obtained 41% of its revenue from 56 41 57 reprints. In The Lancet, industry-funded publications constituted a large proportion 58 of highly reprinted articles (63/88) compared with a sample of control articles from 59 60 the same journal (23/88).42 The generation of revenue for publishers from the selling

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 29

1 2 3 of reprints leaves publishers open to the criticism that bias can be introduced into 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 editorial decisions.41 This concern could be addressed by a transition to open access 6 7 publishing exclusively with a CC BY licence. However, such a transition may need to 8 be managed. 9 10 11 12 Two of the journals included in our analysis, Science and Science Translational 13 14 Medicine, both published by the American Association for the Advancement of 15 Science, do not disclose article processing charges on their websites;43 instead, they 16 17 provide this information upon their acceptance of an article. This practice does not 18 For peer review only 19 comply with the Directory of Open Access Journals guidelines,44 which state that 20 21 processing fees must be stated clearly on journal websites in a place that is easy to 22 find for potential authors prior to submitting their manuscript. The practice is also 23 24 common among predatory journals, potentially reinforcing perceptions held by some 25 26 academics of the association between open access and predatory publishing. 27 28 29 We found that the open access policies of some journals precluded commercially 30 31 funded research from being published open access, even after an embargo period 32 33 and without a Creative Commons licence. Further analyses could therefore be 34 undertaken to clarify the proportion of journals with this policy and the rationale 35 36

behind this position. Future research could also focus on a larger cohort of journals http://bmjopen.bmj.com/ 37 38 than the current study, or on journals from a specific therapy area, to further clarify 39 40 the use of open access variants in the medical publications landscape. 41 42 43 CONCLUSIONS 44 on September 30, 2021 by guest. Protected copyright. 45 The CC BY licence is recommended by open access declarations and funders of 46 47 research as the optimal open access licence. Our analysis shows that although 48 49 journals with a high impact factor provide some form of open access, they restrict 50 commercially funded research from being published with the CC BY licence. 51 52 Approximately half of all medical research is funded by the pharmaceutical industry,1 53 54 27 28 meaning that the research output cannot be reused or built upon if it is 55 56 published in journals with a high impact factor without payment of additional fees, 57 hampering research innovation and collaboration. However, there are concerns that 58 59 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 29 BMJ Open

1 2 3 a rapid transition to publishing exclusively with a CC BY licence will be difficult, given 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 current processes and business models in scientific publishing. 6 7 8 The idea that open access to research articles is beneficial to all stakeholders in 9 10 medical research and publishing is compelling. Open access publishing facilitates 11 12 faster and more thorough disclosure of research, removes barriers for groups 13 14 conducting systematic reviews, increases both the citation counts and Altmetric 15 scores of publications, and benefits patient health by improving informed decision- 16 17 making by doctors and patients.9 Commercial research funders lag behind non- 18 For peer review only 19 commercial funders in the implementation of open access policies, and it is time for 20 21 them to close the gap. Pharmaceutical companies should make clear their open 22 access requirements, for example in a unified position statement, ideally aligned with 23 24 open access declarations,18 31 32 the Horizon 2020 programme and Plan S,35-37 and 25 45 34 26 the International Committee of Medical Journal Editors and GPP3 guidelines, and 27 then work together with publishers to realise the ultimate goal of improved access to 28 29 medical research for all. 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 29

1 2 3

Acknowledgements BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 4 5 6 Robert Kiley (https://orcid.org/0000-0003-4733-2558) is Head of Open Research at 7 8 the Wellcome Trust, London, UK, and contributed to the review of this manuscript. 9 Paul Farrow (https://orcid.org/0000-0002-0569-9688) is an employee of Oxford 10 11 PharmaGenesis, Oxford, UK, and contributed significantly to the review of this 12 13 manuscript. Sarah Stokes (https://orcid.org/0000-0002-8761-8588) and Velissaria 14 Vanna are employees of Oxford PharmaGenesis, Oxford, UK, and contributed to the 15 16 review and editing of this manuscript. This work was presented as a poster at both 17 18 the European MeetingFor of thepeer International review Society for onlyMedical Publication 19 20 Professionals (ISMPP) on 23 January 2018 and the Annual Meeting of ISMPP on 21 2 May 2018, and was posted to bioRxiv as a preprint on 22 January 2018 22 23 (https://www.biorxiv.org/content/early/2018/01/22/250613). 24 25 26 Funding statement 27 28 29 This research was funded by Oxford PharmaGenesis. 30 31 32 Competing interests 33 34 35 Tim Ellison, Tim Koder, Amy Williams and Chris Winchester are employees of 36 Oxford PharmaGenesis, Oxford, UK. At the time of the research and writing of this http://bmjopen.bmj.com/ 37 38 manuscript, Laura Schmidt was an employee of Oxford PharmaGenesis, Oxford, UK. 39 40 Chris Winchester is also a Director and a shareholder of Oxford PharmaGenesis 41 Holdings Ltd. 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 Author contributions 46 47 Conceptualization, project administration, TE (https://orcid.org/0000-0003-0307- 48 49 725X), TK (https://orcid.org/0000-0001-6152-7365), LS (https://orcid.org/0000-0001- 50 51 6117-781X), AW (https://orcid.org/0000-0002-9354-6402); methodology, resources, 52 investigation, formal analysis, TE; writing – original draft, TE and LS; visualization, 53 54 TE; writing – review and editing, TE, TK, LS, AW, CW (https://orcid.org/0000-0003- 55 56 3267-3990); supervision, TK, LS. 57 58 59 60

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 29 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 References 6 7 1. Moses H, 3rd, Matheson DH, Cairns-Smith S, et al. The anatomy of medical 8 9 research: US and international comparisons. JAMA 2015;313(2):174–89. doi: 10 11 10.1001/jama.2014.15939. 12 2. Balas EA, Boren SA. Managing clinical knowledge for health care improvement. 13 14 Yearb Med Inform 2000(1):65–70. 15 16 3. Balas EA, Chapman WW. Road map for diffusion of innovation in health care. Health 17 18 Aff (Millwood)For 2018;37(2):198–204. peer review doi: 10.1377/hlthaff.2017.1155. only 19 4. Maggio LA, Moorhead LL, Willinsky JM. Qualitative study of physicians' varied uses 20 21 of biomedical research in the USA. BMJ Open 2016;6(11):e012846. doi: 22 23 10.1136/bmjopen-2016-012846. 24 25 5. Moorhead LL, Holzmeyer C, Maggio LA, et al. In an age of open access to research 26 policies: physician and public health NGO staff research use and policy 27 28 awareness. PLoS One 2015;10(7):e0129708. doi: 29 30 10.1371/journal.pone.0129708. 31 6. Davis PM, Lewenstein BV, Simon DH, et al. Open access publishing, article 32 33 downloads, and citations: randomised controlled trial. BMJ 2008;337:a568. doi: 34 35 10.1136/bmj.a568. 36 http://bmjopen.bmj.com/ 37 7. Ottaviani J. The post-embargo open access citation advantage: it exists (probably), 38 its modest (usually), and the rich get richer (of course). PLoS One 39 40 2016;11(8):e0159614. doi: 10.1371/journal.pone.0159614. 41 42 8. Piwowar H, Priem J, Lariviere V, et al. The state of OA: a large-scale analysis of the 43 44 prevalence and impact of Open Access articles. PeerJ 2018;6:e4375. doi: on September 30, 2021 by guest. Protected copyright. 45 10.7717/peerj.4375. 46 47 9. Tennant JP, Waldner F, Jacques DC, et al. The academic, economic and societal 48 49 impacts of open access: an evidence-based review. F1000Res 2016;5:632. doi: 50 10.12688/f1000research.8460.3. 51 52 10. Wang X, Liu C, Mao W, et al. The open access advantage considering citation, 53 54 article usage and social media attention. Scientometrics 2015;103(2):555–64. 55 56 doi: 10.1007/s11192-015-1547-0. 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 29

1 2 3 11. Allen HG, Stanton TR, Di Pietro F, et al. Social media release increases 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 dissemination of original articles in the clinical pain sciences. PLoS One 6 7 2013;8(7):e68914. doi: 10.1371/journal.pone.0068914. 8 12. Hopewell S, Clarke M, Moher D, et al. CONSORT for reporting randomised trials in 9 10 journal and conference abstracts. Lancet 2008;371(9609):281–3. doi: 11 12 10.1016/S0140-6736(07)61835-2. 13 14 13. Barbour V, Chinnock P, Cohen B, et al. The impact of open access upon public 15 health. Bull World Health Organ 2006;84(5):337–424. 16 17 14. Leung PTM, Macdonald EM, Stanbrook MB, et al. A 1980 letter on the risk of opioid 18 For peer review only 19 addiction. N Engl J Med 2017;376(22):2194–95. doi: 10.1056/NEJMc1700150. 20 21 15. Pastorino R, Milovanovic S, Stojanovic J, et al. Quality assessment of studies 22 published in open access and subscription journals: results of a systematic 23 24 evaluation. PLoS One 2016;11(5):e0154217. doi: 25 26 10.1371/journal.pone.0154217. 27 16. Tahim A, Bansal H, Goodson AM, et al. Open access publishing: a study of current 28 29 practice in oral and maxillofacial surgery research. J Maxillofac Oral Surg 30 31 2016;15(4):517-20. doi: 10.1007/s12663-016-0898-2. 32 33 17. Creative Commons. About The Licenses: Creative Commons; 2017. Available from: 34 https://creativecommons.org/licenses/ (Accessed 29 October 2018). 35 36

18. Budapest Open Access Initiative. Ten years on from the Budapest Open Access http://bmjopen.bmj.com/ 37 38 Initiative: setting the default to open 2012. Available from: 39 40 http://www.budapestopenaccessinitiative.org/boai-10-recommendations 41 (Accessed 2 January 2018). 42 43 19. Williams L. Best practices in licensing and attribution: What you need to know: 44 on September 30, 2021 by guest. Protected copyright. 45 OASPA; 2016. Available from: https://oaspa.org/best-practices-licensing- 46 attribution-need-to-know/ (Accessed 3 January 2018). 47 48 20. Stodden V, Leisch F, Peng RD. Implementing Reproducible Research: Taylor & 49 50 Francis 2014. 51 52 21. Open access policy: Wellcome Trust. Available from: 53 https://wellcome.ac.uk/funding/managing-grant/open-access-policy (Accessed 3 54 55 January 2018). 56 57 22. Charity Open Access Fund. COAF guidelines: Charity Open Access Fund; 2017. 58 Available from: https://wellcome.ac.uk/funding/managing-grant/charity-open- 59 60 access-fund (Accessed 2 January 2018).

