Commentary: Moving The Journal Forward

Conflicts of Interest

Medical research, and the publications in which that research is presented, continue to be confronted with allegations, if not findings, of research misconduct. Available data indicate that the number of research papers that have been retracted for fraud have increased in the past decade.⁹ Beyond the occurrence of actual fraud such as data fabrication are other forms of misconduct, such as plagiarism, scientific error, and the bias introduced by conflicts of interest. Medical publication through objective peer review procedures has been regarded as one mechanism by which to ensure the integrity of the underlying scientific work.

Biomedical journals vary considerably in their rules and procedures related to conflict of interest and disclosures of such.¹⁰ Conflict of interest considerations apply to authors, editors, peer reviewers, research sponsors, and their respective institutions. Such disclosures were optional to authors in the past, but are now typically mandatory, although their comprehensiveness varies. Disclosures of conflicts of interest, of course, do not negate bias or legitimate the published data and interpretation.¹¹ Some periodicals provide information about required conflict of interest disclosures on their web sites, in print, in the instructions to authors section, or on a required authorship disclosure form submitted with the manuscript. Some journals provide a definition of conflict of interest, while others even provide examples of what constitutes a required conflict of interest disclosure (e.g., direct financial ownership through equities or stock, consultancy, personal relationships, paid expert testimony, and travel grants). Journal disclosure policies vary regarding the threshold of financial involvement to be reported, and the relevant time interval for that involvement.¹² Some journals publish disclosures of financial relationships and other potential conflicts of interest of the editors themselves.¹³ Many journals annually publish names of peer reviewers for the previous year, without conflict of interest disclosures, although the editor may retain those unpublished data. Research in this area has demonstrated that some journals state that they adhere to the conflict of interest provisions of the International Committee of Medical Journal Editors (ICMJE),¹⁴ but do not actually do so.¹⁰

The ICMJE requires that participating journals call for the disclosure of four types of information and provide specific definitions.¹² Beginning November 1, 2010, any “relevant financial activities outside the submitted work, during the 36 months prior to submission,” must be reported on that form, which is completed online and subsequently revised when appropriate. An additional required disclosure includes any “other relationships or activities that readers perceive to have influenced, or that give the appearance of potentially influencing, what you wrote in the submitted work,” for the 36 months before submission.¹⁵

There is less consensus in biomedicine and its publications about identifying and disclosing nonfinancial conflicts of interest. The Institute of Medicine (IOM), in its definition of conflict of interest, identifies these nonfinancial interests as desire for professional advancement, recognition for personal achievement, and favors to colleagues, friends, and family.^16,17 The IOM apparently has overlooked such possible conflicts as political or other advocacy interests. Regarding financial relationships, the IOM suggested that the medical profession develop a consensus regarding standard content, format, and procedures for the disclosure of financial relationships with industry, the categories of relationships that should be disclosed, and the specific information that should be provided.^16,17

Conflict of interest considerations are pervasive in medical practice, research, and subsequent publication. Ethics guidelines to manage these considerations are a moving target and will continue to evolve. The ICMJE adopted conflicts of interest provisions regarding the conduct and reporting of research, and The Journal has similar requirements.

Financial conflicts of interest are less prevalent in forensic psychiatry, given the absence of commercial research support in the field, but they are not absent. The Journal authors often are paid testimonial experts in civil litigation on behalf of corporate interests. If an expert psychiatric witness, repeatedly retained by the tobacco industry in litigation, subsequently attempts to publish an article in The Journal about the voluntariness of tobacco use, tobacco addiction, or related topics, then that expert would be expected to disclose that the underlying work has been funded by the industry.

More broadly troubling for forensic psychiatry publishing is the disclosure of nonfinancial conflicts of interest, whether intellectual, religious, political, academic, or personal. Might the reader, for instance, be interested to know that the author of an article on pre-trial criminal forensic mental health evaluations was exclusively retained by and testified for the prosecution? Might the reader be interested to know that the author of an article on legal disposition of sexual offenders had been personally sexually assaulted? Might the reader want to know that the author of an article on capital punishment has openly advocated for the abolition of the death penalty?

