ѻý

Journal Conflict of Interest Disclosure Policies Are All Over the Map

<ѻý class="mpt-content-deck">— Editors rely on the honor system; is that good enough to disclose authors' possible conflicts?
MedpageToday
 A photo of a stethoscope draped over a stack of magazines.

The nation's leading medical journals have widely varying policies regarding the conflicts of interest they require authors to disclose, a ѻý survey found.

Not only are the time periods for which conflicts should be disclosed different; none of the responding journals said they routinely conduct independent checks, such as reviewing the database, for financial relationships that authors didn't disclose.

Eight of the 16 journals surveyed set a 3-year window in which potential conflicts should be reported, three set a 2-year time period, and two -- Science and the American Journal of Public Health -- impose a 5-year reporting period. Potential conflicts that existed outside those windows would not be required for disclosure.

A disclosure of a potential conflict does not mean an article is biased or inaccurate, and many authors with industry or advocacy organization support are often the most knowledgeable experts in their field. But readers, especially clinicians and patients using these publications to make treatment decisions, argue they need to know a study's financial backing in order to weigh its merit.

Some journals require disclosure of potential conflicts that existed up to the publication date, while others require it as of the article's acceptance or from the conception of the work. Still others said conflicts should be disclosed as of the date the author fills out the disclosure form.

Nature's portfolio of journals does not specify a look-back period for disclosures, but spokesman Michael Stacey wrote, "We expect authors to be transparent about any financial and non-financial conflicts of interest that could directly, or be perceived to, undermine the objectivity, integrity, and value of the publication in question."

The Journal of the American College of Cardiology likewise doesn't specify a time frame but asks for "all current relevant disclosures," said executive managing editor Eileen Cavanagh.

Some journals said disclosure requirements can vary by the type of submission. For example, The Lancet has separate policies for comments, seminars, reviews, series, and therapeutics papers.

Several journal representatives responded that they rely on the honor system and depend on authors to disclose relevant conflicts. Some elaborated that their staffs don't have time to check, and only look further if an issue is brought to their attention by a reviewer or, after publication, a reader.

Said Nature's Stacey, "Although we will facilitate disclosure during the peer review and publication process, we do not expect to police this policy ourselves: the responsibility for appropriately disclosing, managing, and eliminating competing interests rests with the authors and their institutions."

The BMJ's "approach to additional checks varies on a case-by-case basis," Caroline White, BMJ Group's senior media relations executive, told ѻý in an email. "The subject, journal, content, and preferences of the editor will all influence any checks which are done."

Sabine Kleinert, MD, MRCP, deputy editor for The Lancet, said her publication "would only check/ask authors again if a reviewer pointed out undeclared [conflicts of interest] or someone else, i.e., a reader, after publication."

"Unfortunately, we don't have staffing to conduct thorough random checks," wrote American Journal of Public Health strategic communications director Joseph Bremner. "Given the number of authors on papers, this would likely require a full-time position to run background checks." He added that the journal relies "upon the honor system," and goes further "if something is brought to our attention."

Those explanations did not sit well with Barbara Redman, PhD, MBE, of New York University School of Medicine's Grossman division of medical ethics.

"All they have to do is go into the Open Payments database. I mean, how hard is that?" she asked rhetorically.

Unreported conflicts are a "terrible problem" as are the narrow windows of time required by most journals, she continued. "There's plenty of research that shows that what's in the Open Payments database is not reported to journals at all," she told ѻý.

"It means that there's potential bias -- emphasizing potential bias -- in the [submitted] article, but the reader and researchers and patients who follow along don't know there might be," she said.

ѻý is aware of at least one recently published article related to a class of drugs whose author, according to the Open Payments database, received a significant amount of money from a prominent maker of a drug in that class. The article said the author reported no conflicts, apparently because the relationship ended a few years earlier than the period that journal required the author to disclose it.