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 29 BMJ Open

1 2 3 23. Bill & Melinda Gates Foundation. How We Work: Bill & Melinda Gates Foundation 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Open Access Policy 2015. Available from: https://www.gatesfoundation.org/How- 6 7 We-Work/General-Information/Open-Access-Policy (Accessed 3 January 2018). 8 24. Collins E. Publishing priorities of biomedical research funders. BMJ Open 9 10 2013;3(10):e004171. doi: 10.1136/bmjopen-2013-004171. 11 12 25. Marchington J, Panayi A, Baronikova S, et al. Open access licences: what drives 13 14 publisher options? Suppl 1 (26)2017. Available from: 15 http://www.caudex.com/downloads/OA_survey_EU_ISMPP_2017_poster_15.pdf 16 17 (Accessed 3 January 2018). 18 For peer review only 19 26. Medical Research Council UK. RCUK Policy on Open Access and Supporting 20 21 Guidance 2017. Available from: 22 http://www.rcuk.ac.uk/documents/documents/rcukopenaccesspolicy-pdf/ 23 24 (Accessed 2 January 2018). 25 26 27. Dorsey ER, de Roulet J, Thompson JP, et al. Funding of US biomedical research, 27 2003–2008. JAMA 2010;303(2):137–43. doi: 10.1001/jama.2009.1987. 28 29 28. Hakoum MB, Jouni N, Abou-Jaoude EA, et al. Characteristics of funding of clinical 30 31 trials: cross-sectional survey and proposed guidance. BMJ Open 32 33 2017;7(10):e015997. doi: 10.1136/bmjopen-2017-015997. 34 29. The MAP Newsletter. Shire announces new open access policy: ISMPP; 2018 35 36

updated 20 January 2018. Available from: http://ismpp- http://bmjopen.bmj.com/ 37 38 newsletter.com/2018/01/30/shire-announces-new-open-access-policy/ 39 40 (Accessed 20 July 2018). 41 30. Shire. Shire continues to uphold high standards of ethics and transparency with 42 43 adoption of open access policy for publication of Shire-supported research 2018. 44 on September 30, 2021 by guest. Protected copyright. 45 Available from: https://www.shire.com/en/newsroom/2018/january/xajhds 46 (Accessed 16 July 2018). 47 48 31. Open access Max-Planck-Gesellschaft. Berlin Declaration on Open Access to 49 50 Knowledge in the Sciences and Humanities 2003. Available from: 51 52 https://openaccess.mpg.de/Berlin-Declaration (Accessed 2 January 2018). 53 32. Brown PO, Cabell D, Chakravarti A, et al. Bethesda Statement on Open Access 54 55 Publishing Chevy Chase, Maryland, US 2003. Available from: 56 57 http://legacy.earlham.edu/~peters/fos/bethesda.htm (Accessed 3 January 2018). 58 33. WMA Declaration of Helsinki – Ethical principles for medical research involving 59 60 human subjects: World Medical Association. Available from:

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 29

1 2 3 https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles- 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 for-medical-research-involving-human-subjects/ (Accessed 19 January) 2018. 6 7 34. Battisti WP, Wager E, Baltzer L, et al. Good Publication Practice for Communicating 8 Company-Sponsored Medical Research: GPP3. Ann Intern Med 9 10 2015;163(6):461–4. doi: 10.7326/M15-0288. 11 12 35. Science Europe. cOAlition S 2018. Available from: 13 14 https://www.scienceeurope.org/coalition-s/ (Accessed 29 October 2018). 15 36. Guidelines to the rules on open access to scientific publications and open access to 16 17 research data in Horizon 2020 (Version 3.2): EUROPEAN COMMISSION 18 For peer review only 19 Directorate-General for Research & Innovation; 2017. 11. Available from: 20 21 http://ec.europa.eu/research/participants/data/ref/h2020/grants_manual/hi/oa_pil 22 ot/h2020-hi-oa-pilot-guide_en.pdf (Accessed 17 January 2018). 23 24 37. Guidelines on the Implementation of Open Access to Scientific Publications and 25 26 Research Data in projects supported by the European Research Council under 27 Horizon 2020: European Research Council 2017. Available from: 28 29 https://erc.europa.eu/sites/default/files/document/file/ERC%20Open%20Access 30 31 %20guidelines-Version%201.1._10.04.2017.pdf (Accessed 3 January 2018). 32 33 38. Author Services Supporting Taylor & Francis authors. Open access options: Taylor 34 & Francis. Available from: https://authorservices.taylorandfrancis.com/journal- 35 36

list/ (Accessed 20 July 2018). http://bmjopen.bmj.com/ 37 38 39. Siwek J. Permission fees for reproducing tables in journal articles are exorbitant. 39 40 BMJ 2015;351:h5128. doi: 10.1136/bmj.h5128. 41 40. Smith R. Medical journals and pharmaceutical companies: uneasy bedfellows. BMJ 42 43 2003;326(7400):1202–5. doi: 10.1136/bmj.326.7400.1202. 44 on September 30, 2021 by guest. Protected copyright. 45 41. Lundh A, Barbateskovic M, Hrobjartsson A, et al. Conflicts of interest at medical 46 journals: the influence of industry-supported randomised trials on journal impact 47 48 factors and revenue – cohort study. PLoS Med 2010;7(10):e1000354. doi: 49 50 10.1371/journal.pmed.1000354. 51 52 42. Handel AE, Patel SV, Pakpoor J, et al. High reprint orders in medical journals and 53 pharmaceutical industry funding: case-control study. BMJ 2012;344:e4212. doi: 54 55 10.1136/bmj.e4212. 56 57 43. Science Journals: editorial policies: Science. Available from: 58 http://www.sciencemag.org/authors/science-journals-editorial-policies (Accessed 59 60 14 September) 2018.

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 29 BMJ Open

1 2 3 44. Principles of transparency and best practice in scholarly publishing: DOAJ. 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Available from: https://blog.doaj.org/2018/01/15/principles-of-transparency-and- 6 7 best-practice-in-scholarly-publishing-version-3/ (Accessed 18 January) 2018. 8 45. International Committee of Medical Journal Editors. Recommendations for the 9 10 conduct, reporting, editing, and publication of scholarly work in medical journals: 11 12 International Committee of Medical Journal Editors; 2017. Available from: 13 14 http://www.icmje.org/icmje-recommendations.pdf (Accessed 2 January 2017). 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 29

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Tables 6 7 Table 1 Categorization of journals based on the most open variant of open access 8 9 offered. 10 11 Category Version of article Embargo CC BY licence offered 12 13 available period by the journal? 14 15 1 Published None Yes 16 17 2 Published None No 18 For peer review only 19 20 3 Published/accepted ≤ 12 No 21 months 22 23 CC BY, Creative Commons Attribution. 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Page 21 of 29 BMJ Open

1 2 3 Table 2 Access policies of journals with high impact factors that do not provide open access with Creative Commons licences. 4 Publisher Journals included Open access variants available* 5 6 (n=14) 7 8 Embargo period Version of article 9 10 available 11 American Association for Cancer Cancer Discov None VoR† 12 For peer review only 13 Research journals 6–12 months Accepted 14 15 American College of Physicians Ann Intern Med 6 months Accepted 16 http://bmjopen.bmj.com/ 17 American Medical Association JAMA None VoR‡ 18 19 6 months VoR 20 Massachusetts Medical Society N Engl J Med 6 months VoR 21 22 Nature Publishing Group Nature; Nat Biotechnol; Nat Cell Biol; 6 months Accepted 23

24 Nat Genet; Nat Immunol; Nat Mater; on September 30, 2021 by guest. Protected copyright. 25 26 Nat Med; Nat Methods; Nat Neurosci 27 Wiley-Blackwell World Psychiatry 12 months Accepted 28 29 *Available under the terms specified on the journal website. 30 †Upon payment of USD $3500 AuthorChoice fee. 31 32 ‡Available to read on JAMA Network Reader. 33 VoR, version of record. 34 35 36 37 38 39 40 41 42 21 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open Page 22 of 29

