Elizabeth Wager

Introduction: why should publishers be concerned about ethics?

Scholarly publications are influential because they can affect decisions with important consequences for individuals and society. For example, papers in medical journals may influence doctors’ decisions about how to treat individual patients or health policy decisions affecting society at large. If the published reports of medical research are unreliable or unclear, patients and society may be hurt. Perhaps because of this, medical publishing has often taken a lead in discussing ethical issues and setting standards, but there is no reason to think that other fields are immune from ethical problems. Publications in disciplines such as engineering, environmental science and agriculture can have important implications for society. Therefore, those who produce scholarly publications, i.e. publishers, carry ethical responsibilities for the integrity of their publications which they share with other players such as editors, peer reviewers and authors.

Peer-reviewed publications can also affect the lives of academics because they are used to measure the productivity of individual researchers and entire departments. Academic appointments and careers therefore depend heavily on publication records. Those who act as editors or reviewers of scholarly journals therefore wield considerable power as they can influence what gets published. Peer review systems should therefore be as fair as possible and designed to minimise bias and ensure consistent treatment for all authors; publishers can play an important role in ensuring that they are.

Another reason why publishers should be concerned about ethics is because they are professionals. Many of the other key players in peer-reviewed publications, whatever their skills or expertise, will be volunteers or amateurs (i.e. not doing this role as their main occupation and very often doing it without any formal training). Most academic editors take on this role on top of other academic commitments and without any specific training. It often therefore falls to the full-time, professional publishing staff to advise on policies and practices and to assist editors when faced with ethical issues.

Because cases of serious misconduct, such as data fabrication, occur infrequently, few academic editors will have experience of handling these issues more than once in their editorial career. However, companies with long-standing experience of publishing are well placed to develop policies and advise editors on best practice on the basis of this experience.

Lastly, following ethical guidelines can make good business sense, as having good policies can prevent many of the costly and time-consuming problems that can arise in scholarly publishing, or make them easier to resolve, and, in extreme cases, may reduce the risk of litigation.

What can go wrong if scholarly publishing is unethical?

Publication ethics is concerned with various types of misconduct and unethical behaviour committed by authors, editors, reviewers and publishers. The worst forms of publication misconduct committed by authors are generally held to be data fabrication, data falsification and major plagiarism (sometimes termed FFP). Other forms of misconduct or questionable practices by authors include failure to disclose conflicts of interest and misleading authorship. Journal editors and peer reviewers may also abuse their positions and commit misconduct, for example by attempting to suppress or delay the work of rivals or inappropriately promoting or publishing their own work or that of their friends or families.

The publication of unreliable (or fraudulent) data can have adverse consequences for researchers, funders and others, such as doctors, who base decisions on the research. It may even pose a risk to members of the public. For example, Potti et al. (2006) described a technique for predicting patients’ responses to different types of chemotherapy. This technique was used in a clinical trial to determine which treatment cancer patients received. However, while the trial was underway, the research was shown to be unreliable and the trial was stopped (Baggerly and Coombes, 2009). The publication was later retracted (after the work was shown to be fraudulent) but its use as the basis for a clinical trial probably resulted in patients receiving suboptimal treatments, which may have reduced their chances of survival.

Although it is rare to be able to demonstrate serious harm caused directly by a single publication, there is considerable evidence that biases in the published literature can undermine decision-making. For example, national or institutional decisions about which treatment to prescribe for a particular condition are generally based on systematic reviews of the medical literature. If a manufacturer does not publish the less favourable studies of its drug, or suppresses those that suggest harmful effects, the evidence will be incomplete and therefore unreliable. The evidence may be further skewed if manufacturers repeatedly publish positive findings without making it clear that they are presenting the same data several times. Several studies have shown that such publication bias is common and, if undetected, can undermine medical decision-making (Tramer et al., 1997; Melander et al., 2003; Chan et al., 2004).

