Editor’s Note: Today’s post is by Andrew Grey, Alison Avenell, and Mark Bolland. Andrew is a clinical endocrinologist and Associate Professor of Medicine at the University of Auckland. Alison is a medically trained clinical biochemist and a Professor at the University of Aberdeen, Scotland. Mark is a clinical endocrinologist and Associate Professor of Medicine at the University of Auckland.
Consider these facts:
- A meta-analysis that reportedly screened >20,000 titles is completed and submitted to a prestigious international medical journal in under 4 months. Concerns about its reliabilityare immediately raised. The journal verifies the concerns but takes no public action for more than 1 year.
- A journal is notified that the author of research it published has fabricated clinical trials. Subsequently, detailed concerns about the publication integrity of 10 papers are raised with the journal. The editor declines to become involved, and defers to a publisher staff member, who repeatedly fails to answer requests for updates about the concerns. Almost 6 years later none of the concerns about the 10 affected publications has been resolved or readers alerted to the concerns.
- A publisher says it will only assess the integrity of papers authored by the same fabricator when concerns are raised on a publication-by-publication basis (personal communication, Elsevier publishing staff).
- A journal publishes a letter detailing concerns about internally inconsistent results for the primary outcome from a very influential publication. The authors decline to resolve the concerns, and the journal concludes that nothing more can be done and fails to link the paper to these concerns.
- A national research regulatory body investigates widespread concerns about more than 150 publications: it determines that a university committee, featuring several co-authors of the publications in question, should make the final decision(s) about the outcomes for the affected papers (personal communication, Iran National Committee for Ethics in Biomedical Research). Three years later, to our knowledge, there is no publicly available report or details about any of the investigations.
- An academic institution investigates concerns about the integrity of publications by a staff member: it fails to detect or act upon the absence of ethics committee approval for several studies.
- A World Health Organization-approved clinical trials registry accepts unquestioningly a slew of changes to 119 retrospectively registered clinical trial documents which were made shortly after concerns about the published trial reports are raised, and are designed to ‘resolve’ widespread discrepancies between trial registration information and trial publications. Registry staff do not reply to notification of this. When journals and publishers are notified, they do not act.
- A leading academic in osteoporosis admits to falsifying data in multiple clinical studies, after trying for a long time to cover it up and deflect the blame elsewhere. Only 4 papers are retracted. She is barred from clinical practice but then reinstated 2 years later. Meanwhile, at least 112 of her publications remain in the literature without any indication as to whether they are reliable. A co-author of one thought it had already been retracted because it was unreliable. The relevant Universities did not investigate these papers, and her medical college insisted it was not their responsibility to consider information not part of its disciplinary proceedings.
In each case, readers of the affected publications remain unaware of the problems, and the damaging ripples of the flawed papers distort the literature and clinical practice.
No one questions the critical importance of a reliable biomedical literature. Universities teach research integrity, publishers espouse it fulsomely, government agencies debate and endorse it. There are definitions, recommendations, and guidelines. Yet the cases described above are but a few of the many examples of the slow, opaque, inconsistent, frustrating, and unsatisfactory outcomes of tumbling into the rabbit hole of publication integrity. Watching paint dry is ultimately more fulfilling: at least the paint will be dry eventually. Why is achieving and maintaining publication integrity so fraught? Could it be that the main protagonists don’t actually care?
Publishers have the final say in publication integrity, since they decide what to publish and what to correct. Publications are their ‘product’. Most organizations selling products have mechanisms to ensure the quality of their wares. Assessment of publication integrity should be part of that process at publishing companies. Yet even large enormously profitable publishing houses seem to invest little in the practicalities of resolving integrity concerns. Instead, they may devolve responsibility to journal editors who often have little experience, interest, time, or resources. Publishers rarely collaborate with each other in resolving concerns. Publishers usually do not undertake a comprehensive review of publications by researchers with track records of misbehavior and multiple retractions. When decisions about integrity concerns are reached, publishers rarely report the nature of the concerns, the processes undertaken to assess them, or the responses to them. So, the most important parties – readers of the publications and members of the public whose health is affected by them – remain ignorant of the concerns. All of this suggests publishers don’t care.