1 2 3 4 Table 3. Open access policies of journals with high impact factors that offer immediate open access with the CC BY licence (n=21) 5 6 Creative Commons licences are shaded. 7 8 9 Publisher Journals Open access variants available* Funding requirements for obtaining 10 11 included (n=21) open access with a CC BY licence 12 For peer review only 13 Embargo Creative Version 14 15 period Commons of article 16 licence available http://bmjopen.bmj.com/ 17 18 American Science; None CC BY VoR The American Association for the 19 20 Association Sci Transl Med None None Accepted Advancement of Science “will allow authors 21 22 for the 6 months None Accepted funded by the Bill & Melinda Gates 23 Advancement 12 months None VoR Foundation to publish their research with 24 on September 30, 2021 by guest. Protected copyright. 25 of Science a CC BY licence”† 26 27 American J Clin Oncol None CC BY VoR Creative Commons licences available only if 28 Society of CC BY-NC-ND funders are “academic institutions, not-for- 29 30 Clinical 6 months None VoR profit organizations, philanthropic foundations 31 32 Oncology 12 months None VoR or government agencies” 33 34 BMJ BMJ None CC BY VoR CC BY licence available for authors “where 35 36 Publishing CC BY-NC the funder requires it” 37 Group 38 39 Cell Press Cancer Cell; Cell; None CC BY VoR Creative Commons licences “available only to 40 41 42 22 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Page 23 of 29 BMJ Open

1 2 3 Cell Metab; CC BY-NC-ND authors covered by a funding body 4 Cell Stem Cell; agreement” (these non-commercial funding 5 12 months None Accepted 6 Immunity bodies are listed on the journal websites) 7 8 Elsevier Eur Urol; None CC BY VoR Creative Commons licences are available to 9 10 Gastroenterology; CC BY-NC-ND authors funded by specific funding bodies 11 J Am Coll 6 months None VoR (these non-commercial funding bodies are 12 For peer review only 13 Cardiol; Lancet; listed on the journal websites) 14 15 Lancet Diabetes

16 http://bmjopen.bmj.com/ 17 Endocrinol; Elsevier has established agreements and 18 Lancet Infect Dis; developed policies to allow authors who 19 20 Lancet Oncol; publish in Elsevier journals to comply with 21 22 Lancet Neurol; manuscript archiving requirements of various 23

24 Lancet Respir funding bodies on September 30, 2021 by guest. Protected copyright. (these non-commercial 25 Med funding bodies are listed on the journal 26 27 websites) 28 29 European Eur Heart J None CC BY VoR “RCUK/Wellcome Trust-funded authors…can 30 Society of CC BY-NC use the CC BY licence for their articles” 31 32 Cardiology CC BY-NC-ND 33 34 None None Accepted 35 36 12 months None Accepted 37 Lippincott Circulation None CC BY VoR “Note that authors funded by RCUK or the 38 39 Williams & CC BY-NC Wellcome Trust may choose the CC BY 40 41 42 23 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open Page 24 of 29

1 2 3 Wilkins CC BY-NC-ND licence if they agree to pay the article 4 processing charge and commercial reuse of 5 6 6–12 months None Accepted the article is not a factor” 7 8 Wiley- CA Cancer J Clin None CC BY VoR “All RCUK and Wellcome Trust-funded 9 10 Blackwell CC BY-NC authors will be directed to the CC BY licence” 11 CC BY-NC-ND 12 For peer review only 13 12–24 months None Accepted 14 15 *Available under the terms specified on the journal website. 16 http://bmjopen.bmj.com/ 17 †The American Association for the Advancement of Science's pilot open access partnership with the Gates Foundation concluded on 30 June 2018.43 18 CC BY, Creative Commons Attribution; NC, Non-Commercial; ND, No Derivatives; RCUK, Research Councils UK; VoR, version of record. 19 20 21 22 23

24 on September 30, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 24 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 25 of 29 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Figure legends 6 7 Figure 1 Flow chart of journals included in this study. 8 9 10 11 12 Figure 2 Medical journals categorized by impact factor and their most open variant of 13 14 open access available (n=35). (A) Impact factor, ≥15.0; (B) Impact factors, 15.0–19.9, 15 16 20.0–29.9 and ≥30.0. 17 CC BY, Creative Commons Attribution licence. 18 For peer review only 19 20 21 22 Figure 3 Article processing charges of journals that offer immediate open access with a 23 24 CC BY licence (n=21). 25 *Details on processing fees are provided at acceptance.43 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

25 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 29

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 N=53 n=18 8 Journals with an Journals removed that impact factor of ≥15.0 9 did not meet inclusion criteria 10 Journals that exclusively 11 publish review articles 12 n=35 (n=16) 13 Journals included Non-medical journals (n=2) 14 in the analysis 15 16 17 Email contact round one 18 Journals that were contactedFor peer review only 19 to clarify information missing/ 20 not clear from journal websites 21 (14/15 journals provided confirmation) 22 23 24 Email contact round two 25 Journals that were contacted 26 to confirm tabulated results 27 (34/35 journals replied and 28 provided confirmation) 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 29 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Category 1: published version of record available upon publication with a CC BY licence 7 8 Category 2: published version of record free to read upon publication (no Creative Commons licence) 9 10 Category 3: published version of record or accepted version first available 6–12 months after publication (no Creative 11 Commons licence) 12 13 A 14 15 16 37% 17 (n=13) 18 For peer review only 60% 19 (n=21) 20 21 22 3% 23 (n=1) 24 25 B 26 16 27 14 28 29 12 30 10 31 32 8 33 6 34 35 Number of journals 4 36 2 http://bmjopen.bmj.com/ 37 38 0 39 15.0–19.9 20.0–29.9 ≥30.0 40 Impact factor range 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 29

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 8 Unknown* 9 USD $3000 10 USD $3750 11 10% 12 (n=2) GBP £3000 13 USD $4700 14 14% USD $5000 15 (n=3) 16 17 5% (n=1) 18 62% For peer review only 19 (n=13) 20 5% (n=1) 21 22 5% (n=1) 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 29 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 1 2 3 4 Supplemental information 5 6 7 8 Table S1 Definitions of CC licences commonly used by medical journals17 9 10 Type Definition 11 12 CC BY Free to distribute and adapt the original work, even commercially, if the 13 original creation and authors are credited 14 15 CC BY-NC Free to adapt the original work non-commercially and, although 16 For peer review only 17 derivative works must also acknowledge the authors and be non- 18 19 commercial, they do not have to be licensed on the same terms 20 CC BY-NC-ND Free to download the original work and share it if the authors are 21 22 credited, but the work cannot be adapted or used commercially 23 24 CC, Creative Commons; CC BY, Creative Commons Attribution; NC, Non-Commercial; ND, 25 26 No Derivatives. 27 28 29 30 31

32 http://bmjopen.bmj.com/ 33 34 35 36 37 38 39

40 on September 30, 2021 by guest. Protected copyright. 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Open access policies of leading medical journals: a cross- sectional study

Journal: BMJ Open ManuscriptFor ID peerbmjopen-2018-028655.R1 review only Article Type: Research

Date Submitted by the 14-Mar-2019 Author:

Primary Subject Medical publishing and peer review Heading:

Secondary Subject Heading: Communication

article processing charges, CC BY, Creative Commons, funding, open Keywords: access, pharmaceutical http://bmjopen.bmj.com/

on September 30, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 34 BMJ Open

1 2 3 Open access policies of leading medical journals: a cross- 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 sectional study 6 7 8 9 Authors: 10 1 11 Tim S Ellison (https://orcid.org/0000-0003-0307-725X), 12 2 13 Tim Koder (https://orcid.org/0000-0001-6152-7365), 14 Laura Schmidt2 (https://orcid.org/0000-0001-6117-781X), 15 16 Amy Williams1 (https://orcid.org/0000-0002-9354-6402), 17 2 18 Christopher C WinchesterFor peer (https://orcid.org/0000-0003-3267-3990) review only 19 20 21 1PharmaGenesis London, London, UK 22 23 2Oxford PharmaGenesis, Oxford, UK 24 25 26 27 28 Correspondence to: 29 30 Dr Tim Ellison 31 Address: PharmaGenesis London, 9 Whitehall, 4th Floor, London, SW1A 2DD 32 33 Email: [email protected] 34 35 Phone: +44 203 675 7065 36 http://bmjopen.bmj.com/ 37 38 39 40 Keywords: article processing charges, commercial, CC BY, Creative Commons, 41 42 funding, open access, pharmaceutical. 43 44 Word count (excluding abstract, references and figures and tables): 3776 on September 30, 2021 by guest. Protected copyright. 45 Abstract word count: 300 46 47 Figures and tables: 6 (excluding 1 supplemental table) 48 49 References: 53 50 51 52 53 54 55 56 57 58 59 60

