4. Research and publication
35. Publication ethics: obligations of authors, peer-reviewers, and editors
Gail A. Van Norman and Stephen Jackson
An anesthesia pain management researcher submits abstracts for two studies. It is discovered that he did not obtain prior approval for human subjects research. In addition, a colleague finds he is listed as a co-author on one of the studies, although he did not participate in the research. Investigation reveals that the anesthesiologist had fabricated data in some or all of 21 published studies dating back over 13 years. This uncovered what arguably is one of the most wide-ranging research fraud cases in history. His reports had been favorable to several drugs under investigation as “replacements” for postoperative narcotic therapy, and also had refuted animal studies suggesting that one of these medications interfered with bone healing after orthopedic surgery.
Not only are all of this researcher’s findings now in question, but also those of numerous studies by other investigators that relied upon these findings for subsequent research design and comparisons. A medical journal editor remarked that “these retractions clearly raise the possibility that we might be heading in the wrong directions or toward blind ends in attempts to improve pain therapy.”1 A prominent academic anesthesiologist stated that the fraudulent research had affected “millions of patients worldwide,” and possibly led to the sale of “billions of dollars worth of potentially dangerous drugs.”2
Ethical principles of beneficence and nonmaleficence require that physicians strive to improve medical knowledge to improve patients’ lives (beneficence) and avoid harmful or ineffective treatments (nonmaleficence). Medical research seeks to find truths that support these principles, and medical literature should be a map illuminating that search. The integrity of research involves more than simply the just and honest conduction of research, but also honest and fair reporting and analysis of results, peer review, and publication of research findings. Investigators and authors have primacy over the ethical conduct and reporting of research, but the publication process also involves ethical obligations on the part of reviewers, editors, and publishers. Fraudulent research and publication practices divert the search for truth, and corrupt the medical literature.
Publication also serves other critical processes in promoting the integrity and efficacy of the medical profession, and thereby in promoting patient well-being. Scientific enterprise is built on a foundation of trust. “If science is to flourish and attain its appropriate role in aiding human progress, it is incumbent upon…the scientific community to help provide a research environment that, through its adherence to high ethical standards…will attract and retain individuals of outstanding intellect and character.”3 Synthesis, debate, and discussion are all integral to the intellectual process required for analyzing and validating the relevancy of scientific findings in the context of medical care. Medical publications educate physicians through reviews and synopses of complex research findings, distilled for application in clinical practice.
Publication is a critical part of academic medicine, and sets scholarly work apart from the practice of medicine. Authorship is integral to obtaining credit for one’s research, creative thoughts, and educational efforts, and it is essential for achieving a successful academic career. Publications influence promotions and can affect recruitment opportunities and future research funding. As the gatekeepers to publication, reviewers and editors wield power over whether academic credit is justly apportioned, and thereby can affect academic careers. They also determine which research is “relevant” enough to see the light of publication, and may thereby indirectly determine which patient groups benefit from research efforts.
Investigation of scientific fraud
Scientific misconduct is internationally recognized as a serious problem, but few countries have regulatory systems in place to evaluate scientific fraud. Widely publicized cases of scientific misconduct led the US Congress to establish the Office of Scientific Integrity (later named the Office of Research Integrity, or ORI) in 1989. Denmark, Norway, Finland, and Sweden all established formal review councils in the early 1990s to investigate scientific fraud. In the UK, medical fraud is reported to the General Medical Council (GMC). A voluntary organization, the Committee on Publication Ethics (COPE), which includes publishers and editors of over 300 journals in Europe and Asia, reviews instances of misconduct. However, it functions as an advisory body and cannot apply sanctions. Other countries calling for institutional or other reviews and sanctions include Canada, Japan, India, Croatia, Germany, and China, although none has regulatory agencies.4
In practice, the reporting, investigation, and (potential) sanctioning of fraudulent research and publication practices falls largely to professional peers, individual institutions, and the actions of journal reviewers and editors.