Improper behaviour by reviewers can harm other researchers, especially if it goes undetected. In the 1980s, scientists working for Cistron, a US biotechnology company, sequenced the gene for interleukin-1 and submitted the results to Nature. Their paper was reviewed by a scientist working for Immunex, a rival company, who recommended rejection. The paper was therefore rejected by Nature and submitted to the Proceedings of the National Academy of Sciences (PNAS) which accepted it. Meanwhile, both companies filed a patent for the gene sequence.

However, the sequence filed by Immunex contained seven errors that had appeared only in the manuscript submitted to Nature, and not in the final paper published in PNAS, strongly implicating the peer reviewer in stealing the sequence. After a lengthy legal battle, Immunex agreed to pay $21 million to Cistron in an out-of-court settlement (Rennie, 2003). While such extreme misconduct is probably rare, many academics believe their work has been held back by rivals during the peer review process, or their ideas have been stolen.

Ethical issues may have legal implications for journals and therefore for their publishers. If peer review and publication are not handled appropriately there is a risk that the publisher may be sued. In 2006, an authorship dispute that was inappropriately handled by the British Journal of Obstetrics & Gynaecology caused the journal to be named in a US court case costing the publisher considerable amounts in legal fees and expenses (Anon., 2006).

What can publishers do to prevent, detect and respond to research and publication misconduct?

Because scholarly journals provide a permanent record of research findings and these findings can influence important decisions affecting individuals and society, publishers and editors have a duty to ensure the integrity of everything they publish. This means they should endeavour to prevent and detect research and publication misconduct such as data fabrication and plagiarism. As prevention is not always possible, and honest errors may also occur, editors and publishers should also have systems in place to minimise the harm caused by the publication of fraudulent or unreliable work.

Journal policies and instructions to contributors may encourage good practice and discourage misconduct. Many journals publish their policies on ethical issues such as plagiarism and authorship in the hope of educating and guiding authors. Journal policies can also have a direct influence on researchers’ behaviour. For example, many commentators have expressed concern about the non-publication of negative studies, redundant publication of positive findings and misleading reporting of trial designs and suggested these problems could be reduced if clinical trials were registered at inception. The first proposal for a public trial register was made by Simes (1986) but widespread registration did not occur until the International Committee of Medical Journal Editors announced in 2004 that their journals would not publish findings from clinical trials unless they had been registered (De Angelis et al., 2004). The influence of the journals’ policies was clearly shown by the marked rise in registrations around the cut-off date for registering studies that had started before the policy was announced (Zarin et al., 2005).

Journals may actively screen for some forms of publication misconduct. Text-matching software can be used to detect plagiarism and duplicate publication. The usefulness of this technology for scholarly publishing has been greatly enhanced by the CrossCheck initiative which, using the iThenticate® system from iParadigms, enables text to be compared against a large database of academic publications that would not be included in a simple search (such as one using Google™), which is limited to searching for text on publicly available websites and therefore cannot ‘see’ text from journals that is behind access barriers (http://www.crossref. org/crosscheck/index.html). The creation of the CrossCheck database by CrossRef was only possible due to cooperation between publishers and it works because those that use the system to check submissions also contribute their publications to the database. As well as detecting cases of plagiarism and redundancy before publication, publicising journal screening policies may act as a deterrent (Kleinert, 2011).

Software can also be utilised to detect inappropriately manipulated (i.e. falsified) images. These techniques were pioneered by the Journal of Cell Biology (JCB) and have been adopted by journals in other disciplines (Rossner and Yamada, 2004). As with other forms of misconduct, routine screening revealed that inappropriate manipulation of images was more common than many editors had imagined, occurring in about 1 per cent of submissions to the JCB (Rossner, 2006).

Publishers need to understand methods for preventing and detecting misconduct so that they can advise editors on their use, establish appropriate policies and take decisions about what tools to fund.