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 34

1 2 ABSTRACT 3 4 Objectives BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Academic and not-for-profit research funders are increasingly requiring that the 7 research they fund must be published open access, with some insisting on 8 9 publishing with a Creative Commons Attribution (CC BY) licence to allow the 10 11 broadest possible use. We aimed to clarify the open access variants provided by 12 13 leading medical journals and record the availability of the CC BY licence for 14 commercially funded research. 15 16 17 18 Methods For peer review only 19 20 We identified medical journals with a 2015 impact factor of ≥15.0 on 24 May 2017, 21 then excluded from the analysis journals that only publish review articles. Between 22 23 29 June 2017 and 26 July 2017, we collected information about each journal’s open 24 25 access policies from their websites and/or by email contact. We contacted the 26 journals by email again between 6 December 2017 and 2 January 2018 to confirm 27 28 our findings. 29 30 31 32 Results 33 Thirty-five medical journals publishing original research from 13 publishers were 34 35 included in the analysis. All 35 journals offered some form of open access allowing 36 http://bmjopen.bmj.com/ 37 articles to be free-to-read, either immediately on publication or after a delay of up to 38 39 12 months. Of these journals, 21 (60%) provided immediate open access with a 40 CC BY licence under certain circumstances (e.g. to specific research funders). Of 41 42 these 21, 20 only offered a CC BY licence to authors funded by non-commercial 43 44 organizations and one offered this option to any funder who required it. on September 30, 2021 by guest. Protected copyright. 45 46 47 Conclusions 48 49 Most leading medical journals do not offer to authors reporting commercially funded 50 51 research an open access licence that allows unrestricted sharing and adaptation of 52 the published material. The journals’ policies are therefore not aligned with open 53 54 access declarations and guidelines. Commercial research funders lag behind 55 56 academic funders in the development of mandatory open access policies, and it is 57 58 time for them to work with publishers to advance the dissemination of the research 59 they fund. 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 34 BMJ Open

1 2 3

4 Strengths and limitations of this study BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6  This manuscript includes a cross-sectional analysis of open access policies of 7 8 medical journals with a high impact factor, including society-owned journals, 9 10 from multiple publishers. 11  The open access policies of all journals analysed were clarified, and 12 13 confirmation of our findings was received by email from 97% of the contacted 14 15 journals. 16 17  Open access policies of the journals and publishers analysed are subject to 18 change, soFor the information peer presented review here may only change in the future. 19 20  By selecting journals with a high impact factor, our analysis does not include 21 22 prestigious journals from specialized therapy areas and regional or non- 23 24 English language journals, which may have lower impact factors. 25  Some of the journals included in our analysis (e.g. Science, Nature) could be 26 27 considered as interdisciplinary journals rather than exclusively medical 28 29 journals. 30 31 32 INTRODUCTION 33 34 35 Hundreds of billions of US dollars are invested in medical research by governments, 36 charities, and philanthropic and commercial organizations each year, with the aim of http://bmjopen.bmj.com/ 37 38 extending and improving human lives.1 Publication plays an important role in the 39 2 3 40 dissemination of scientific innovation. However, translation of medical research 41 into clinical practice is slow; one study has suggested that it takes an average of 42 43 17 years for research evidence to reach 50% adoption in clinical practice, with the 44 on September 30, 2021 by guest. Protected copyright. 45 longest delays occurring after successful publication of clinical trial results.2 3 46 47 Implementation of research published using the traditional subscription publication 48 model is hindered by copyright restrictions that prohibit reuse of the published 49 50 content and paywalls that prevent public access. 51 52 53 54 Open access publishing has the potential to improve innovation and speed up its 55 adoption. Complete access to research literature encourages viewing of more 56 57 articles than partial access,4 5 and open access articles appear to be downloaded 58 59 more often and receive more citations than subscription articles, indicating a greater 60

3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 34

1 2 academic impact.6-9 There is also evidence suggesting that open access articles 3

4 have a broader societal impact based on altmetric data that measure the attention BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 publications receive in the news and social media.9-11 Depending on the 6 7 restrictiveness of its licensing, open access can facilitate public and commercial 8 9 reuse of research results, which is beneficial for collaboration, education and 10 9 11 innovation. Access to the full text of research articles also increases transparency, 12 benefitting the public by helping both doctors and patients to find complete and 13 14 current evidence to inform treatment decisions, and by preventing potentially harmful 15 16 decisions being made based on the abstracts of paywalled articles.9 12-14 The 17 publishing model used by a journal (i.e. open access or subscription) has no impact 18 For peer review only 19 on the quality of articles published.15 16 20 21 22 23 ‘Open access’ is a broad term that encompasses a range of definitions, from ‘free-to- 24 read’ (full text available to read on demand, without charge to the reader) to ‘free-to- 25 26 read and reuse’ (with the additional ability to reuse text, tables and figures in different 27 28 formats). The Budapest Open Access Initiative,17 the Berlin Declaration,18 the 29 19 20 30 Bethesda Statement and open access advocates define ‘open access’ 31 exclusively as published content that can be read free-of-charge immediately at the 32 33 time of publication with unrestricted reuse rights providing that the original source is 34 35 attributed. Therefore, these open access advocates and declarations recommend 36 http://bmjopen.bmj.com/ 37 open access publishing under a Creative Commons Attribution (CC BY) licence, 38 which allows sharing and adaptation of published materials for any purposes (both 39 40 commercial and non-commercial), subject only to attribution of the original source. 17 41 21 22 42 Common alternatives to the CC BY licence include CC BY Non-Commercial 43 (CC BY-NC), which restricts commercial reuse; CC BY No Derivatives (CC BY-ND), 44 on September 30, 2021 by guest. Protected copyright. 45 which restricts adaptation; and CC BY-NC-ND, which restricts both (table S1).21 23 46 47 48 49 When a journal offers open access, it has wide scope in the choice of policy or 50 policies it will apply, using one of the Creative Commons licences that allow reuse 51 52 under specific terms, or offering free-to-read access without a licence.21 The 53 54 Directory of Open Access Journals (DOAJ) requires journals indexed in the directory 55 to state on their websites clearly and precisely the terms of use and reuse that 56 57 readers and authors have when they submit an article or use the published content. 58 59 60

4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 34 BMJ Open

1 2 DOAJ has a strong preference for the use of Creative Commons licenses, especially 3 24 4 the CC BY licence. BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 At prominent journals, open access with a Creative Commons licence is typically 8 9 facilitated by an article processing charge. Following payment by the research 10 11 author, institution or funder, articles are usually made available on the journal’s 12 website at the time of publication in the publisher’s typeset format (Version of 13 14 Record). Open access articles that do not include a Creative Commons licence at 15 16 the time of publication typically involve an embargo period before the published 17 articles are freely accessible and may allow access only to the accepted manuscript 18 For peer review only 19 (a version that has not been edited and typeset by the journal), which is made 20 21 available on the author’s institutional website, PubMed Central or Europe PubMed 22 23 Central without a requirement for payment. It is noteworthy that the accepted version 24 of a manuscript as well as the Version of Record are sometimes required to bear a 25 26 Creative Commons licence, often including the -NC and/or -ND clause.25 27 28 29 30 There has been an increasing trend towards open access publishing over the last 31 20 years, and almost 50% of articles were published open access in 2015.8 32 33 However, a study analysing global health research articles published between 2010 34 35 and 2014 showed that 69% of these articles were not freely available on the journal’s 36 http://bmjopen.bmj.com/ 37 website and 61% of researchers do not self-archive their work even when journal 38 policy allows them to do so free of charge.26 Many academic and not-for-profit 39 40 research funders now require the research they fund to be published open access.9 41 27-32 42 Prominently, the Wellcome Trust and the Bill & Melinda Gates Foundation insist 43 on publishing with a CC BY licence to allow the broadest possible use.27 29 44 on September 30, 2021 by guest. Protected copyright. 45 Commercial research funders, which fund approximately half of all medical 46 47 research,1 33 34 have been more hesitant to require open access publishing but now 48 30 49 commonly pay for open access when the option is available. Commercial research 50 funders are defined here as pharmaceutical companies and other medical industries 51 52 that fund research for commercial purposes. The proportion of articles authored by 53 54 large pharmaceutical companies that were published open access doubled between 55 2009 and 2016.35 In January 2018, Shire (now part of Takeda) became the first 56 57 commercial research funder to require all research manuscripts it funds to be 58 59 60

5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 34

1 2 published open access.36 37 One year later, Ipsen committed to making its published 3 38 4 scientific research freely accessible to everyone. BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 We set out to clarify the open access variants provided by leading medical journals 8 9 for research in general, and commercially funded research in particular, and 10 11 establish the availability of the CC BY licence for commercially funded research. 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 34 BMJ Open

1 2 3

4 METHODS BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Using Journal Selector (Sylogent, Newtown, PA, USA), we identified medical 7 8 journals with a 2015 impact factor of at least 15.0 (accurate on 24 May 2017). To 9 focus on journals publishing original medical research, we excluded journals that 10 11 only publish review articles. We collected information on the open access variants 12 13 provided by the included journals from their websites and by email contact when 14 15 information was missing or unclear, making up to three attempts between 16 29 June 2017 and 26 July 2017. 17 18 For peer review only 19 20 For each journal, we recorded the following information: 21 22  for immediate open access, whether a CC BY licence or other Creative 23 Commons licence was provided 24 25  for delayed open access, the length of embargo period for open access 26 27  for both immediate and delayed open access, which version of the article 28 29 would be available (published Version of Record or accepted). 30 31 32 For journals that provided a CC BY licence, we additionally collected information on: 33 34  the requirements for obtaining a CC BY licence (e.g. dependence on funding 35 36 source) http://bmjopen.bmj.com/ 37  article processing charges. 38 39 40 41 Between 6 December 2017 and 2 January 2018, we emailed the journals’ editorial 42 43 offices requesting confirmation of our findings. Once open access variants were 44 recorded, we categorized the most open variant provided by each included journal on September 30, 2021 by guest. Protected copyright. 45 46 using our own classification, as shown in table 1. 47 48 49 To gather general information on open access licences and charges available from a 50 51 larger selection of medical journals, we carried out a search on the DOAJ website 52 53 (https://doaj.org/search) on 21 February 2019. 54 55 56 57 58 59 60