Ethical obligations of authorship and author misconduct
The International Committee of Medical Journal Editors (ICMJE) defines an author as one who has made substantial contributions in all of the following areas: concept, design and acquisition of data or analysis and interpretation of data, drafting or critical revision of the publication, and final approval of the version that is published.5
Medical authors have ethical obligations of veracity, or to be truthful and nonmanipulative, in the reporting of research results. In addition, authors have obligations to be just in allocating credit for what often is a collaborative effort among colleagues. Author misconduct has the potential to cause detrimental treatment of patients, as well as to mislead other researchers in the construction and implementation of their subsequent research designs. It also may harm academic colleagues by failing to properly credit their work, or by inappropriately crediting them with work they did not do.
Types of author misconduct in research publication include, but are not limited to: (1) fabrication (the invention of false results); (2) falsification (the manipulation or omission of critical data so as to inaccurately represent the research record or results); (3) plagiarism (appropriation of another’s ideas, processes, results or words without appropriate attribution); (4) ghost authorship (the acceptance of credit for publications written by another or the assignment of credit to a non-author); and (5) redundant publication (the submission of the same study in more than one publication, or the deliberate splitting of results from one study into several publications).
Fabrication and falsification of data
The harm caused by falsification and fabrication of data is self-evident. False information may cause practitioners to expose patients to ineffective, or worse, harmful treatments. Other researchers may be diverted from productive paths of inquiry in the pursuit of fictional results. One study of researchers suggests that nearly 2% admit to at least one career episode of fabricating, falsifying or modifying data. However, when asked about the behavior of colleagues, about 14% admitted to knowing of falsifications. Medical and pharmacological researchers more frequently report misconduct than other researchers.6 Another study confirmed these discouraging results, finding that 4.7% of authors surveyed stated they had participated in research that involved fabrication or misrepresentation. In only about half of cases had the misconduct ever been discovered. Over 17% of the authors indicated that they knew of instances of falsification, fabrication or misleading reporting involving other academics. In a large proportion of these cases, the conduct also remains undiscovered, suggesting that the authors themselves had not reported the misconduct.7
Plagiarism: theft of intellectual property, or the sincerest form of flattery?
A basic definition of plagiarism is the appropriation of someone else’s words and/or ideas as one’s own. It is nearly always wrong,a because it damages the true author (by denying credit and disrespecting their efforts), it harms readers (by deceiving them, and by making it harder for them to trace the true route by which an idea was developed) and it accumulates benefits to the undeserving (the plagiarist). Plagiarism raises questions of trust in scholarly work: if an author copies some material, how do we know they didn’t copy something else?
A review conducted at Harvard found that complaints over medical authorship, including the plundering of noncredited work of junior academicians, had more than quadrupled over a period of one decade. A larger total proportion of complaints came from females, although they represented a minority of the academic population. There was also a trend toward more complaints from non-US citizens,8suggesting that plagiarism may more often “victimize” academically vulnerable populations. A notorious example of scientific plagiarism is the surreptitious acquisition by Watson and Crick of unpublished data, “photograph 51,” from Rosalind Franklin’s research on DNA structure as well as of a confidential research progress report about her work. Both elements were essential to their development of the double helix model.9,10 Crick later acknowledged that their work was based on Franklin’s data in 1961, sadly 3 years after her death.11
Plagiarism violates ethical principles of nonmaleficence and justice. While it is easy to condemn plagiarism, it nonetheless is technically challenging to identify precisely what constitutes plagiarism. It also is questionable whether all forms of plagiarism are equally culpable.
Words versus ideas
The verbatim copying of an entire thesis of another is easy to identify as plagiarism, because it steals both the words and ideas that comprise the creative work of the original author. But words and ideas do not in all instances have similar weight in determining originality. For example, a great love poem is not great because of an original concept (love), but is exceptional for its expression of the concept. Paraphrasing such a poem would not constitute plagiarism because it is the original wording that makes it unique. Bouville points out that science and scholarship, on the other hand, are about newknowledge and ideas. “An experimental result that is described using different words is not a different result.”12 Therefore, in medicine and science, the wording might be less important than the ideas that the words convey in determining if a scientific thesis has been plagiarized.
Is all plagiarism equally wrong?
