Another kind of ethics: from corrections to retractions

Authors


This issue of the Journal, just over a year into my term as Editor-in-Chief, contains a landmark publication, not one that I or any other Editor or Editor-in-Chief would relish: it’s a Notice of Retraction [1], the first in Anaesthesia’s long history – though not, alas, the first in an anaesthetic journal. But more of that later.

Ask most of our readers what they understand by ‘ethics’ in the context of research and I’d imagine that almost all would talk about ethical approval of studies, informed consent and the general conduct of studies, perhaps expanding into the differences (real or imaginary [2]) between audit and research and the barriers to conducting studies of any type in today’s hard-pressed health service. These issues are important and, some would say, fundamental to the conduct of research, and in that regard, we’ve come a long way since the early voices (key amongst them Henry K. Beecher – an anaesthesiologist, as it happens, [3]) raised concerns about the apparent disregard for such principles despite their enshrinement in the Nuremberg Code of 1947. This Journal, like all other reputable ones, demands adherence to these basic principles as a prerequisite for all material submitted to it [4].

However, there is another kind of ethics apart from ‘research ethics’, although they overlap to a degree, that is more often in the forefront of editors’ minds than researchers’. This is ‘publication ethics’, the principles by which authors (and editors and publishers too, for they are not immune from the potential for bad behaviour) are obliged to act to promote and maintain the integrity of the scientific record. Like research ethics, the standards and principles of publication ethics have evolved largely in response to deviations from what could be considered acceptable behaviour. In the UK, a small group of medical journal editors met in 1997 to discuss how they should react to cases of apparent misconduct on the part of authors, following four cases encountered by the then Editor of Gut [5]. This led to the formation of the Committee on Publication Ethics (COPE), a body that now – through the membership of publishers and publishing groups en masse, as well as individual journals – has over 6200 members and is truly international (and not only medical) in the makeup of both its membership and its Council.

Anaesthesia and its publisher, Blackwell Publishing Ltd, are members of COPE and strive to meet the obligations that such membership demands, including making its processes open and dealing with possible cases of misconduct responsibly, according to COPE’s guidelines and flowcharts (see http://www.publicationethics.org). To this end, a new set of ‘Editorial Policies’ accompanies our guidance to authors, setting out what can be expected by us of authors – and vice versa – regarding such matters (see [4]). These policies complement our general guidance, which, over the last few years, has been expanded to cover many of the issues encompassed by research ethics and some of the ones pertaining to publication ethics.

So what are these areas of misconduct we might encounter, and how do we (i) detect them and (ii) deal with them? As with most things, there’s a range of possible misconduct, with a range of possible outcomes: some result in no more than strained personal relationships; some waste considerable amounts of editors’ and journals’ time, and potentially mislead readers; others indicate that research ethical standards were breached and patients were possibly put at risk; and others result in a distorted scientific record that may lead to harm to future patients. Let’s look at all of these in turn.

One of the most difficult issues to resolve concerns authorship. Author A may feel that he has been unfairly left off the list of contributors, Author B may accuse Author C of taking her idea or data without permission, Author D may have a vendetta against Author B, Author E may not even know that his name has been included on the study, Author F may be the wife of Author A and also an academic competitor of Author B, etc. In such situations, it’s impossible for the editor to know what really happened, so in most cases, the best course of action is to put the disputed manuscript on ice and wait until the authors have sorted it out between them, with or without the involvement of their department(s) and institution(s). We can’t stop such disputes from occurring, but we’d rather the authors all agreed before sending the affected manuscript to us. However, there’s a balance between trying to pick up disputes at the submission stage on the one hand, and making the submission process too unwieldy on the other; investigators will know what a chore it is when a journal requires every author to sign a hard copy of the submission letter, especially for multinational collaborative studies. At Anaesthesia, like many other journals, we require the corresponding author to declare on behalf of all the authors that the criteria for authorship have been met, but we also ask him/her to provide the email addresses of all the named authors so that the initial communication from the Journal advising provisional acceptance goes to everyone on the list. Whilst this will alert authors that their name has been included in a submitted manuscript, it doesn’t help an omitted author, or prevent a miscreant from entering bogus addresses. Within the last year or so, we have had to deal with two cases where authorship dispute has been a major feature.