7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 34

1 2 Patient and public involvement 3

4 Although patients and/or the public were not directly involved in the design and BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 conduct of this study, patients’ perspectives were sought during the reporting of our 6 7 findings and are included in the Supplemental information. 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 34 BMJ Open

1 2 3

9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 34

1 2 3

4 The cost of open access with a CC BY licence BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 Of the 21 journals that offered a CC BY licence, 19 (90%) disclosed article 7 processing charges on their websites. Across these journals, charges ranged from 8 9 USD $3000 to $5000; the most common article processing charge was $5000 (in 10 11 13 [62%] of journals; figure 3). Of the six journals disclosing an article processing 12 charge of less than $5000, five had an impact factor of less than 20.0, indicating that 13 14 the cost of article processing charges may depend on impact factor. Details of the 15 16 fees charged by the remaining two journals (10%), Science and Science 17 18 Translational Medicine,For were peer not available review from their websitesonly because the details 19 were only provided when the article was accepted.39 20 21 22 23 Relationship between funding source and the availability of open access 24 25 variants 26 Table 3 shows the open access policies of the journals offering open access with a 27 28 CC BY licence. Of the 21 journals listed, 20 journals allowed open access with a 29 30 CC BY licence for research funded by specific non-commercial organizations, and 31 32 only The BMJ offered it to any organization who required it, regardless of the nature 33 of the funding source. 34 35 36 http://bmjopen.bmj.com/ 37 Availability of open access from a larger selection of medical journals 38 39 Of 713 medical journals indexed in the DOAJ database on 21 February 2019: 689 40 (96.6%) supported publishing with a Creative Commons licence and 24 (3.4%) 41 42 supported publishing with the publisher’s own licence; 227 (31.8%) journals charged 43 44 article processing charges for open access; and 257 (36.0%) journals listed CC BY on September 30, 2021 by guest. Protected copyright. 45 as their most restrictive licence regardless of whether there was an associated article 46 47 processing charge. Of the 257 journals that allowed open access publishing with a 48 49 CC BY licence, 108 (42.0%) charged an article processing charge for the opportunity 50 51 and two (0.8%) did not have available information on publication charges. 52 53 54 DISCUSSION 55 56 Here, we present a cross-sectional analysis of open access policies of medical 57 58 journals with a high impact factor, including society-owned journals, from multiple 59 60 publishers. We met our objective to clarify the open access policies of all journals

10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 34 BMJ Open

1 2 3 analysed and received confirmation of our findings by email from 97% of the 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 contacted journals. We found that all leading medical journals in this study provided 6 7 some form of open access, but there was little consistency across their policies. 8 Over half of the included journals provided a CC BY licence; however, with the 9 10 exception of one journal, this option was available only to authors funded by non- 11 12 commercial organizations. One journal (The BMJ) allowed authors to obtain a CC BY 13 14 licence when the work was supported by funders who required its use. Therefore, if 15 commercial organizations, such as pharmaceutical companies had a policy that 16 17 required open access with a CC BY licence, The BMJ would currently be the only 18 For peer review only 19 compliant medical journal with an impact factor greater than 15.0. At the time of our 20 21 analysis, no commercial research funder required open access with a CC BY 22 licence. However, the company at which the analysis was performed, Oxford 23 24 PharmaGenesis, has since updated its publication policy to require open access with 25 40 26 a CC BY licence for the research it funds. 27 28 29 Limitations of this study are that we investigated journals listed in the Journal 30 31 Selector database with an impact factor of at least 15.0, and that, because impact 32 33 factors and the open access policies of journals and publishers are subject to 34 change, the information may change in the future. The validity of the impact factor 35 36

metric is contentious, and its use in this analysis may have led to exclusion of http://bmjopen.bmj.com/ 37 38 prestigious journals from specialized therapy areas and regional or non-English 39 40 language journalsthat have impact factors under 15.0. Furthermore, some of the 41 journals included in our analysis (e.g. Science, Nature) can be considered 42 43 interdisciplinary journals rather than exclusively medical journals. Although our study 44 on September 30, 2021 by guest. Protected copyright. 45 covers only a small number of journals, extending such a manual analysis to a 46 greater number of journals without loss of detail and verification of all results would 47 48 take more time and increase the scope of the study. If more extensive mining of 49 50 journal metadata becomes feasible, , this study could be more easily repeated for a 51 52 bigger cohort of journals. To gather general information on open access licences and 53 charges available from a larger selection of medical journals, we carried out a search 54 55 on the DOAJ website on 21 February 2019. Unlike our manual analysis of medical 56 57 journals with a high impact factor, the search of medical journals indexed in the 58 DOAJ included only those that met the DOAJ criteria to be considered an open 59 60 access journal. Therefore, it is not surprising that the proportion of journals that

11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 34

1 2 3 provided open access with a Creative Commons licence in our manual analysis 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 (60.0%) was lower than that found in the DOAJ search (96.6%). To our surprise, the 6 7 proportion of journals that provided open access with a CC BY licence was higher for 8 the journals in our manual analysis (60.0%) than for the journals identified in the 9 10 DOAJ search (36.0%). However, the DOAJ currently only lists one licence for each 11 12 journal and asks publishers to choose the most restrictive licence, so there is a 13 14 possibility that the CC BY licence is available from a greater proportion of medical 15 journals indexed in the DOAJ. In our manual analysis, all included medical journals 16 17 that provided open access with a CC BY licence required payment of an article 18 For peer review only 19 processing charge, but less than half of the medical journals in the DOAJ charge for 20 21 a CC BY licence. This finding suggests that medical journals with a high impact 22 factor charge more for publishing open access with a CC BY licence than the 23 24 average medical open access journal. Information on whether the availability of the 25 26 CC BY licence is dependent on the funding source could not be easily found using 27 the DOAJ search. 28 29 30 31 To our knowledge, this is the first report showing that the availability of open access 32 33 options depends on the source of funding. A previous study by Solomon and Björk 34 analysing the source of funding for open access publishing across 74 open access 35 36

journals of different disciplines showed that 50% of the open access publications in http://bmjopen.bmj.com/ 37 38 Health Sciences, Biology, and Life Sciences were funded by a grant/contract or 39 41 40 national funding and 30% of the publications were funded by an institution. 41 However, the study did not show that the availability of open access was dependent 42 43 on whether the source of funding is commercial or non-commercial.41 In line with our 44 on September 30, 2021 by guest. Protected copyright. 45 results, the analysis by Solomon and Björk showed that journals with the highest 46 impact factor tended to charge the highest article processing charges.41 Limitations 47 48 on the availability of the CC BY licence depending on the research funder are not in 49 50 line with statements such as the Budapest Declaration,17 the Berlin Declaration18 and 51 19 52 the Bethesda Statement, which aim to provide end users with immediate access to 53 research articles and to give them the opportunity to reuse material without 54 55 restrictions. Furthermore, placing restrictions on access to medical research owing to 56 57 its source of funding is not in line with the key principles of human research ethics 58 laid out in the Declaration of Helsinki.42 59 60

12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 34 BMJ Open

1 2 3 Good Publication Practice 3 (GPP3) guidelines state that authors should take 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 responsibility for the way research findings are published.43 In line with these 6 7 recommendations, commercial companies can and, we believe, should advise 8 authors to reach a consensus on which journal to publish with, to avoid predatory 9 10 journals, and to adhere to sponsor guidelines and regulations. In the authors’ 11 12 experience, some pharmaceutical companies already have internal guidelines 13 14 recommending open access publishing, and two (Shire, now part of Takeda, and 15 Ipsen) now requires it.37 38 16 17 18 For peer review only 19 Our research shows that one-third of the journals with a high impact factor do not 20 21 offer immediate access to the published version of a manuscript upon publication, 22 even though the open access policies of many funders with respect to embargo 23 24 periods echo the recommendations set out by open access declarations 25 17-19 27-29 32 44 26 worldwide. Of note, Horizon 2020, which is supported by the European 27 Research Council, requires its beneficiaries to make publications open access no 28 29 later than 6 months after the official publication date and to make every effort to 30 31 allow for maximum reuse of the materials, whether that be copying, distributing, 32 45 46 33 searching, linking, crawling, mining or some other use. Furthermore, cOAlition S, 34 a group of national research funders with the support of the European Commission 35 36

and the European Council, has committed to Plan S, the key principle of which is http://bmjopen.bmj.com/ 37 38 that scientific publications on research funded by participating national and European 39 44 40 funders must be published open access by 2020. Under the terms of Plan S, 41 authors must retain copyright of their publication with no restrictions, and all 42 43 publications must be published under an immediate open licence (preferably CC BY) 44 44 45 on September 30, 2021 by guest. Protected copyright. 45 that fulfils the requirements defined by the Berlin Declaration. 46 47 48 Policies vary between publishers but also across journals at the same publisher, and 49 50 this is also the case for journals not included in this analysis, as shown, for example, 51 47 52 by Taylor & Francis in their table of the policies of all their journals. Differences in 53 policy have many underlying factors, including the choices of the journals’ academic 54 55 editorial boards and societies. A potential disincentive to publishers offering CC BY 56 57 licences to commercial research funders is the revenue generated from copyright 58 fees and reprints. Permission to reproduce copyrighted materials can cost hundreds 59 60 or even thousands of dollars; for example, the permission fee requested for reuse of