Redman added that more important than the span of time in which author conflicts are supposed to be disclosed is the value or amount of that relationship, how long it went on, how recently it was active, and whether the author or family members have any secondary interest, such as investments in company stocks.

Two medical publications reported that they depend on the peer review process to discover undisclosed conflicts. That too riled Redman.

"Peer reviewers don't get paid anything. They should expect that a journal -- not necessarily the editor but someone on the staff -- has actually checked the Open Payments database to see if the disclosure is correct, or if there's no disclosure, whether there should be," she said.

Peter Lurie, MD, MPH, president and executive director of the Center for Science in the Public Interest, had a similar view. For most if not all journals, he said, "It's all on the honor system. And at a minimum it seems to me that any journal should be running Open Payments against all of its authors. It does not seem to me that's a lot of work."

Most of the journals said they use the International Committee of Medical Journal Editors (ICMJE) and ask submitting authors to fill out its form. The form asks authors to "disclose all relationships/activities/interests listed below that are related to the content of your manuscript."

Two publications, the Journal of Clinical Endocrinology and Metabolism and Clinical Infectious Diseases, said they use the process guidelines from the (COPE) in cases of undisclosed potential conflicts.

Nancy Chescheir, MD, COPE's vice chair and a retired ob/gyn from North Carolina, said the majority of COPE's work involves setting guidelines for how journals should be run, such as what their editorial boards should look like, and how to deal with copyright issues.

But COPE also provides guidance for how journals should deal with conflicts, "when the whole system goes awry for one reason or another," she said.

Bernard Lo, MD, is director emeritus of the Program in Medical Ethics at the University of California San Francisco and chaired the on conflicts of interest. He told ѻý he thinks medical journals should require authors to divulge potential conflicts over much longer periods of time than most now do.

For clinical trials, those disclosures should date back to the original discussions about study design, "certainly longer than 2 years," he said.

And he thinks those requirements should extend to authors of commentaries and viewpoints, especially since many doctors in busy clinical practices "are overwhelmed; they don't have time to read the journals critically." They may instead rely on media articles that include quotes from study commentaries when recommending treatments for their patients.

Should all of the important medical journals agree to enforce a much tighter disclosure policy? Perhaps a uniform one? Lo said that would be great, "but the problem is they all have little fiefdoms. No one tells the New England Journal what to do. No one tells The Lancet what to do," he said.

Lo, who with deputy JAMA editor Deborah Grady, MD, MPH, authored a 2017 JAMA on how payments to physicians can influence decisions, also scoffed at excuses some journals gave that they don't have time to go online to check authors' funding sources.

The journals "could ask a college student, or use AI. You could program a computer to do it," he said.

In a 2018 in JAMA, Jeffrey Botkin, MD, MPH, former associate vice president for research integrity at the University of Utah who oversaw the school's conflict of interest office, suggested that researchers who knowingly or negligently falsify significant conflict of interest disclosures in publications or grant applications should risk a finding of misconduct.

In an email to ѻý, Botkin elaborated. He said conflict of interest rules and requirements should be guided by data, but said there has been little research done on the topic.

"Policies tend to be guided by opinions, often committee opinions, about what seems like a reasonable number," he wrote. "Does a $50,000 interest create substantially more bias than $20,000? Which creates more bias, a $20,000 financial interest in the sponsor [covered by the rules] or prospects of publication in a premiere journal [not covered by the rules]? I don't think we know."

For this survey, ѻý sought responses from 16 leading medical journals: The American Journal of Psychiatry, American Journal of Public Health, Annals of Internal Medicine, The BMJ, Clinical Infectious Diseases, Health Affairs, JAMA, Journal of Clinical Endocrinology and Metabolism, Journal of Clinical Oncology, Journal of the American College of Cardiology, The Lancet, Nature, New England Journal of Medicine, Neurology, Pediatrics, and Science. All responded except the Pediatrics, which declined to participate.

  • author['full_name']

    Cheryl Clark has been a medical & science journalist for more than three decades.