Is self-plagiarism (when an author makes duplicate statements in two different publications) ethically as concerning as plagiarizing the work of another? Self-plagiarizing original research results is the same as redundant or duplicate publication and it is harmful in ways that are discussed below. But what about scholarly reviews, synthesis, and opinion publications? Ideas can only be expressed in so many ways. While some have labeled the practice “intentional fraud,”13 Chalmers points out that an author may deliberately use similar styles and wording in different venues to express particularly important concepts, simply because repetition also represents emphasis, and thereby underscores an idea’s importance.14 The self-plagiarist is not stealing original ideas or words from anyone else. Publishers have reason to expect that their copyrights will be honored, but concerns about the integrity of the scholarly work itself appear to be less serious in such cases.
Redundant publication includes the practice of publishing the same results in different journals, publishing a review of those results nearly simultaneously in a different journal, or splitting a study into two or more parts to publish in separate journals. While the last practice may be acceptable when a study involves large populations studied over many years, overlapping or split manuscripts are usually not justified.
Redundant publication may be undertaken to distort academic accomplishments in a “publish or perish” system. Almost 1 in 20 scientists admit to publishing the same data in two or more sources in order to enhance their curriculum vitae.15 Some authors claim that they merely were trying to reach different audiences. However, duplicate publication has detrimental effects that may be visited on patients, and on colleagues trying to sort through a mass of information for meaningful results. Redundancy adds inconsequential material to the medical literature, may wrongly emphasize the importance of findings, and burdens an already over-laden publication review system.
Ghost authorship and honorary authorship
“Ghost writing” and “honorary” authorship both involve the improper attribution of authorship or credit to someone who did not actually participate in the research or the review and synthesis of ideas on which a publication is based. “Ghost writing” is the practice of failing to name an individual who substantially contributed to a publication. The term often refers to the practice of attaching a researcher’s name to a paper that was written by a professional writer who is not named. In “honorary authorship,” the name of (often) a senior academic is included among the true authors although that academician did not have a principle role in the publication. Both of these practices are common. In one study, approximately 29% of articles had honorary authors, ghost authors, or both.16 In the case of industry-related trials, one study demonstrated a prevalence of ghost authorship approaching 91%.17
Improperly assigning or concealing authorship violates ethical guidelines for authorship, and is harmful to the medical publication process in a number of ways. It may falsely elevate the perceived significance and reliability of a study if well-respected researchers are inappropriately credited with the findings. Ghost writing can hide conflicts of interest that may affect the balance and veracity of the material presented – as, for example, when the writer is a pharmaceutical company employee discussing an important new drug that the company is developing. Knowing the identities of the true author(s) of a study is important with respect to both accountability, and also to the ability to retrospectively examine results in light of new data, such as newly recognized adverse outcomes. Basic trial data can be difficult to review after the fact, particularly if the authors of the record did not write the original report and no longer have access to original trial data.18 An author who knowingly agrees to allow his or her name to be attached to work in which he/she did not participate is engaging in an act of fraud.
Ethical obligations of peer reviewers
Peer review is encountered throughout medicine. It is used to judge clinical performance and quality outcome measures, as well as to determine the allocation of research grants and the quality of research design. It is also a key process in medical publications. Peer review satisfies the practical need to determine if research and reviews are well-designed, ethically executed, and present significant new information that will lead to better patient care. Peer review also serves professional interests in maintaining autonomy within the specialty to evaluate professional performance. Peer reviewers have the power not only to impact career advancement among authors, but also to directly affect which studies and opinions find their way into the medical literature and thereafter influence the practice of medicine.
Although it seems self-evident that peer reviewers require appropriate expertise in the field being reviewed, incompetent review is one of the most common complaints among researchers. More than 60% of researchers polled in one study indicated that they had experienced incompetent reviews, defined as the reviewer being unfamiliar with the subject matter, not carefully reading the article, or making mistakes of fact or reasoning in the review.19
Confidentiality and trust
Reviewers are in a position of trust, receiving advance notice of potentially important discoveries prior to their publication. Breach of confidentiality exposes the author to the risks of plagiarism, and also to the risks that commercially sensitive information could be inappropriately released, either to the public or to investor interests. Examples of plagiarism and theft by peer reviewers, while not common, are not very difficult to find.20 One-tenth of researchers who were surveyed indicated they believed that a reviewer had deliberately delayed approval of one of their manuscripts so that the reviewer could publish an article on the same topic.19
Respect and protection of dignity of colleagues
No author relishes a poor review or rejection, but surprisingly, abusive reviews, that is ones in which the author, and not the idea, comes under attack, are not uncommon.19 One article reports abusive reviewer comments that included an unsupported accusation that a new primary investigator had submitted a senior researcher’s ideas as their own, a statement that revisions to a paper should “start by burning the entire manuscript,” and actual name calling.21 Clearly, peer reviewers have obligations not only to confine their critiques to the manuscript, but also to respect the dignity of their colleagues, lest the peer review process become mired in personal attacks and retributive behavior.