13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 34

1 2 3 a single table containing 40 words in the journal American Family Physician was 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 $4400.48 Reprints can cost significantly more than permissions charges; for example, 6 7 reprint sales from a single clinical trial can total $1 million or more, with a large profit 8 margin.49 9 10 11 12 Research by Lundh et al.50 aimed to quantify reprint revenues as a proportion of 13 14 journal income. Of the six journals investigated, the two European journals, The BMJ 15 and The Lancet, owned by Elsevier, disclosed the information requested. The editors 16 17 of the US journals Archives of Internal Medicine, Annals of Internal Medicine, JAMA 18 For peer review only 19 and the New England Journal of Medicine did not provide the data. For The BMJ, 20 21 reprint revenues constituted 3% of its overall income; The Lancet obtained 41% of its 22 revenue from reprints.50 In The Lancet, commercially funded publications constituted 23 24 a large proportion of highly reprinted articles (63/88) compared with a sample of 25 51 26 control articles from the same journal (23/88). The generation of revenue for 27 publishers from the selling of reprints leaves publishers open to the criticism that bias 28 29 can be introduced into editorial decisions.50 This concern could be addressed by a 30 31 transition to open access publishing exclusively with a CC BY licence. 32 33 34 Two of the journals included in our analysis, Science and Science Translational 35 36

Medicine, both published by the American Association for the Advancement of http://bmjopen.bmj.com/ 37 38 Science, do not disclose article processing charges on their websites;39 instead, they 39 40 provide this information upon their acceptance of an article. This practice does not 41 comply with the DOAJ guidelines,52 which state that processing fees must be stated 42 43 clearly on journal websites in a place that is easy to find for potential authors prior to 44 on September 30, 2021 by guest. Protected copyright. 45 submitting their manuscript. 46 47 48 We found that the open access policies of some journals precluded commercially 49 50 funded research from being published open access, even after an embargo period 51 52 and without a Creative Commons licence. Further analyses could therefore be 53 undertaken to clarify the proportion of journals with this policy and the rationale 54 55 behind this position. Future research could also focus on a larger cohort of journals 56 57 than the current study, or on journals from a specific therapy area, to clarify further 58 the use of open access variants in the medical publications landscape. Future work 59 60 could also involve collecting information on whether medical journals with a high

14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 34 BMJ Open

1 2 3 impact factor allow commercial funders to use preprints or registered reports, which 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 speed up research dissemination and remove publication bias, respectively. For 6 7 example, it would be interesting to see whether journals that do not provide 8 immediate open access options to commercial funders allow research manuscripts 9 10 to be posted as preprints, and therefore support immediate dissemination of the 11 12 results, albeit in a manuscript that has not yet been peer reviewed. 13 14 15 CONCLUSIONS 16 17 18 The CC BY licenceFor is recommended peer byreview open access onlydeclarations and funders of 19 research as the optimal open access licence. Our analysis shows that although 20 21 medical journals with a high impact factor provide some form of open access, they 22 23 restrict commercially funded research from being published with the CC BY licence, 24 25 meaning that the research output cannot be reused or built upon if it is published in 26 journals with a high impact factor without payment of additional fees. These 27 28 restrictions hamper the further development and implementation of the 29 30 approximately half of all medical research that is funded by commercial research 31 funders.1 33 34 32 33 34 35 Open access publishing facilitates faster and more thorough disclosure of research, 36 http://bmjopen.bmj.com/ 37 removes barriers for groups conducting systematic reviews, increases both the 38 citation counts and altmetric scores of publications, and benefits patient health by 39 40 improving informed decision-making by doctors and patients.9 Commercial research 41 42 funders lag behind non-commercial funders in the implementation of open access 43 44 policies, and we believe that it is time for them to close the gap. Commercial on September 30, 2021 by guest. Protected copyright. 45 companies could, and we believe should, make clear their open access 46 47 requirements, for example in a unified position statement, ideally aligned with open 48 17-19 44-46 49 access declarations, the Horizon 2020 programme and Plan S, and the 50 International Committee of Medical Journal Editors53 and GPP343 guidelines, and 51 52 then work together with publishers to realise the ultimate goal of improved access to 53 54 medical research for all. 55 56 57 58 59 60

15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 34

1 2 3

Acknowledgements BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 4 5 6 Robert Kiley (https://orcid.org/0000-0003-4733-2558) is Head of Open Research at 7 8 the Wellcome Trust, London, UK, and contributed to the review of this manuscript. 9 Paul Farrow (https://orcid.org/0000-0002-0569-9688) is an employee of Oxford 10 11 PharmaGenesis, Oxford, UK, and contributed significantly to the review of this 12 13 manuscript. Sarah Stokes (https://orcid.org/0000-0002-8761-8588) and Velissaria 14 Vanna are employees of Oxford PharmaGenesis, Oxford, UK, and contributed to the 15 16 review and editing of this manuscript. The authors also thank Alan Thomas and 17 18 Elizabeth Kinder forFor their reviewpeer of this review article from the only patient perspective. This work 19 20 was presented as a poster at both the European Meeting of the International Society 21 for Medical Publication Professionals (ISMPP) on 23 January 2018 and the Annual 22 23 Meeting of ISMPP on 2 May 2018, and was posted to bioRxiv as a preprint on 22 24 25 January 2018 (https://www.biorxiv.org/content/early/2018/01/22/250613). 26 27 28 Funding statement 29 30 31 This research was funded by Oxford PharmaGenesis. 32 33 34 Competing interests 35 36 Tim Ellison, Tim Koder and Chris Winchester are employees of Oxford http://bmjopen.bmj.com/ 37 38 PharmaGenesis, Oxford, UK. At the time of the research and writing of this 39 40 manuscript, Laura Schmidt and Amy Williams were employees of Oxford 41 PharmaGenesis, Oxford, UK and are currently employed by Comradis and dna 42 43 Communications, respectively. Chris Winchester is also a Director and a shareholder 44 on September 30, 2021 by guest. Protected copyright. 45 of Oxford PharmaGenesis Holdings Ltd. 46 47 48 Author contributions 49 50 51 Conceptualization, project administration, TE (https://orcid.org/0000-0003-0307- 52 725X), TK (https://orcid.org/0000-0001-6152-7365), LS (https://orcid.org/0000-0001- 53 54 6117-781X), AW (https://orcid.org/0000-0002-9354-6402); methodology, resources, 55 56 investigation, formal analysis, TE; writing – original draft, TE and LS; visualization, 57 58 TE; writing – review and editing, TE, TK, LS, AW, CW (https://orcid.org/0000-0003- 59 3267-3990); supervision, TK, LS. 60

16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 34 BMJ Open

1 2 3

Data availability statement BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 4 5 6 All data relevant to the study are included in the article or uploaded as 7 8 supplementary information. 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 34

1 2 3

18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 34 BMJ Open

1 2 3 11. Allen HG, Stanton TR, Di Pietro F, et al. Social media release increases 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 dissemination of original articles in the clinical pain sciences. PLoS One 6 7 2013;8(7):e68914. doi: 10.1371/journal.pone.0068914. 8 12. Hopewell S, Clarke M, Moher D, et al. CONSORT for reporting randomised trials in 9 10 journal and conference abstracts. Lancet 2008;371(9609):281–3. doi: 11 12 10.1016/S0140-6736(07)61835-2. 13 14 13. Barbour V, Chinnock P, Cohen B, et al. The impact of open access upon public 15 health. Bull World Health Organ 2006;84(5):337–424. 16 17 14. Leung PTM, Macdonald EM, Stanbrook MB, et al. A 1980 letter on the risk of opioid 18 For peer review only 19 addiction. N Engl J Med 2017;376(22):2194–95. doi: 10.1056/NEJMc1700150. 20 21 15. Pastorino R, Milovanovic S, Stojanovic J, et al. Quality assessment of studies 22 published in open access and subscription journals: results of a systematic 23 24 evaluation. PLoS One 2016;11(5):e0154217. doi: 25 26 10.1371/journal.pone.0154217. 27 16. Tahim A, Bansal H, Goodson AM, et al. Open access publishing: a study of current 28 29 practice in oral and maxillofacial surgery research. J Maxillofac Oral Surg 30 31 2016;15(4):517-20. doi: 10.1007/s12663-016-0898-2. 32 33 17. Budapest Open Access Initiative. Ten years on from the Budapest Open Access 34 Initiative: setting the default to open 2012. Available from: 35 36

http://www.budapestopenaccessinitiative.org/boai-10-recommendations http://bmjopen.bmj.com/ 37 38 (Accessed 2 January 2018). 39 40 18. Open access Max-Planck-Gesellschaft. Berlin Declaration on Open Access to 41 Knowledge in the Sciences and Humanities 2003. Available from: 42 43 https://openaccess.mpg.de/Berlin-Declaration (Accessed 2 January 2018). 44 on September 30, 2021 by guest. Protected copyright. 45 19. Brown PO, Cabell D, Chakravarti A, et al. Bethesda Statement on Open Access 46 Publishing Chevy Chase, Maryland, US 2003. Available from: 47 48 http://legacy.earlham.edu/~peters/fos/bethesda.htm (Accessed 3 January 2018). 49 50 20. Carroll MW. Why full open access matters. PLoS Biol 2011;9(11):e1001210. doi: 51 52 10.1371/journal.pbio.1001210. 53 21. Creative Commons. About The Licenses: Creative Commons; 2017. Available from: 54 55 https://creativecommons.org/licenses/ (Accessed 29 October 2018). 56 57 22. Williams L. Best practices in licensing and attribution: What you need to know: 58 OASPA; 2016. Available from: https://oaspa.org/best-practices-licensing- 59 60 attribution-need-to-know/ (Accessed 3 January 2018).