The next group of misdemeanours concerns the word ‘duplicate’: duplicate submission, when authors submit to more than journal at the same time; duplicate (or redundant) publication, when they write up the same data with a different slant or context, or write up a study as several small papers rather than a single larger one to increase the number of publications (‘salami’ publication); and plagiarism, when authors copy text from elsewhere and present it as original work. (Many authors do not realise that copying their own previously published work into a new manuscript still constitutes plagiarism. Typically, ‘self-plagiarism’ and redundant publication go hand-in-hand). At best, such practices cause unnecessary extra work for editors and reviewers; at worst, the scientific record may be distorted by falsely suggesting that, for example, 400 patients have been described in the literature when the true number may be 200, but reported twice. The crime of plagiarism has a more philosophical basis and a legal one too: the former because readers have the right to expect that what they are reading is original, and the presentation of unoriginal work is therefore misleading them; and the latter because of the risk of infringing copyright law. Within the last year, we have had several instances of duplicate submission/publication/plagiarism to consider; in the majority of cases (but not all), we have concluded that no wrongdoing occurred (e.g. the nature of the study did indeed warrant separate analysis and presentation) or that its occurrence represented error through ignorance rather than a deliberate attempt to misbehave. In our authors’ declaration form, we require confirmation that the presented work has not been published or submitted elsewhere, in whole or in part – a simple enough requirement, you’d have thought, but these cases still occur.

Readers might wonder why we’ve noticed so many cases of duplication in the last year. One possibility is that the incidence is increasing, although we have no way of telling. Another is that we are now scanning manuscripts submitted to Anaesthesia using a specific anti-plagiarism service, CrossCheck (CrossRef, Lynnfield, MA, USA) powered by iThenticate® (iParadigms, Oakland, CA, USA) software. This service compares text strings from the scanned manuscript against a huge database of published work from over 50 publishers (Fig. 1). The Editorial Board of Anaesthesia decided that rather than limit scans to only those manuscripts ‘suspected’ of plagiarism (whatever that may mean), or scan a fixed (random) proportion of manuscripts, it would be better to scan every manuscript, and this has been our practice for the past year, during which time we have rejected around 4% of manuscripts outright (i.e. before review) – in most cases, aiming to educate the authors concerned rather than to punish them, especially when the infringement is relatively minor.

Figure 1.

 CrossCheck scan powered by iThenticate® software, showing that this Editorial has only a 4% overlap with other published work (arrowed). Further screens indicate the source articles and the percentage overlap with each. Scan conducted on 06/10/2010. Image reproduced with permission.

The most serious types of misconduct, and the ones that are dealt with most severely, are those involving gross plagiarism (e.g. duplication of large sections of text from someone else’s work), those in which patients may have been harmed (including a breach of their autonomy) by the conduct of unethical research, and those in which data have been fabricated. Unethical research may be apparent as such at the time of submission and/or review, and Anaesthesia’s reviewers and editors are specifically asked to consider the ethical aspects of the manuscripts they review, in addition to the scientific aspects. Sometimes, however, this only comes to light after acceptance or publication. We have rejected a small but steady minority of manuscripts on the basis of their ethics, including withdrawal after initial acceptance, and one reason for the retraction in this issue is the finding that ethical approval was not obtained as originally stated.

Another aspect of misconduct that can have subtle but potentially severe ramifications is that of competing interests, whereby authors may be biased, either consciously or subconsciously, by their allegiances or aversions to factors (e.g. financial, intellectual or professional) that might influence their conduct of the study and/or its analysis. At worst, an author may wish to hide a major financial or other competing interest that editors, reviewers and readers would otherwise want to know, and putting the results of a study in doubt. The International Committee of Medical Journal Editors (ICMJE) has recently issued a ‘standard’ competing interests form, that it encourages all journals to adopt [6]; Anaesthesia has so far decided against this as it doesn’t fully suit our requirements and we cover competing interests in our authors’ declaration form, but we will keep this under regular review. Our reviewers are likewise asked to declare any competing interests that might influence their review of manuscripts.

Fabrication of data is generally seen as the worst of all crimes [7], as a particular treatment may be given to, or withheld from, countless patients on the basis of a single, influential study or series of studies; if the data supporting the studies’ conclusions are made up, then the potential harm to these patients may be enormous. Furthermore, when the fabrication comes to light, there may be great damage to public confidence in research and medicine. Finally, the consequence for the investigator (and by implication, his/her institution) may also be huge – as Scott Reuben, an anaesthesiologist and prolific researcher, recently sentenced to a six-month prison sentence for healthcare fraud, has discovered [7, 8].