Despite editors’ and publishers’ best efforts, and despite the most stringent peer-review systems, fraudulent or unreliable work will occasionally get published. Therefore, journals need policies for issuing corrections and retractions; these policies will often be guided by the publisher. Journals not only have responsibilities to their readers (to ensure the quality and integrity of what they publish) but also to their authors. Retracting a published article can affect the reputation and career of the author so it is important that retractions are handled fairly and consistently. Following research suggesting that journals’ retraction practices were not consistent, and showing that retractions were not always clearly indicated to readers and sometimes failed to distinguish the reasons for the retraction (e.g. whether they were due to misconduct or honest error) (Wager and Williams, 2011), COPE (the Committee on Publication Ethics) produced guidelines on this topic to assist editors and publishers (Wager et al., 2009).

Despite publishers’ best efforts, several studies have shown that retracted articles do not disappear but continue to be cited and may therefore continue to mislead (Budd et al., 1998; Steen, 2011). Even when articles are sound they usually undergo various revisions so readers may be confused to find several versions on the Internet, for example on an institutional repository, a pre-print server and a journal website. A new initiative which may help to address these problems is CrossMark (http://www.crossreforg/crossmark/index.html) which will allow publishers to identify the current version of a publication and provide information about its status (e.g. if an article has been corrected or retracted). According to the CrossMark website this is designed to ‘Highlight that the scholarly publisher is responsible for both the initial certification of a publication, as well as the ongoing stewardship of said certified publication’. CrossMark builds on CrossRef’s application of the DOI (Digital Object Identifier) system, which creates permanent links to online content. It was launched in Spring 2012 and is already being used by several major publishers.

When allegations or suspicions of misconduct arise, journals usually need to liaise with authors’ or reviewers’ institutions. It is helpful to have policies for handling such allegations and for responding to institutions and research integrity organisations (such as the Office for Research Integrity or the Office of the Inspector General of the National Science Foundation in the US). Again, because these are relatively rare occurrences, publishers can play an important role in setting policies and advising editors on how to handle cases.

How should editors and publishers respond to allegations or suspicions of fraud and misconduct?

Research and publication misconduct, such as data fabrication and plagiarism, often surface at submission or after publication. Therefore editors and publishers may find themselves having to handle suspicions of misconduct (e.g. reviewers saying that findings are ‘too good to be true’) or allegations of misconduct (e.g. from alert readers or aggrieved researchers). When the concerns relate to published material, there is a clear responsibility to address the issues and, if necessary, correct the record. When they relate to submitted material it may be tempting simply to reject the submission and pass the problem onto somebody else, but the COPE Code of Conduct advises against this (COPE, 2011a).

COPE publishes a series of flowcharts which set out recommended steps for handling most types of misconduct (http://publicationethics.org/ resources/flowcharts). In most cases, they recommend that those accused or suspected of misconduct should be contacted, in neutral terms, asking for an explanation. If this proves unsatisfactory, or no response is received, the case should be investigated further. In some cases, such as plagiarism or image manipulation, the available evidence (e.g. from text-matching software) may be sufficient to determine clearly that misconduct has occurred (but it is still important to let the accused give their account, as first appearances can be misleading). However, in other cases, such as data falsification or disputed authorship, it will be necessary to refer the case to the institution where the research was done. In such cases, publishers and editors should not be expected to act as adjudicators or to conduct their own investigation. However, trying to contact research institutions and instigate an appropriate investigation may be challenging as shown by cases presented to COPE (Wager, 2011).

Publishers can support editors in responding to possible misconduct by offering advice, including legal advice, and administrative support (as getting responses can be time-consuming). While editors should be encouraged to take misconduct seriously, publishers may also need to advise editors not to attempt to investigate or resolve contentious cases by themselves. However well-intentioned they may be, editors (often in a different country or even continent) should not attempt to investigate or arbitrate on most types of serious research misconduct but should focus their energies on ensuring that an appropriate investigation is performed and then acting on the outcome. On the other hand, editors may sometimes need to be protected against bullying and unreasonable behaviour by aggrieved or campaigning parties, who may demand retractions and threaten legal action. In each case, working closely with a knowledgeable and helpful publisher can not only make the editor’s life easier but is also likely to ensure that a good outcome is reached and further problems, such as litigation, are avoided.

What role should publishers play in setting journal policies?

Publishers can play an important role in ensuring their journals have sound policies and that these policies are put into practice. The relative contributions of publishers, editors and journal owners (e.g. academic societies) to developing such policies varies and probably depends on the history of the journal, the relations between the parties and the degree of individual (or corporate) interest in ethical issues. As publishers normally have more experience of running journals than individual editors, they are well placed to offer advice or propose policy templates. Policies should be revised in the wake of problems, and publishers can ensure that experience gained at one journal is shared with others. While detailed instructions to authors may vary for each journal, many publishers set common ethical policies for all their journals and may provide resources for editors, authors and reviewers to explain and elaborate on these. Examples of some publishers’ ethical policies and resources are shown in Table 14.1.

One area in which publishers’ policies should provide guidance is the handling and disclosure of competing interests. Most individuals, and especially academic researchers, like to think they are totally objective and therefore uninfluenced by financial or other factors. However, it is important to realise that editors, reviewers, authors and publishers may be subject to competing interests. WAME (the World Association of Medical Editors) notes that: ‘Conflict of interest (COI) exists when there is a divergence between an individual’s private interests (competing interests) and his or her responsibilities to scientific and publishing activities such that a reasonable observer might wonder if the individual’s behaviour or judgment was motivated by considerations of his or her competing interests.’ (WAME, 2011a). This definition is helpful because it stresses that conflicts relate to what ‘a reasonable observer’ might think rather than whether an editor or author believes their actions have actually been affected. In other words, conflict (like beauty) lies ‘in the eye of the beholder’. The WAME statement goes on to note: ‘Everyone has COIs of some sort. Having a competing interest does not, in itself, imply wrongdoing. However, it constitutes a problem when competing interests could unduly influence one’s responsibilities in the publication process (or be reasonably seen to do so).’

To retain confidence in the peer review process, it is important that conflicts of interest are correctly handled. Publishers should work with editors to decide which conflicts of interest are either so serious or so easily avoidable that they should be prevented and which are less serious, or unavoidable, but must be disclosed. For example, editorial board members should never be involved in decisions concerning their own research papers, nor should they handle papers from close colleagues or family members, so journals need to have systems in place to prevent this. However, editors are usually also researchers, so they are likely to be receiving funding for their research (or to have received this in the past) and this may create conflicts of interest. If editors have strong links to a company or organisation which might be affected by a publication, then they should ensure such papers are handled by another editor without such interests. In 2009, the editor of an orthopaedic journal was heavily criticised in the media for failing to disclose that he had received over $20 million in royalties from a medical device company (Fauber, 2011). Some commentators suggested that he should not be a journal editor given his close ties to one company (Lenzer, 2010). Indeed, the Association for Medical Ethics (2011) recommends that editorial board members should not have ‘substantive financial interests’ in a commercial company (which they define as receiving more than $50 000 in personal compensation).

While having good policies is clearly important, making sure they can be, and are, followed is even more so and this may also depend on publishers. Not only should publishers show leadership in setting high standards for their journals, they will have to decide how much resource to invest in various processes. Actively screening manuscripts for problems such as plagiarised text or inappropriately manipulated images consumes time (usually of paid staff) and may require specialised tools, such as CrossCheck. Therefore, publishers should be involved in decisions about how and when to screen. Training of staff and editors, possibly even of authors and reviewers, may also be needed to ensure that ethical policies are implemented, so publishers should consider funding this. Holding annual meetings for editors from particular disciplines or regions may also provide opportunities for training and raising awareness of ethical issues.

Many publishers pay for their journals to belong to COPE (see www.publicationethics.org and specific documents referenced at the end of the chapter), which provides guidance on all aspects of publication ethics via its public website and advice on specific cases to members.

Another way of increasing compliance with policies is to do an audit. COPE makes audit guidelines available to its members and some publishers have used this for their journals.

While publishers can do much to promote and support initiatives aimed at making their publications ethical, they also need to realise the ethical responsibilities of their own roles. Many of these ethical duties are related to handling conflicts of interest and, in particular, in separating editorial from commercial decision-making as far as possible.

The relationship between publishers and journal editors

The COPE Code of Conduct for Editors notes that ‘The relationship of editors to publishers and owners is often complex but should be based firmly on the principle of editorial independence.’ It also states that ‘Editors should make decisions on which articles to publish based on quality and suitability for the journal and without interference from the journal owner/publisher’ and recommends that ‘Editors should have a written contract(s) setting out their relationship with the journal’s owner and/or publisher’. The COPE Code of Conduct for Publishers (2011b) reiterates this by stating that publishers should ‘foster editorial independence’.

The World Association of Medical Editors (WAME, 2011b) notes that ‘Owners (whether professional associations or for-profit companies) … are ultimately responsible for all aspects of publishing the journal, including its staff, budget, and business policies.’ However, ‘Owners should not interfere in the evaluation, selection or editing of individual articles, either directly or by creating an environment in which editorial decisions are strongly influenced.’ Indirect influences might include bonuses paid to editors linked to journal profitability.

Both COPE and WAME recognise that editorial independence is not absolute; for example, a publisher may legitimately refuse to publish something they consider libellous or likely to bring the publication into disrepute. As Richard Smith (former editor of the BMJ) has written ‘editorial freedom . cannot be total. I couldn’t turn the BMJ into a soccer magazine because I’d got bored with medicine. Freedom must be accompanied by accountability.’ (Smith, 2004).

The WAME Policy Statement notes that ‘The limits of editorial freedom are difficult to define in the general case’ and recognises that ‘Owners have the right to hire and fire editors-in-chief’.

Smith (2004) also commented that ‘everybody supports editorial independence in principle, although it sometimes feels to editors as if the deal is "you can have it so long as you don’t use it".’ He also quoted a similarly cynical interpretation (attributed to journalist Hannen Swaffer) that ‘freedom of the press . means freedom to print such of the proprietor’s prejudices as the advertisers don’t object to’ (Swaffer, 2002).

For commercial publishers, potential conflicts of interest are most likely to arise from commercial aspects. They should therefore ensure that, as far as possible, activities such as selling advertising space or reprints are not allowed to influence editorial decisions. However, when journals are owned (or published) by academic societies, conflicts can also arise if the editor disagrees with a position taken by the society or wishes to publish something that is critical of the society. Several high-profile disputes between the owners and editors of prominent medical journals, which have resulted in editors being fired or resigning, involved this type of dispute. Commenting on one such well-publicised dismissal (of George Lundberg, editor of the Journal of the American Medical Association) two other former editors wrote ‘there is an inherent friction between the society’s journal editor and its executive officer. The mindset and mission of editors are frequently at odds with the understandable wish of the executive to control the society’s affairs and realize as much income as possible for other activities.’ (Fletcher and Fletcher, 1999). In a more recent case (in 2006), two senior editors of the Canadian Medical Association Journal were fired following a dispute over editorial freedom (Spurgeon, 2006).

Separating commercial and editorial decisions is not always straightforward. The WAME Policy Statement (2011b) reminds publishers that they need to avoid ‘creating an environment in which editorial decisions are strongly influenced’. This might include considering how editors are paid or setting up systems to ensure that editors are unaware of any financial implications of publication. For example, if a journal charges authors a publication fee, but waives this for authors who cannot pay, then editorial decisions might be influenced if editors know which authors had requested a waiver. Similarly, reprint transactions (which can be highly lucrative for journals) should be handled separately from editorial discussions – although even if effective ‘firewalls’ are created, it may still be possible for editors to predict which articles are most likely to generate reprint sales (Lundh et al., 2010).