19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 34

1 2 3 23. Stodden V, Leisch F, Peng RD. Implementing Reproducible Research: Taylor & 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Francis 2014. 6 7 24. Directory of Open Access Journals. Information for publishers. Available from: 8 https://doaj.org/publishers#licensing (Accessed 21 February 2019). 9 10 25. Elsevier. Open access licences. Available from: 11 12 https://www.elsevier.com/about/policies/open-access-licenses (Accessed 21 13 14 February 2019). 15 26. Smith E, Haustein S, Mongeon P, et al. Knowledge sharing in global health 16 17 research - the impact, uptake and cost of open access to scholarly literature. 18 For peer review only 19 Health Res Policy Syst 2017;15(1):73. doi: 10.1186/s12961-017-0235-3. 20 21 27. Open access policy: Wellcome Trust. Available from: 22 https://wellcome.ac.uk/funding/managing-grant/open-access-policy (Accessed 3 23 24 January 2018). 25 26 28. Charity Open Access Fund. COAF guidelines: Charity Open Access Fund; 2017. 27 Available from: https://wellcome.ac.uk/funding/managing-grant/charity-open- 28 29 access-fund (Accessed 2 January 2018). 30 31 29. Bill & Melinda Gates Foundation. How We Work: Bill & Melinda Gates Foundation 32 33 Open Access Policy 2015. Available from: https://www.gatesfoundation.org/How- 34 We-Work/General-Information/Open-Access-Policy (Accessed 3 January 2018). 35 36

30. Collins E. Publishing priorities of biomedical research funders. BMJ Open http://bmjopen.bmj.com/ 37 38 2013;3(10):e004171. doi: 10.1136/bmjopen-2013-004171. 39 40 31. Marchington J, Panayi A, Baronikova S, et al. Open access licences: what drives 41 publisher options? Suppl 1 (26)2017. Available from: 42 43 http://www.caudex.com/downloads/OA_survey_EU_ISMPP_2017_poster_15.pdf 44 on September 30, 2021 by guest. Protected copyright. 45 (Accessed 3 January 2018). 46 32. Medical Research Council UK. RCUK Policy on Open Access and Supporting 47 48 Guidance 2017. Available from: 49 50 http://www.rcuk.ac.uk/documents/documents/rcukopenaccesspolicy-pdf/ 51 52 (Accessed 2 January 2018). 53 33. Dorsey ER, de Roulet J, Thompson JP, et al. Funding of US biomedical research, 54 55 2003–2008. JAMA 2010;303(2):137–43. doi: 10.1001/jama.2009.1987. 56 57 34. Hakoum MB, Jouni N, Abou-Jaoude EA, et al. Characteristics of funding of clinical 58 trials: cross-sectional survey and proposed guidance. BMJ Open 59 60 2017;7(10):e015997. doi: 10.1136/bmjopen-2017-015997.

20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 34 BMJ Open

1 2 3 35. Yegros-Yegros A, van Leeuwen T. Production and uptake of Open Access 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 publications involving the private sector: the case of big pharma (preprint). 6 7 SocArXiv 2019 doi: https://doi.org/10.31235/osf.io/zt6kc 8 36. The MAP Newsletter. Shire announces new open access policy: ISMPP; 2018 9 10 updated 20 January 2018. Available from: http://ismpp- 11 12 newsletter.com/2018/01/30/shire-announces-new-open-access-policy/ 13 14 (Accessed 20 July 2018). 15 37. Shire. Shire continues to uphold high standards of ethics and transparency with 16 17 adoption of open access policy for publication of Shire-supported research 2018. 18 For peer review only 19 Available from: https://www.shire.com/en/newsroom/2018/january/xajhds 20 21 (Accessed 16 July 2018). 22 38. Ipsen commits to making all its published scientific research freely accessible to 23 24 everyone: Ipsen. Available from: https://www.ipsen.com/ipsen-commits-to- 25 26 making-all-its-published-scientific-research-freely-accessible-to-everyone/ 27 (Accessed 25 February 2019). 28 29 39. Science Journals: editorial policies: Science. Available from: 30 31 http://www.sciencemag.org/authors/science-journals-editorial-policies (Accessed 32 33 14 September) 2018. 34 40. Oxford PharmaGenesis launches open access publication policy: Oxford 35 36

PharmaGenesis. Available from: https://www.pharmagenesis.com/oxford- http://bmjopen.bmj.com/ 37 38 pharmagenesis-launches-open-access-publication-policy/#.XHP-LIj7SUk 39 40 (Accessed 25 February 2019). 41 41. Solomon DJ, Björk B. Publication fees in open access publishing: sources of 42 43 funding and factors influencing choice of journal. Journal of the American 44 on September 30, 2021 by guest. Protected copyright. 45 Society for Information Science & Technology 2012;63(1):98–107. 46 42. WMA Declaration of Helsinki – Ethical principles for medical research involving 47 48 human subjects: World Medical Association. Available from: 49 50 https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles- 51 52 for-medical-research-involving-human-subjects/ (Accessed 19 January) 2018. 53 43. Battisti WP, Wager E, Baltzer L, et al. Good Publication Practice for Communicating 54 55 Company-Sponsored Medical Research: GPP3. Ann Intern Med 56 57 2015;163(6):461–4. doi: 10.7326/M15-0288. 58 44. Science Europe. cOAlition S 2018. Available from: 59 60 https://www.scienceeurope.org/coalition-s/ (Accessed 29 October 2018).

21 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 34

1 2 3 45. Guidelines to the rules on open access to scientific publications and open access to 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 research data in Horizon 2020 (Version 3.2): EUROPEAN COMMISSION 6 7 Directorate-General for Research & Innovation; 2017. 11. Available from: 8 http://ec.europa.eu/research/participants/data/ref/h2020/grants_manual/hi/oa_pil 9 10 ot/h2020-hi-oa-pilot-guide_en.pdf (Accessed 17 January 2018). 11 12 46. Guidelines on the Implementation of Open Access to Scientific Publications and 13 14 Research Data in projects supported by the European Research Council under 15 Horizon 2020: European Research Council 2017. Available from: 16 17 https://erc.europa.eu/sites/default/files/document/file/ERC%20Open%20Access 18 For peer review only 19 %20guidelines-Version%201.1._10.04.2017.pdf (Accessed 3 January 2018). 20 21 47. Author Services Supporting Taylor & Francis authors. Open access options: Taylor 22 & Francis. Available from: https://authorservices.taylorandfrancis.com/journal- 23 24 list/ (Accessed 20 July 2018). 25 26 48. Siwek J. Permission fees for reproducing tables in journal articles are exorbitant. 27 BMJ 2015;351:h5128. doi: 10.1136/bmj.h5128. 28 29 49. Smith R. Medical journals and pharmaceutical companies: uneasy bedfellows. BMJ 30 31 2003;326(7400):1202–5. doi: 10.1136/bmj.326.7400.1202. 32 33 50. Lundh A, Barbateskovic M, Hrobjartsson A, et al. Conflicts of interest at medical 34 journals: the influence of industry-supported randomised trials on journal impact 35 36

factors and revenue – cohort study. PLoS Med 2010;7(10):e1000354. doi: http://bmjopen.bmj.com/ 37 38 10.1371/journal.pmed.1000354. 39 40 51. Handel AE, Patel SV, Pakpoor J, et al. High reprint orders in medical journals and 41 pharmaceutical industry funding: case-control study. BMJ 2012;344:e4212. doi: 42 43 10.1136/bmj.e4212. 44 on September 30, 2021 by guest. Protected copyright. 45 52. Principles of transparency and best practice in scholarly publishing: DOAJ. 46 Available from: https://blog.doaj.org/2018/01/15/principles-of-transparency-and- 47 48 best-practice-in-scholarly-publishing-version-3/ (Accessed 18 January) 2018. 49 50 53. International Committee of Medical Journal Editors. Recommendations for the 51 52 conduct, reporting, editing, and publication of scholarly work in medical journals: 53 International Committee of Medical Journal Editors; 2017. Available from: 54 55 http://www.icmje.org/icmje-recommendations.pdf (Accessed 2 January 2017). 56 57 58 59 60

22 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 34 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Tables 6 7 Table 1 Categorization of journals based on the most open variant of open access 8 9 offered. 10 11 Category Version of article Embargo CC BY licence offered 12 13 available period* by the journal? 14 15 1 Published None Yes 16 17 2 Published None No 18 For peer review only 19 20 3 Published/accepted ≤ 12 No 21 months 22 23 *None=immediate open access; > 0 months=delayed open access. 24 25 CC BY, Creative Commons Attribution. 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

23 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open Page 24 of 34

1 2 3 Table 2 Access policies of journals with high impact factors that do not provide open access with Creative Commons licences. 4 Publisher Organisation Journals included Open access variants 5 6 status (n=14) available* 7 8 Embargo Version of 9 10 period† article 11 available 12 For peer review only 13 American Association for Nonprofit Cancer Discov None VoR‡ 14 15 Cancer Research journals society