So much for the range of offences. I have mentioned some of the preventive measures taken within this journal to try and reassure ourselves and the readership that what we publish is sound and follows accepted practice. We also aim to follow COPE’s guidelines when it comes to acting on misconduct, or suspected misconduct. The first step is to contact the author(s) seeking an explanation, going on to correct deficiencies in manuscripts, where that’s appropriate, rejecting them, where that’s appropriate, and for the more serious cases of misconduct, informing the authors’ institution(s) and requesting that they conduct an investigation – for it is not within the Journal’s remit or capability to investigate such matters itself. Typically, such requests fall on deaf ears, at least initially, and considerable effort is required on the part of the editors – as demonstrated by accounts of previous cases [9, 10]. For manuscripts that have already been published, readers are notified of simple errors, omissions or lesser offences through the publications of Corrections (‘Errata’ when it’s our fault and ‘Corrigenda’ when it’s the authors’), both in print and online. For more serious matters such as unreliable, unethical or fraudulent studies, or severe cases of plagiarism, a Notice of Retraction (if there is conclusive evidence) or an Expression of Concern (if there isn’t) will be published, which brings me back to the opening of this Editorial.

In the modern age of literature databases and repositories, an unreliable, unethical or fraudulent article cannot simply be removed from existence, and there is a moral argument that attempting to do so (as has been attempted for various writings in recent history) is fundamentally wrong anyway. Retractions are intended to alert readers and users that they should disregard the publication concerned, and each retracted article should be labelled as such in all electronic databases, for all time. As I have said, the retraction in this issue of Anaesthesia is not the first in an anaesthetic journal, although a search of PubMed suggests that relatively few publications in anaesthetic journals have ever been retracted, with two clusters accounting for the vast majority (Table 1). Each Notice of Retraction should provide the reason for retraction, and readers should not assume that just because an article has been retracted, this must necessarily imply intentional wrongdoing on the part of the authors (for example, there may have been an honest error with data collection etc. that became apparent only after publication). As to the Notice in this issue, readers are encouraged to take a look and decide for themselves.

Table 1.   Retracted publications arising from eight anaesthetic journals. Results of a PubMed search (http://www.ncbi.nlm.nih.gov/pubmed) using the search string: ‘Retraction of Publication[Publication Type] AND (“Acta Anaesthesiologica Scandinavica”[Journal] OR “Anaesthesia”[Journal] OR “Anaesthesia and Intensive Care”[Journal] OR “Anesthesia and Analgesia”[Journal] OR “Anesthesiology” [Journal] OR “British Journal of Anaesthesia”[Journal] OR “Canadian Journal of Anaesthesia = Journal Canadien d’Anesthésie”[Journal] OR “European Journal of Anaesthesiology”[Journal])’, conducted on 06/10/2010 (Resulting in 12 ‘hits’, some Notices of Retractions referring to more than one publication). A search using just ‘Retraction of Publication[Publication Type]’ on the same date revealed 1594 ‘hits’, an increase of 10 from the previous week.
JournalNumber of publications retracted Year of retraction(s)*
  1. *The original date of publication may precede the retraction by several years.

  2. †Including seven related publications concerning a series of pharmacokinetic studies, relying on a particular pharmacological assay.

  3. ‡Including 11 publications with Scott Reuben as a co-author.

  4. Note added in proof: a further article has just been retracted by Anesthesia and Analgesia (http://www.aaeditor.org/NoticeofRetraction.pdf (accessed 29/10/2010)).

Acta Anaesthesiologica Scandinavica12009
Anaesthesia00
Anaesthesia and Intensive Care00
Anesthesia and Analgesia2†2010
11‡2009
12008
Anesthesiology5†2010
2‡2009
11999
11998
11993
British Journal of Anaesthesia00
Canadian Journal of Anesthesia00
European Journal of Anaesthesiology1†2010

Competing interests

I am Editor-in-Chief of Anaesthesia and a member of COPE Council and of Wiley-Blackwell’s Ethics Advisory Panel. No external funding is declared.

Ancillary