Similarly, while the timing of publication should normally be a purely editorial decision, uninfluenced by the publisher, in some cases journals may agree to publish a study to coincide with a major meeting at which the results will be presented. Such timing is likely to increase media attention, which may be good for the journal, but may also benefit the study’s sponsor, who may also buy reprints to distribute at the meeting. Publishers should endeavour to ensure that the initial editorial decision (i.e. whether or not to publish a particular article) is based solely on the study’s importance and validity, and is not influenced by commercial considerations. However, being responsible for running the business side of the journal, it is reasonable for publishers to allow some coordination between editorial and sales departments after an article has been accepted. Likewise, if a journal offers a rapid publication service for a fee, the editor must be aware of which articles fall into this category (as rapid peer review will have to be organised). Because, in this case, it is impossible for the editor to be ‘blinded’ to the category of submission, it is particularly important that the editor does not have a direct interest (such as a bonus scheme or shareholding) in the journal’s profitability in order to minimise bias in the publication decision.

Why being ethical makes commercial sense

Breaking ethical codes and conventions can have either short- or long-term effects on the commercial viability of a publication. The acute (i.e. short-term) problems, especially if they end up in court, receive most attention but occur less often, while insidious (i.e. long-term) problems are less dramatic, but may still cause the demise of a publication. Publishers need to remember that either scenario, be it a sudden, violent and unexpected blow, or a slow, lingering decline, may be fatal to the journal.

Conclusions and outlook

Publishers need to appreciate the ethical implications of their work. In particular, those producing scholarly (peer-reviewed) publications that report research findings need to consider the responsibilities that this entails. Such responsibility is carried jointly by publishers and editors (and, to some extent also by authors and reviewers). The relationship between publishers and journal editors is, ideally, a close one and should be based on trust and respect, but nevertheless be formalised in a written contract defining the roles and responsibilities of the different parties. Publishers can play an important role in setting journals’ policies and enabling these to be followed. They also have an important role in educating editors and providing guidance based on their more extensive experience.

Whether done by a commercial company or a learned society, publishing is usually undertaken as a business designed to make money or at least cover costs. Following ethical practices represents good business, as ethical breaches can damage confidence in publications and, on occasion, threaten the solvency of a journal or publishing company if disputes have to be resolved in court.

While some ethical issues have been present ever since researchers started to publish their work, developments in technology and changes in societal attitudes mean that new issues and challenges continue to emerge. Advances in detecting or preventing misconduct (such as the availability of software for text matching and image screening) offer opportunities to improve publishing practices. Publishers need to be alert to new developments and ensure their policies and practices are up to date and reflect current realities.

Sources of further information

Perhaps because their publications can have a direct effect on healthcare decisions, and may therefore, directly or indirectly, harm patients, many guidelines on ethics come from the editors and publishers of medical journals. Although some ethical considerations (such as the protection of participants in clinical trials) are not relevant to publishers working in other fields, much advice on general issues such as conflict of interest, authorship and editorial independence applies to all areas of academic publishing. Therefore, do not be put off by guidelines that appear to relate to medical journals even if you are working in another field.

COPE, Committee on Publication Ethics

www.publicationethics.org

An international not-for-profit organisation that advises editors and publishers on all aspects of publication ethics and covers all academic disciplines (although it was originally founded by medical editors). It has published a Code of Conduct for publishers as well as guidance and flowcharts on topics such as retractions and how to handle plagiarism, redundant publication, authorship disputes, reviewer misconduct, etc.

ICMJE, International Committee of Medical Journal Editors

www.icmje.org

An influential group of editors of major medical journals who have produced guidance on many aspects of publishing, notably authorship and dealing with the press. Their guidelines have been widely adopted by many journals, including some outside the medical field.

CSE, Council of Science Editors

www.councilscienceeditors.org

A US-based organisation of science editors (with members around the world). It has published thoughtful guidance, including a useful White Paper on Promoting Integrity in Scientific Journal Publications. (http://www.councilscienceeditors.org/i4a/pages/index.cfm?pageid=3313)

WAME, World Association of Medical Editors

www.wame.org

An international organisation of medical editors. It has published useful policy statements on many aspects of publication ethics.

References