16 http://bmjopen.bmj.com/ 17 6–12 months Accepted 18 American College of Nonprofit Ann Intern Med 6 months Accepted 19 20 Physicians society 21 22 American Medical Nonprofit JAMA None VoR§ 23

24 Association society on September 30, 2021 by guest. Protected copyright. 25 6 months VoR 26 27 Massachusetts Medical Nonprofit N Engl J Med 6 months VoR 28 29 Society society 30 31 Nature Publishing Group Commercial Nature; Nat Biotechnol; Nat 6 months Accepted 32 Cell Biol; 33 34 Nat Genet; Nat Immunol; Nat 35 36 Mater; 37 38 Nat Med; Nat Methods; Nat 39 Neurosci 40 41 42 24 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Page 25 of 34 BMJ Open

1 2 3 Wiley-Blackwell Commercial World Psychiatry 12 months Accepted 4 *Available under the terms specified on the journal website. 5 6 †None=immediate open access; > 0 months=delayed open access. 7 ‡Upon payment of USD $3500 AuthorChoice fee. 8 9 §Available to read on JAMA Network Reader. 10 VoR, version of record. 11 12 For peer review only 13 14 15

16 http://bmjopen.bmj.com/ 17 18 19 20 21 22 23

24 on September 30, 2021 by guest. Protected copyright. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 25 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open Page 26 of 34

16 period† Commons of article http://bmjopen.bmj.com/ 17 licence available 18 19 American Nonprofit Science; None CC BY VoR The American Association for the Advancement of 20 Association society Sci Transl Science “will allow authors funded by the Bill & 21 22 for the Med None None Accepted Melinda Gates Foundation to publish their research 23

24 Advancement 6 months None Accepted with a CC BY licence”‡ on September 30, 2021 by guest. Protected copyright. 25 of Science 12 months None VoR 26 27 American Nonprofit J Clin None CC BY VoR Creative Commons licences available only if funders 28 29 Society of society Oncol CC BY-NC-ND are “academic institutions, not-for-profit 30 Clinical 6 months None VoR organizations, philanthropic foundations or 31 32 Oncology 12 months None VoR government agencies” 33 BMJ Non-profit BMJ None CC BY VoR CC BY licence available for authors “where the 34 35 Publishing society CC BY-NC funder requires it” 36 37 Group 38 Cell Press Commerci Cancer None CC BY VoR Creative Commons licences “available only to 39 40 al Cell; Cell; CC BY-NC-ND authors covered by a funding body agreement” 41 42 26 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

Page 27 of 34 BMJ Open

1 2 3 Cell Metab 12 months None Accepted (these non-commercial funding bodies are listed on 4 ; the journal websites) 5 6 Cell Stem 7 Cell; 8 9 Immunity 10 11 Elsevier Commerci Eur Urol; None CC BY VoR Creative Commons licences are available to authors 12 al GastroentFor peerCC BY-NC-ND reviewAccepted§ fundedonly by specific funding bodies (these non- 13 14 erology; /VoR commercial funding bodies are listed on the journal 15 J Am Coll 6 months None VoR websites) 16 http://bmjopen.bmj.com/ 17 Cardiol; 18 19 Lancet; Elsevier has established agreements and developed 20 Lancet Dia policies to allow authors who publish in Elsevier 21 22 betes journals to comply with manuscript archiving 23 Endocrinol requirements of various funding bodies (these non- 24 on September 30, 2021 by guest. Protected copyright. 25 ; commercial funding bodies are listed on the journal 26 Lancet Inf websites) 27 28 ect Dis; 29 30 Lancet 31 Oncol; 32 33 Lancet Ne 34 urol; 35 36 Lancet Re 37 38 spir Med 39 40 41 42 27 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from

BMJ Open Page 28 of 34

1 2 3 European Nonprofit Eur Heart None CC BY VoR “RCUK/Wellcome Trust-funded authors…can use 4 Society of society J CC BY-NC the CC BY licence for their articles” 5 6 Cardiology CC BY-NC-ND 7 None None Accepted 8 9 12 months None Accepted 10 11 Lippincott Commerci Circulation None CC BY VoR “Note that authors funded by RCUK or the Wellcome 12 Williams & al For peerCC BY-NC review Trustonly may choose the CC BY licence if they agree to 13 14 Wilkins CC BY-NC-ND pay the article processing charge and commercial 15

6–12 None Accepted reuse of the article http://bmjopen.bmj.com/ is not a factor” 16 17 months 18 19 Wiley- Commerci CA None CC BY VoR “All RCUK and Wellcome Trust-funded authors will 20 Blackwell al Cancer J CC BY-NC be directed to the CC BY licence” 21 22 Clin CC BY-NC-ND 23

12–24 None Accepted on September 30, 2021 by guest. Protected copyright. 24 25 months 26 27 *Available under the terms specified on the journal website. 28 †None = immediate open access; > 0 months = delayed open access. 29 ‡The American Association for the Advancement of Science's pilot open access partnership with the Gates Foundation concluded on 30 June 2018.39 30 31 §Accepted manuscripts can be self-archived and are required to attach a CC BY-NC-ND licence.25 32 CC BY, Creative Commons Attribution; NC, Non-Commercial; ND, No Derivatives; RCUK, Research Councils UK; VoR, version of record. 33 34 35 36 37 38 39 40 41 42 28 43 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 29 of 34 BMJ Open

1 2 3

4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Figure legends 6 7 Figure 1 Flow chart of journals included in this study. 8 9 10 11 12 Figure 2 Medical journals categorized by impact factor and their most open variant of 13 14 open access available (n=35). (A) Impact factor, ≥15.0; (B) Impact factors, 15.0–19.9, 15 16 20.0–29.9 and ≥30.0. 17 CC BY, Creative Commons Attribution licence. 18 For peer review only 19 20 21 22 Figure 3 Article processing charges of journals that offer immediate open access with a 23 24 CC BY licence (n=21). 25 *Details on processing fees are provided at acceptance.39 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

29 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 30 of 34

1 2 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 31 of 34 BMJ Open

1 2 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 32 of 34

1 2 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 33 of 34 BMJ Open

1 2 3

4 Supplemental information BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 6 7 Table S1 Definitions of CC licences commonly used by medical journals17 8 9 Type Definition 10 11 CC BY Free to distribute and adapt the original work, even commercially, 12 if the original creation and authors are credited 13 14 CC BY-NC Free to adapt the original work non-commercially and, although 15 16 derivative works must also acknowledge the authors and be non- 17 18 Forcommercial, peer they doreview not have to beonly licensed on the same terms 19 CC BY-NC-ND Free to download the original work and share it if the authors are 20 21 credited, but the work cannot be adapted or used commercially 22 23 CC, Creative Commons; CC BY, Creative Commons Attribution; NC, Non-Commercial; ND, No 24 Derivatives. 25 26 27 28 Patients’ perspectives 29 After reviewing a draft of our manuscript, Alan Thomas, a patient advocate for rare 30 31 diseases, shared his experiences. He said, “I regularly encounter paywalls while 32 33 trying to find high quality and important information on a rare disease for myself and 34 to share with other patients.” Because patients can build up substantial 35 36 understanding during a lifetime living with their rare disease, they are often able to http://bmjopen.bmj.com/ 37 38 read and understand articles in peer-reviewed journals. However, “when we hit a 39 40 paywall, we typically give up and look for information elsewhere,” he said. Elizabeth 41 Kinder, a patient with multiple sclerosis, highlighted the importance of patients 42 43 knowing about the latest research to make informed choices about treatments. She 44 on September 30, 2021 by guest. Protected copyright. 45 said “It is absolutely essential that validated scientific findings are as easily and 46 freely available as all the dangerous false-hope-fake-news. A friend of mine spent 47 48 his life savings on a pseudo-scientific procedure and after a brief respite, was 49 50 considerably more disabled than before. And broke. Which increased the negative 51 52 impact on his health. If publishers refuse to share information they are as immoral as 53 those snake oil salesman and as dangerous to public health. Patients like me have 54 55 to become our own experts. Faced with a steadily crippling disease like multiple 56 57 sclerosis, it is not just in our own interests but in the interests of society as a whole 58 59 that we are able to work, to contribute, to play our part, and for our families to do so, 60 to not become our carers.”

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 34 of 34

1 2 3 4 BMJ Open: first published as 10.1136/bmjopen-2018-028655 on 20 June 2019. Downloaded from 5 Elizabeth Kinder is mindful of the risks she took when taking part in a clinical trial, 6 7 saying “I think ‘Sudden death’ was no. 3 in the long list of side effects on which I 8 9 signed off).” She believes that “by withholding free access to research findings, 10 journals are scorning the funders, the researchers, and people like me (patients) who 11 12 risk their lives participating in a clinical trial”. Alan Thomas noted that many patients 13 14 do not realize that scientific content is put behind paywalls by the journals in which it 15 is published, and that they could misconstrue such information as being hidden by 16 17 pharmaceutical companies or others. He wanted to encourage pharmaceutical 18 For peer review only 19 companies to be more open and collaborative by making their research open 20 21 access. He said, “open access to medical documents enables communication with 22 patients as well as to patients, and is vital to patients’ wellbeing and peace of mind”. 23 24 25 26 27 28 29 30 31 32 33 34 35 36 http://bmjopen.bmj.com/ 37 38 39 40 41 42 43 44 on September 30, 2021 by guest. Protected copyright. 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml