乌鸦传媒

More often than not, conflict-of-interest disclosures in three major rheumatology journals didn't match records in the U.S. government's , researchers found.

About two-thirds of a random selection of recent papers in these journals included authors who had received relevant industry payments recorded in Open Payments, and in more than 90% of those papers such payments were either underreported or not reported at all, according to Aryeh Abeles, MD, of NYU Langone Orthopedic Hospital in New York City, and colleagues.

It was not just a few authors who underdisclosed. Some 87% of individual authors who had industry relationships failed to report them accurately, Abeles's group . Moreover, all 14 clinical trial reports included in the analysis were marred by faulty disclosures.

"Clearer, journal-provided guidelines should define PCOI [potential conflicts of interest] and reporting expectations both broadly and explicitly, and expectations of adherence must be enforced by the community," the researchers wrote. "Although journals may lack the resources to investigate every author and every PCOI, given the influence of published clinical trials on real-world practice, journals might consider rigorously verifying the accuracy of disclosures for clinical trials ahead of acceptance or publication."

Over the past 15 years or so, upper-tier journals have ramped up their requirements for authors to disclose their relationships with industry, spurred in part by a in 2009 slamming then-current practices that largely left it to authors' discretion whether to report potential conflicts. However, the effectiveness of stricter rules has been questioned: an investigation in 2018 found that many prominent authors in major journals were still neglecting to report industry relationships that paid hundreds of thousands of dollars.

Abeles and colleagues noted that industry may have special interest in influencing rheumatologists because medications they frequently prescribe, such as biologics, are immensely profitable. As such, drugmakers seek to retain them as consultants and speakers, and pay them handsomely to run clinical trials. While the recipients can still maintain their scientific integrity, it's widely agreed that their research products come with disclosures of potential conflicts in the interest of transparency.

To determine the extent to which rheumatology researchers are adhering to current standards, Abeles and colleagues analyzed the first 50 papers published in each of three U.S.-based journals -- Arthritis & Rheumatology, Arthritis Care & Research, and Seminars in Arthritis & Rheumatism -- starting in January 2019. Authors' Open Payment records were examined for payments within the previous 36 months from a company with a marketed product related to the paper's topic, or that was conducting mid- to late-stage clinical trials with a product in that field, as indicated on .

Importantly, payments listed as research grants, or for travel expenses, were not counted as PCOIs because they could have been necessary to support research as opposed to personal remuneration. Also, Abeles and colleagues only looked at papers' first, second, and last authors, as these would likely be the ones most influential on the papers' content.

Additionally, the first two journals listed above had a requirement that authors disclose whether payments exceeded $10,000. For papers there, Abeles's group checked on the accuracy of authors' reports of dollar amounts.

Of the 150 papers in total, 101 included authors receiving payments qualifying as PCOI, and of these, 93 had at least one author whose industry payments were not declared accurately. For most of these (86%) a relevant relationship was not disclosed at all; in 25% (from the two journals with a dollar-amount requirement) the value was underreported. About three-quarters of the papers with inaccurate disclosures had just one author with a faulty report.

Only payments to physicians are recorded in Open Payments, so of the 150 papers' total of 409 authors, 292 were eligible to have their disclosures checked. A total of 135 had payments qualifying as PCOIs; for 118 of them, their disclosures were either nonexistent or undervalued. The aggregate value of completely undisclosed payments was $5.2 million.

Another dismal statistic emerging from the study was that, of 656 individual payments listed in Open Payments, just 29% were reported accurately.

As well, while the 14 clinical trial summaries in the analysis represented 9% of all papers, they accounted for 15% of disclosure inaccuracies and 25% of the undisclosed dollar amounts. And every author on those trial reports who had industry payments failed to declare them properly. (One could take heart, however, from the finding that seven published disclosures did not correspond to any Open Payments record or else overvalued them, and thus could be considered instances of overreporting.)

Limitations to the study included its restriction to U.S.-based authors and an inability to account for the likelihood that authors made innocent mistakes in their disclosures. Abeles and colleagues also observed that journals' instructions to authors had "wording [that] leaves room for interpretation," such that the journals themselves might have considered the resulting disclosures to be accurate even though they were not as defined in the current study.

On the other hand, Abeles and colleagues asserted, their exclusion of several common types of industry payment and their strict definition of topical relevancy "allowed authors greater latitude in what was deemed accurate" for disclosure.

Comment