The Dark Side of Publishing

Abstract

In 2003, we began our journey as Editor-in-Chief and Senior Associate Editor of The American Journal of Pathology (AJP), with a Scientific Editor joining us in 2004. Determined to make a mark on the publishing landscape of Pathology, we introduced several features. Among our personal highlights were implementing a strict triage process, hiring one of us (A.E.C.) as the Scientific Editor, subdividing articles into categories, and bringing publish-ahead-of-print technology to the fore. In the end, however, our final legacy became the implementation of a formal Journal Scientific Integrity Policy. Although the Journal always followed ethical guidelines and principles and expected our authors and reviewers to do the same, the generation and publication of a detailed policy was the natural response to the changing dimensions of today's publishing backdrop. When we entered into our editorships, we didn't anticipate that, beyond our desire to attract and publish the highest quality manuscripts in cellular and molecular mechanisms of disease, an increasing portion of our time (admittedly small, even at the end) and mental unease would be focused on the ethics of our colleagues in the writing, reviewing, and publishing of scholarship. The AJP had long required authors to follow the Uniform Requirements for Manuscripts Submitted to Biomedical Journals prepared by the International Committee of Medical Journal Editors (http://www.icmje.org). However, simply asking for adherence was no longer sufficient. Now to be fair, this scenario wasn't unique to the AJP or to the discipline of Pathology. Ethical questions in research and writing have long affected all disciplines. John Lambert summed it up nicely when he wrote, “It's a strange juncture in the history of moral education when a student can just as conveniently purchase a term paper for his or her Business Ethics class as sit down and write one.” (http://www.highbeam.com/doc/1G1-244246841.html, last accessed July 11, 2012). The added challenge is that the ethics are, in fact, a moving target; one that is highly fluid, often journal-specific, and open to interpretation. In fact the Public Health Service (ie, NIH) and the National Science Foundation's 1990 Final Rule clarified misconduct as “practices that seriously deviate from those commonly accepted”1U.S. Department of Health and Human Services, 1989a. 42 CFR 50, Policies of General Applicability. Notices. Federal Register 54:32449Google Scholar (emphasis added); this phrasing was eventually removed due to its controversial nature. That aside, the question remaining for us became “Could we define for ourselves what we term appropriateness?” To do this required the advice and consent of our stakeholders: the society that owns the journal (the American Society for Investigative Pathology), the Publisher (ASIP during our tenure; currently Elsevier), the editors and editorial board members, the authors, the readers, the payers of the research (often the government), other journals that had grappled with similar issues—the list goes on. We tried to issue best practice guidelines and modified our Instructions to Authors. We purposely tried to steer a middle ground between the highly regimented and dogmatic approach taken by the NIH for handling misconduct to the laissez faire attitude of many of our competitors. The Office of Research Integrity (ORI; http://ori.hhs.gov/documents/rcrintro.pdf, last accessed July 11, 2012) provides a very clear and concise definition of scientific misconduct: “fabrication, falsification, or plagiarism in proposing, performing, or reviewing research, or in reporting research results.” Each type of offense is also defined in more detail: “Fabrication is making up data or results and recording or reporting them; falsification is manipulating research materials, equipment, or processes, or changing or omitting data or results such that the research is not accurately represented in the research record; plagiarism is the appropriation of another person's ideas, processes, results, or words without giving appropriate credit.” These definitions sound easy enough to follow on the surface. The problems reside in the detection of them. CrossRef has a plagiarism tool called CrossCheck, developed by iThenticate, which allows manuscripts to be compared to other papers in its database. The limit to such tools, however, is the source database itself, relying on journals to allow deposit of their materials to be used to search against the submitted manuscript. Imaging fraud, on the other hand, is much more difficult to detect. One notable change that occurred in this millennium revolves around the preparation of scientific manuscripts for submission to journals—the loss of medical illustrations departments. Long gone are the days of submitting your blots, gels, and graphs to experienced graphic designers who would compose your image panels to produce photograph copies of the new figure layout, complete with labels. Further, journals no longer accept hard-copy figures but instead require that all files be submitted digitally. The tools for preparing figures for publication now lie in the hands of those performing the research, whether principle investigator, postdoc, or graduate student. Now, the usefulness and power of such tools is without question, allowing researchers more control over their figure preparation and saving on outsourcing costs. However, without proper training—both technical and ethical—there remains the opportunity for introduction of innocent errors as well as purposeful manipulation. Mike Rossner, previous Managing Editor of the Journal of Cell Biology (JCB), has worked and written extensively on the issue of image manipulation. In 2002, after the JCB switched to electronic-only manuscript submission, Rossner requested that authors and reviewers carefully look at figures in a digital, rather than printed, manner to enable magnification of images to expose any improper manipulation. 