Ethical obligations of journal editors
Editors have considerable power: their decisions about what gets published – whether research results, reviews or editorial opinion – significantly impact perceptions and practice in medicine and medical research. They control what criticisms may be launched against authors, and how authors may respond when criticized. As the ultimate gatekeepers on publication, they hold considerable sway over academic careers. Despite such influence, the ethical responsibilities of editors only recently have been considered in depth.
In parallel to authors and peer reviewers, editors have ethical obligations to assure as far as possible that published material is accurate and not fraudulent, that research adheres to ethical guidelines with regard to protection of human and animal subjects, that confidentiality of manuscript submissions is maintained,
Table 35.1. Responsibilities of journal editors*
* summarized from the Committee on Publication Ethics
and that any potential personal conflicts of interest are avoided or at least made transparent.
COPE has outlined a code of conduct for journal editors, listing essential responsibilities (see Table 35.1). While much of this effort concentrated on the administrative duties and transparency of medical editorship, the code included important responsibilities to assure that allegations of misconduct are properly investigated.
A critically important, and difficult responsibility of medical editors is one of “cleansing” the medical literature when instances of publishing misconduct are proven. This is especially important if the reliability of data is in question, since erroneous or fraudulent results has ramifications for both patient care and future research. Although the COPE published code of conduct for editors charges them with retraction of fraudulent or unethical research (see Table 35.2), barriers to such retractions may include protracted investigation processes, author disagreement, threats of litigation, and misunderstandings by editors themselves of appropriate actions to take. COPE emphasizes that the purpose of retraction should always be correction of the literature, and not punishment of the author,
Table 35.2. Guidelines for retracting articles*
* summarized from the COPE guidelines on retracting articles
because problems in publishing can occur due to honest mistakes as well as willful misconduct. Investigation and punishment are best referred back to the author’s institution for further action.
The introductory case is an example of fabrication of data. Moreover, the anesthesiologist improperly involved an unknowing colleague in the fraud by attaching his name to the research. The concerns about the anesthesiologist’s research were raised at his home institution, which then notified journals in which his publications appeared. All of the researcher’s articles were retracted. However, “scrubbing” the literature of all citations of this research as well as managing derivative publications from related research is an ongoing task that may never be complete. The effects on patients are unknown. Because the research results encouraged the use of medications that may have had detrimental effects on bone healing, both physical and financial harms to patients on a large scale remain possible.
• Publication serves physicians’ ethical duties to improve medical knowledge, provide continuing benefits to patients, and avoid harmful or ineffective treatment.
• Publication is an integral part of academic medicine – crediting research, creative thoughts, and educational efforts.
• Few regulatory bodies exist globally to investigate scientific fraud.
• Authors have ethical obligations to be truthful regarding credit for the work and outcomes of research.
• Fabrication and falsification of data, plagiarism, misleading assignment of authorship and redundant publications all are detrimental to the mission of medical publication.
• Peer reviewers have obligations to be competent, fair, and balanced, and free of conflicts of interest in reviewing medical manuscripts. Breaches of confidentiality and abusive review tactics are not consistent with ethical review.
• Journal editors have parallel responsibilities to assure accuracy in the medical literature. They have additional responsibilities to try to assure that appropriate investigation occurs if fraud is suspected, and to retract suspect material when it is discovered.
Notesa Authors’ comment: plagiarism that is encouraged and condoned by the original author may not always be wrong.
1 Winstein, K.J. and Armstrong, D. Top pain scientist fabricated data in studies, hospital says. The Wall Street Journal. March 11, 2009. http://online.wsj.com/article/SB123672510903888207.html.