Guidance for addressing concerns about publication integrity suggests, very reasonably, that the publisher seek a response from the authors and, depending on the circumstances, an investigation by the authors’ institutions. However, most integrity concerns about publications can be assessed without input from institutions, whose contribution is in any case largely confined to determining researcher behavior, i.e., whether ‘misconduct’ occurred. In correspondence with us, publishers frequently complain that it is difficult to engage with authors and institutions, which suggests that neither of the latter really cares about resolving problems in publications. Although there are notable individual exceptions, many publishing staff seem willing to allow assessments to drift along in a torpid state, repeating the mantra that they are “following the COPE guidelines” while declining to report the status of their assessments. Presumably they hope that eventually a resolution will present itself. Or perhaps that the problems will just disappear. Maybe they think that the research they so profitably publish is of so little importance that it doesn’t matter? It certainly suggests that they don’t care.
The Committee on Publication Ethics (COPE) and its guidelines are often referenced when publication integrity is at stake. COPE seems a strange organization. After beginning as an ad hoc collection of concerned journal editors, COPE has evolved into a charity (definition: an organization set up to provide help and raise money for those in need) whose funding is by subscriptions from publishers and member journals, so it is essentially industry-funded. Membership is restricted to those with direct links to journals and publishers. COPE is therefore strongly journal/publisher-focused. It sets out guidance without involving other important parties (e.g., readers) and focuses on journal processes, not methodology, for integrity assessment. If a journal/publisher claims it followed COPE guidelines, the organization does not appear to question the outcomes. COPE’s guidelines are sufficiently ambiguous that journals and publishers who claim to follow them reach quite different conclusions when considering very similar concerns about publications from the same researchers. COPE makes no recommendations about timeframes for resolution of concerns, and does not sanction members who do not follow its recommendations, perhaps because it lacks either the authority or the will to do so. In our experience, COPE’s responses to requests from journal readers for advice and assistance are erratic, unhelpful, and anonymous, in that they are conducted entirely via an administrative staff member. While it has undoubtedly made some important contributions to publication integrity, COPE shows little sign of evolving into an inclusive organization, with journal readers’ needs foremost, that facilitates timely and rigorous assessments of publication integrity rather than merely the appearance of doing so.
Faced with concerns about an employee’s research, institutions often appear to care first and foremost for their reputations, not the integrity of what is published under their aegis. Commonly, they conduct secretive, prolonged, and incomplete internal investigations, focused on damage control and determining whether ‘misconduct’ occurred, rather than whether all the publications in question are reliable. The reports of their investigations are seldom made public, nor shared with publishers. The key affected parties, readers of the publications, are almost never made aware of the existence of an investigation or its findings. That suggests academic institutions don’t care about publication integrity.
If the leading actors in this grim drama really cared about the reliability of published research rather than their bottom lines — profit and reputation — they would acknowledge the serious flaws in the current approaches to publication integrity, and commit to improvement. But they haven’t. The current processes are so opaque that it is impossible to be reassured that strategies to meaningfully improve the assessment and resolution of publication integrity are even being considered. Academics who work in the field, freely performing the role of quality control officers for publishers, are still treated with suspicion and disdain. The implicit paternalism (“don’t worry, we’ll handle this, but we won’t bother you or our readers with the details”) of publishers and academic institutions is surprising, considering the issues at stake. Perhaps each has forgotten that flawed publications impact adversely on the health and wellbeing of those who fund, participate in, and conduct the research that publishers rely upon for their existence. Or perhaps they don’t care.
Is it harsh to suggest that key organizations do not care about publication integrity? Some may argue that egregious incentives, such as career advancement from quantity and self-defined ‘impact’ of publications, and the competitive business model of academic publishing fuel compromised publication integrity. But many of these incentives are sustained or driven by institutions and publishers, who therefore can influence them.
Improvements are clearly possible. Publishers can review and publish new submissions within a few weeks so should also address integrity concerns in a timely fashion, rather than awaiting the deliberations of lengthy ‘misconduct’ investigations. Publishers could also require more of authors at manuscript submission: agreement to participate in any investigation and/or provide raw data should concerns be raised about the publication, and agreement that failing to participate will lead to withdrawal of the publication. Publishers could also require and audit raw data, support the establishment of independent panels to assess publication integrity, collaborate with academics with relevant expertise, and invest more appropriately in quality control. Institutions can disengage from perverse incentives to publish, and audit research practices among their staff. Both can commit to transparent, non-conflicted, comprehensive, and efficient assessment and resolution of concerns about publication integrity. Such sensible improvements, however, require that each protagonist actually cares. Based on our experience, it is not clear that they do.