2Rossner M. Figure manipulation: assessing what is acceptable.JCB. 2002; 158: 1151Crossref Scopus (12) Google Scholar In a 2004 article,3Rossner M. Yamada K.M. What's in a picture? The temptation of image manipulation.JCB. 2004; 166: 11-15Crossref PubMed Scopus (292) Google Scholar he very clearly outlines the temptations that exist in preparation of images for publication. He provides explicit examples of the kinds of image manipulation that can easily be perpetrated with the use of the tools in Adobe Photoshop. Some of these deceptions are patently obvious, whereas others are less easily detectible with the naked eye. For editors and reviewers, detection of such manipulations can be very difficult. During our tenure, reviewers and readers occasionally alerted us of such image malfeasance. There are tools available now to assist in detecting author malfeasance. One example being droplets, small applications that allows one to ‘drop and drag’ files for examining scientific images in excruciating detail at a level of resolution much greater than the unaided human eye. With use of this and similar techniques, a skilled user can detect manipulation of images and areas of discontinuities of gels, dots, and blots, for example. A whole new lexicon of “solarize edges,” “colorize shapes,” and “visualize backgrounds” came into use. At one point we discussed the purchasing and utilization of such forensic detection tools. We ultimately dropped the idea, as implementing a standard process for assessing all manuscript images at each round of submission and resubmission was not practical at the time. At the end we felt the need to publish a Scientific Integrity Policy.4McDonald J.M. Eisemann M.G. Cox A.E. A matter of trust: introducing the journal scientific integrity policy.Am J Pathol. 2007; 171: 1Abstract Full Text Full Text PDF PubMed Scopus (2) Google Scholar We made use of the myriad resources described above in addition to others not named here but attributed in our final policy (http://www.journals.elsevierhealth.com/periodicals/ajpa/content/integrity; last updated January 2012). We clearly defined such matters as authorship, ghostwriting, and appropriate and inappropriate image preparation. The major change in the submission process was that we asked our authors on submission to verify their adherence to a few principles, among them that the contribution was original and submitted solely to us, that all persons listed as authors had seen and approved its submission, and that the ethical guidelines of the ICMJE and the Journal Scientific Integrity Policy had been adhered to. Further, because scientific publication involves external peer review, the Policy also defined the required conduct of reviewers, editors, and office staff, thus clarifying our expectations of all parties involved. Finally, the policy included a section that detailed procedures for handling allegations of misconduct, a section that has sadly been referred on several occasions. As the saying goes, however, the devil is in the details. In one case a person not affiliated with any university or trained as a scientist had his lawyers contact us to claim that he actually held the copyright over the intellectual property presented in a manuscript for which he was not an author. How could we adjudicate the validity of the claim? Unfortunately many other examples could be cited involving plagiarism, dual submissions, salami slicing (the practice of splitting up what was really one study into multiple smaller, minimal publishable units to inflate the number of papers published), and failure to include all appropriate authors or exclude inappropriate ones (honorific publication). The most egregious, however, is and was that of falsifying data. During our tenure, we were faced with a situation involving fabricated patient data together with falsified supporting figures. After our preliminary assessment, we turned the case over to the author's institution, as it is not a journal's responsibility to make final determinations of scientific misconduct. The result of the institution's formal investigation was that the lead author did, in fact, fabricate data, and the institution requested retraction of the paper. We complied and published a retraction shortly thereafter.5Matsuyama W. Watanabe M. Shirahama Y. Mitsuyama H. Higashimoto I. Osame M. Arimura K. Discoidin domain receptor 1 contributes to the survival of lung fibroblast in idiopathic pulmonary fibrosis.Am J Pathol. 2006; 168: 866-877Abstract Full Text Full Text PDF PubMed Scopus (21) Following this, we were contacted by the institution once again, sadly to inform us that the author had committed suicide. This was a very sobering moment for all of those involved. We do not relay this episode to tell a sensational story but to emphasize the human element of these cases. The simple allegation of misconduct has serious implications for a researcher that are tempered by scientific standing, livelihood, and cultural pressures. It is for this reason that every case must be handled confidentially, with integrity, and as promptly as feasible. So given the threat to one's reputation and career, why would anyone gamble with falsifying or fabricating their data? Ironically, these are the same reasons for committing scientific/publishing fraud. David Goodstein6Goodstein D: Scientific Misconduct. Academe Online 2002, Vol 88, No 1Google Scholar identified three main motives, or risk factors, that in his experience appear to influence the offender. “In all of the cases, the perpetrators i) were under career pressure, ii) knew, or thought they knew, what the result would be if they went to all of the trouble of doing the work properly, and iii) were in a field in which individual experiments are not expected to be precisely reproducible.” We're all aware of the role that tenure practices play in the journal selection game, trying to find the highest impact factor journal that will publish your manuscript. Not to mention the pressure of writing grants to secure funding for one's lab. The downstream consequences of scientific misconduct reach well beyond the personal. Case in point, the long-awaited retraction7Retracted, see Wakefield AJ: MMR vaccination and autism.Lancet. 1999; 354 (The Lancet 2010, 375:44): 949-950PubMed Google Scholar of Wakefield's autism paper, which led to an inordinate number of parents not vaccinating their children for fear of inducing autism (http://seedmagazine.com/content/article/the_cost_of_scientific_misconduct, last accessed July 11, 2012). And the financial costs associated with investigating scientific misconduct have been reported on extensively. Michalek et al8Michalek A.M. Hutson A.D. Wicher C.P. Trump D.L. The Costs and Underappreciated Consequences of Research Misconduct: A Case Study.PLoS Med. 2010; 7: e1000318Crossref PubMed Scopus (54) Google Scholar categorize the financial impact as the fraudulent research itself, the ensuing investigation, and the final remediation. In one case they estimated $525,000 in direct costs, with indirect costs related to time taken by senior administrative faculty, lost grant funding, transfer of postdocs and students to new labs (which had to foot the bill), etc totally over $1 million. And this is just one case; the national costs of such investigations are hard to imagine. The public is closely following these matters and noticing what appears to be an increase in retractions and misconduct cases. Just follow the blog Retraction Watch (http://retractionwatch.wordpress.com), and you'll see daily examples of misconduct allegations, final determinations, and retractions. Of course, one can argue that with the increase digital savvy of our readers and reviewers, we are better able to detect such misconduct (just as the increased knowledge increases the temptation, as discussed above). But there also appears to be a disconnect in scientific ethical culture. In one metaanalysis of fraud surveys, 2% of researchers admitted committing fraud at least once, and 14% acknowledged seeing it in their colleagues.9Fanelli D. How many scientists fabricate and falsify research? A systematic review and meta-analysis of survey data.PLoS One. 2009; 4: e5738Crossref PubMed Scopus (1041) Google Scholar A recent New York Times article also highlights the sharp rise in retractions and scientific misconduct (http://www.nytimes.com/2012/04/17/science/rise-in-scientific-journal-retractions-prompts-calls-for-reform.html, last accessed July 11, 2012). So what are we as editors to do? Our best defense is to educate our authors, editors, and reviewers. And then we must deal with these issues as they arise to the best of our abilities and sensibilities. Having gone through this difficult exercise, for us the world didn't end. We moved on to other assignments and new opportunities,10McDonald J.M. Siegal G.P. In the Blink of an Eye.Am J Path. 2008; 172: 1443-1444Abstract Full Text Full Text PDF Scopus (1) Google Scholar but we would be less than truthful in saying this did not have a profound effect on us, heightening our sensitivity to the importance of our jobs as editors.11Siegal G.P. Who's on first?.Lab Invest. 2009; 89: 4-5Crossref PubMed Scopus (1) Google Scholar, 12Siegal G.P. Navigating interpersonal conflict and peer review.Lab Invest. 2009; 89: 256-258Crossref PubMed Scopus (1) Google Scholar We did not give up on the system (peer review) or the individuals (scientists who participate in the creation of new knowledge and its dissemination), and we maintain our respect for the many thousands of volunteer reviewers and editors who give freely of their time to assure that the nature of science can be uncovered through honest assessment and humankind can thus advance. In the end our goal is to have a “more enduring impact than encomiums inscribed on a plaque” (http://www.tabletmag.com/jewish-life-and-religion/87010/turns-of-praise, last accessed July 11, 2012). (Retracted, see Am J Pathol 2007, 171:1728)

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 47 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.