2 Borrell, B. A medical Madoff: anesthesiologist faked data in 21 studies. Scientific American, March 10, 2009. http://www.scientificamerican.com/article.cfm?id=a-medical-madoff-anesthesiologist-faked-data.
3 Alberts, B. (2010). Promoting scientific standards. (Editorial). Science, 327, 12–3.
4 Council of Science Editors. White paper on promoing integrity in scientific journal publications. Approved by the CSE Board of Directors, March 29, 2009.http://www.councilscienceeditors.org/editorial_policies/whitepaper/3–2_international.cfm.
5 Uniform requirements for manuscripts submitted to biomedical journals; ethical considerations in the conduct and reporting of research: authorship and contributorship. International Committee of Medical Journal Editors. ICMJE.org. http://www.icmje.org./ethical_1author.html.
6 Fanelli, D. (2009). How many scientists fabricate and falsify research? A systematic review and metaanalysis of survey data. PLoS One, 4(5), e5738.
7 Gardner, W., Lidz, C.W., and Hartwig, K.C. (2005). Authors’ reports about research integrity problems in clinical trials. Contemp Clin Trials, 26(2), 244–51.
8 Wilcox, L. (1998). Authorship. The coin of the realm, the source of complaints. JAMA, 280(3), 216–17.
9 Fuller, W. (2003). Who said ‘helix’? Nature, 424, 876–8.
10 Watson, J.D. (1968). The Double Helix: A Personal Account of the Discovery of the Structure of DNA. New York: Athenaeum.
11 Zallen, D.T. (2003). Despite Franklin’s work, Wilkins earned his Nobel. Nature, 425(6951), 15.
12* Bouville, M. (2008). Plagiarism: words and ideas. Sci Eng Ethics, 14, 311–22.
13 Editorial: Self-plagiarism: unintentional, harmless, or fraud? (2009). Lancet, 374(9691), 664.
14 Chalmer, I. (2009). Intentional self-plagiarism. Lancet, 374(9699), 1422.
15* Khanyile, T.D., Duma, S., Fakude, L.P., et al. (2006). Research integrity and misconduct: a clarification of the concepts. Curationis, 29(1), 40–5.
16 Flanagin, A., Carey, L.A., Fontanarosa, P.B., et al. (1998). Prevalence of articles with honorary authors and ghost authors in peer-reviewed medical journals. JAMA, 280(3), 222–4.
17 Gotzsche, P.C., Hrobjartsson, A., Johansen, H.K. et al. (2007). Ghost authorship in industry-initiated randomized trials. PLoS Med, 4, e19.
18 Johansen, H.K. and Gotzsche, P.C. (1999). Problems in the design and reporting of trials of antifungal agents encountered during meta-analysis. JAMA, 282, 1752–9.
19 Resnik, D.B., Gutierrez-Ford, C., and Peddada, S. (2008). Perceptions of ethical problems with scientific journal peer review: an exploratory study. Sci Eng Ethics, 14, 305–10.
20 Dalton, R. (2001). Peers under pressure. Nature, 413, 102–4.
21 Hadjistavropoulos, T. and Bieling, P.J. (2000). When reviewers attack: ethics, free speech, and the peer review process. Can Psychol, 41(3), 152–9.
Committee on Publication Ethics (COPE) http://publicationethics.org/ International Committee of Medical Journal Editors http://www.icmje.org.
Ross, J.S., Hill, K.P., Egilman, D.S., and Drumholz, H.M. (2008). Guest authorship and ghost writing in publications related to rofecoxib. JAMA, 299(15), 1800–12.
Sox, H. and Rennie, D. (2006). Research misconduct: retraction and cleansing the medical literature: lessons learned from the Poehlman case. Ann Int Med, 144(8), 609–13.
Wager, E., Barbour, V., Yentis, S., and Kleinert, S., on behalf of COPE Council. Retractions: guidance from the Committee on Publication Ethics. http://publicationethics.org/guidelines (accessed Dec 1,2009).
World Association of Medical Editors. Dalla Lana School of Public Health, University of Toronto, Ontario, Canada http://www.wame.org/.