Arguably, any personal or professional experience that could bias an author could be said to present a conflict of interest. Of course, there are no bright lines that define what might be considered potential or actual biases in this regard (i.e., family member victimization). This is not merely a hypothetical concern; readers may be familiar with some of the literature regarding the sexual victimization of psychotherapists.

As we know from our work as forensic evaluators, no one is free of bias, and no published article is completely free of bias.¹⁸ Asking authors to state whether they have a conflict of interest, which is one type of bias, is especially problematic, since people are known to be poor judges of their own biases. Cognitive research has shown that people are prone to rationalize their beliefs and actions and thereby tend not to believe that conflicts of interest affect them.¹⁹ These concerns place journal editors, and ultimately the reader, in a challenging situation.

The Journal has not formally adopted the Uniform Requirements for Manuscripts Submitted to Biomedical Journals¹⁴ published by the ICMJE, but could easily do so. Rather, it incorporates some of that group's provisions into its own. The Journal's website contains no information about conflict of interest definitions or procedures. The time frame for conflicts of interest is not explained. There is no source of information about what might constitute a disclosable conflict of interest, and no illustrative examples or discussion of this is provided. The process of resolving an unreported conflict of interest is not published, as other journals have done.²⁰ The editors' own conflicts of interest disclosures are not provided online or in print.

My perusal of authors' disclosures of financial or other potential conflicts of interest in The Journal located remarkably few such statements in any of the articles, a practice that began in 2010 with Volume 38. In that volume, there were three articles with disclosures in the first issue from four authors, two articles with disclosures in the second issue from four authors, and no disclosures in the remaining two issues. Most of the disclosures related to a committee membership or office in a professional organization. Thus, there were disclosures in just 5 of 88 articles in the entire volume, excluding the Legal Digest. The reader should consider the possible explanations for the absence of disclosures to this magnitude. Perhaps such potential author influences and biases do not exist in our field, or perhaps they are not recognized or publicly acknowledged. Disclosures of conflicts of interest in other medical journals or even in the development of clinical practice guidelines are sometimes omitted by authors and deans.^21,22

The Journal's editors, with little effort or controversy, could readily rectify these conflict of interest omissions, both online and in print. Beyond those changes, an additional disclosure should include the professional and personal sources of income for each co-author, regardless of whether the co-author believes that there is a potential conflict of interest with the substance of the manuscript. This disclosure is analogous to the presentation of the credentials of the expert witness in which the expert's professional experience and credentials are outlined for the trier of fact to assess the expert's credibility and objectivity. Let the reader be the judge of whether the author's interests and activities constitute potential conflicts. Recall that voting members of the American Psychiatric Association and American Psychological Association are provided with the income sources of its candidates for office and their professional activities before each election. If there are a large number of co-authors in a given article, then space limitations may dictate that the information be made available only online. Contemplating this, the ICMJE disclosure form can be completed by the author online and then stored there for future review by readers or real-time revision by the author.¹²

We know that we cannot eliminate bias in our evaluations, testimony, publications, or our lives, but can only strive to limit and minimize bias. Perhaps, alternatively, in lieu of objectivity, we can make these biases transparent to others. Should the disclosure default rules err on the side of over-reporting or under-reporting? Yet, biomedicine and its publications must determine how far to go and how much time, energy, and resources are to be devoted to this effort. One survey of several types of health care professionals, for instance, concluded that 88 percent of respondents believed that commercial support of continuing education programs introduces bias, but only 42 percent were willing to pay increased registration fees to eliminate or reduce that commercial support.²³ At a minimum, biomedical journals should pursue more consistency in disclosure standards, so that authors and readers can become accustomed to a broader disclosure practice.^17,24

Biomedical Authorship

Kapoor and colleagues² raise the complex area of biomedical authorship and indicate that the editors are unaware of authorship disputes in manuscript submissions. The Journal continues to apply the three-pronged ICMJE standard for scientific authorship and requires that all co-authors execute in writing their adherence to this standard. The editors have not routinely asked co-authors to substantiate their compliance with these criteria and presumably have not had reason to do so in a given situation. Here, too, editors rely on self-reported information from contributors without corroboration. Biomedical journal editors avoid becoming involved in authorship disputes among co-authors, but ghostwriting and guest authorship have regularly occurred in the recent past, to the detriment of a journal and its editors.²⁵

