Ethics and the publication of research

Publication represents the convergence of the activities of researchers, research subjects, research ethics boards, sponsors (industrial, government and others), officials of research institutions, editors, peer reviewers and publishers. For many of these players and their representatives, publication builds careers and leads to personal financial gains. For research subjects and consumers of health care, publication in peer-reviewed journals validates the introduction of new beneficial treatments and new clinical standards. Each of the above groups has its own interests and its own conventional values, as does each individual within each group. Each group and individual may have ethical blind spots, conflicts of interest and temptations, which may lead to violations of ethical principles and practices. Of the examples of misconduct that have marred publication, those that are based on conflicts of interest have attracted much recent attention (1). Here I will focus on violations by researcher-physicians of the privacy of research subjects. The scope of invasions of privacy that can result from publication is illustrated by the three anecdotes below. Between 1965 and 1971 the Hoyt family in New York State lost five children. The findings in the two youngest Hoyt siblings who died in 1970 and 1971 were published in Pediatrics (parental consent for publication is not mentioned) in 1972 together with the findings on three unrelated infants as a case-series of recurrent unexplained apnea (2). The two Hoyt siblings were reported by their parents to have had life-threatening events after discharge from hospital, for which they were readmitted. The deaths of the two infants on the day after their subsequent respective discharges from the hospital were attributed to sudden infant death syndrome (SIDS) (The crux of the paper was the ‘evidence’, provided by the two Hoyt siblings, that apneic episodes may be a precursor of SIDS. It was inferred that infants with apnea who are predisposed to SIDS may be identified early and their deaths prevented. Considered to be ‘trail blazing’ at the time, the publication stimulated the development of home monitoring). In an unrelated case of suspected homicide of three infant siblings, Dr Linda Norton, a pathologist providing expert opinion to the prosecution, suggested that the two siblings in the above paper and their three siblings who predeceased them were also the victims of homicide (3). The assistant district attorney resolved to re-examine the causes of death of the Hoyt siblings. The investigation was initiated in 1992. Based on the data in Pediatrics, it was easy to identify the two deceased infants, and to locate their parents. On April 21, 1995, the mother, Ms Waneta Hoyt, was convicted of second degree murder of the five children. The detection of the homicides and the conviction of Ms Hoyt were made possible by the publication of the above paper. The perpetrator of the homicides was, in this story, also the victim of a violation (by today's standards) of privacy. The author of the paper enjoyed a successful career for a number of years as a physician and from involvement in the industry for producing home monitor for infants considered to be at risk of SIDS (3). According to the district attorney who followed the published clues that led to the conviction of Ms Hoyt, ‘the chicanery’ of the cited paper had protected other killers from justice (3); homicides could masquerade as SIDS. The paper had far-reaching effects on the lives of many parents and infants who had apneic episodes or were related to victims of SIDS. Moreover, by focusing on this etiological theory, the energy and resources of researchers were diverted away from other (what later proved to be) more fertile avenues of research. This parable, drawn to illustrate the violation of privacy by publication, also exemplifies the potential for conflicts of interest of physicians with stakes in industries that make products used in health care. It also illustrates the power of publication to influence clinical practice and the directions of clinical research. The second anecdote presents a not dissimilar violation of privacy, with less consequential effects. A patient consented to the publication of his photographs in a specialty journal. Unknown to both authors and patient, the journal was published electronically in the public domain as well as in the usual hard copy. Incidentally, the journal claimed to have the highest number of ‘hits’ of any medical journal accessible to the public, perhaps motivated by voyeurism. An acquaintance of the patient saw the photographs on the Internet and informed the patient. The patient was very offended by the violation of his privacy. The third anecdote shows that the consequences of violations of privacy may be quite unexpected. A birth-injured child and her parent sued their obstetrician for damages. A medical expert, reviewing the literature on the subject, found a published case series that included the injured child and her radiographs. The mother of the easily identifiable ‘case’ had not been informed of the publication. The published report of the radiographs included the results of additional analyses that refined the original findings, but were not available in the hospital record. In this case, the refined findings were used to support the plaintiffs' case against the obstetrician. The invasion of a patient's privacy by publishing her data could benefit one or the other party in a lawsuit. The three examples above, taken from short case series or case reports, show that it may be easy to identify persons in publications, and that such information may be used in unpredictable ways. The use of published genetic data to deny health or life insurance or employment to individuals and family members is now well recognized (4), but when first revealed, came as a surprise. The effects on a person's livelihood and quality of life may be profound. Violations of privacy involving stigmatizing conditions such as sexually acquired diseases and suspected child abuse may cost persons their social position and self-respect. Violations of privacy are most likely to occur when rare conditions are published. Although individual rare conditions are, by definition, rare, the large number of such conditions and their intrinsic interest increase their likelihood of publication and of the consequent risks. That violations of privacy are unintentional does not help the victim. On the other hand, the publication of small case series and even individual case reports can be important. Published case series may provide generalizable information and valuable insights; even a single case can shake or initiate credible hypotheses. Publications of case series or single cases are based on information collected for clinical purposes, not for research purposes. Provided that clinical subjects are not transformed prospectively into research subjects without their consent (and without research ethics board approval), the ethical issues associated with the publication of cases are privacy and confidentiality. This involves the question of who can have access to health records, as well as what can be published about the research subjects, with or without their informed consent. The Canadian Tri-Council Policy Statement on the “Ethical Conduct for Research Involving Humans” is explicit about obtaining research ethics board approval for the secondary use of data (data contained in records that were collected for purposes other than the research in question), where identifying information is involved (5). As indicated by the anecdotes above, identification of individuals with rare conditions may be easy. It may be safer to rely on the judgment of the research ethics board regarding the identifiability of individuals in a potential publication rather than that of the researcher. If individuals are potentially identifiable, then obtaining their informed consent for publication may be an ethical option, depending on the circumstances. The publication of fabricated, flawed and plagiarized findings (6), redundant publication (7); partial publication (8); the publication of unethical human experiments including those conducted in developing countries (9); failure to obtain the approval of a research ethics committee to conduct the research; failure to obtain the informed consent of subjects; and the unethical handling of authorship (10) are all well recognized examples of research misconduct. The impact of publication misconduct on clinical practice and the subsequent directions of research, as in the example for SIDS research given above, can cost lives. Withholding negative results (1) or the adverse events of research are other well recognized sources of bias. The challenge is to recognize potential ethical traps and to prevent their execution. Societal controls of research with human subjects have intensified with disclosures of unethical behaviour of researchers. Since the 1970s, local committees that review and approve the scientific and ethical aspects of research at its inception (research ethics boards) have been established. Ongoing controls include the review of the documents generated by the study, the establishment of expert safety monitoring committees for individual research projects, and random audit and formal review of the informed consent process (11), in addition to the reporting and review of adverse events, the submission of proposed amendments to the protocol for review and approval, and the submission of periodic reports as part of the annual reapproval process (11). In parallel with these controls, the education of human subject researchers is being enhanced, and researchers are being certified in some jurisdictions. National guidelines and laws have been created. In the future, it is possible that research ethics committee members will need to be certified and the committees accredited. Journal editors, formerly reluctant to engage in ethical ‘censorship’ (12), have escalated their controls and adopted the role of ethical watchdogs. Statements of research ethics board approval and of individual informed consent have been demanded for two or more decades. More recently, authors have been asked to justify their authorship (10). New standards for the declaration of conflicts of interest of authors, peer reviewers and editorial staff of journals are being established (13). Now, major journals are mandating that authors make themselves accountable for publications on which their names appear (1). Novel ethical situations come up fairly frequently. A recent example is the question that has been raised about the ethics of publishing comparisons of mortality rates of a condition without naming the participating institutions (published performance reports on quality of health care are often called ‘report cards’ [14]) (15). To compare mortality rates objectively, the rates are adjusted statistically for differences in the severity of illness before patient admission. It was suggested that papers withholding the names of the institutions could justifiably be censored by the editor (15). This ethical stance is based on the assumption that such studies constitute ‘high quality report cards’. If report cards are indeed high in quality, then most would agree that the institutions should be named, although it is questionable that the activities that produce such cards can be categorized as ‘research’. On the other hand, if comparisons were unreliable, then it surely would be unethical to publish the names of the institutions. It is tragic that the unethical behaviour of a few researchers has had adverse effects on the human subjects of medical research, on the research endeavour, on patient-physician relationships and on patient outcomes. It seems inevitable that more constraints and obligations will be placed on researchers in the future. Canadian researchers would do well to become familiarized with the Tri-Council Policy Statement, and keep abreast of changes that are resulting from the global overheating of the human research environment.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.