The Journal has retained its practice of an authorship model rather the contributorship model advocated elsewhere and practiced by JAMA and other high-impact publications.^24,26,27 In the contributorship model, each co-author identifies his or her specific contribution to the research and publication to promote accuracy, transparency, and accountability. One author, for instance, could indicate that he wrote the first draft of the paper and then added the comments of the co-authors and peer reviewers. In this model, one of the authors is identified as a guarantor who takes overall responsibility for the endeavor. Academic promotions committees, research funding sources, and readers may have legitimate interest in the specific contribution of each co-author. Kapoor et al. unconvincingly explain their reluctance to adopt the contributorship model by its “lack of wide acceptance in academia and concerns about its applicability outside of laboratory research” (Ref. 2, p 337). Neither reservation seems cogent. Imagine attending any professional entertainment performance (e.g., film, theater, or dance) with a long list of hundreds of names but no specific reference to their individual roles. As forensic psychiatrists, we are especially knowledgeable and sensitive to responsibility and accountability and not only can process such information, but arguably are entitled to more author-specific data than a simple byline.

The Journal's Instructions for Authors and the Authorship Responsibility form could be enhanced by a more detailed description of authorship rules and expectations. Other medical journals, for instance, provide a definition of plagiarism with examples of what submissions are, or are not, acceptable.²⁸ Procedures for resolving authorship disputes and other matters could be established in advance and published as recommended by the Committee on Publication Ethics (COPE).^24,29 This information could be provided both in print and online.

Authorship concerns in forensic mental health publishing could be the subject of empirical study. Authorship decision-making has been the subject of empirical investigation in other fields.³⁰ We do not know the prevalence and significance of gift authorship, ghost authorship, authorship disputes over sequencing, actual conflicts of interest, method of establishing authorship, satisfaction of authors with actual decision-making practices, and other matters in the forensic mental health field. These concerns may or may not differ between publishing in forensic mental health and other fields.

Ethics of Forensic Publishing

In their paper, Kapoor and colleagues² bravely initiated an inquiry regarding the underlying ethics principles that guide publishing in forensic mental health. They wonder whether forensic publishing should be guided by a physician-patient ethic, a forensic evaluator-evaluee ethic (whatever that may be), the “virtuous physician” model proposed by Pellegrino (Ref. 31, pp 14–15) or some variation or combination of these.

Recall that the treating physician-patient relationship is primarily dyadic, as is the forensic evaluator-evaluee. Our collective struggle to identify an ethics model for the forensic evaluator-evaluee relationship derives in part from the triadic, or at least the nondyadic, nature of forensic work, with the justice system and the larger society introduced as other parties. Similarly, in the publication realm, interested parties include editors, peer reviewers, publishers, research funding sources, academic promotions committees, the general reading public, and the mental health field generally. Given these many voices and audiences, we are unlikely to identify a simple model or ethic of medical publishing. Some may analogize biomedical publishing to journalism, but the practice ethics in journalism differs from that in medical research.³²

Recall, too, that the American Medical Association Principles of Medical Ethics, with the American Psychiatric Association's Annotation,³³ in Section 5, command us to pursue scientific research, continuing medical education, a lifetime of learning, and publication of information to patients and the public. How else do we demonstrate our professionalism than by research, teaching, and publication? The vignettes discussed in Kapoor et al. demonstrate the three concepts of professionalism denoted by Belitz: interpersonal professionalism, intrapersonal professionalism, and public professionalism (Ref. 34, p 77). The increasing attention to medical professionalism as behavioral rather than attitudinal may also assist us to identify critical ethics concepts in publishing.³⁵

In conclusion, we are grateful that Kapoor et al. have opened the door to the workings of The Journal and have begun a dialogue about the larger ethics-related concerns. Publication practices and ethics guidelines are a moving, continuously evolving target. No doubt the editors will be confronted with many challenges ahead, especially in the realm of new medical diagnostic and assessment technology, Internet communication, and the appearance of open-access journals.^36,–,40 The forensic applications of technological advancements, such as new and more powerful forensic assessment instruments and neurobiological and genetic assessment tools, will introduce new conflicts of interest and other concerns for the editors, of which we cannot even conceive at this point.