Why do medical journals keep taking authors at their word?

The recent revelation that a leading official at Memorial Sloan Kettering Cancer Center failed for years to disclose lucrative financial conflicts of interest might have been surprising in its scale. But it’s old news that many researchers aren’t fully transparent when it comes to their financial relationships with industry.

So why should we keep up the charade? And why, given the clarity of the problem, do medical journals continue to take authors at their word — only to wind up looking like dupes?

According to an investigation by ProPublica and The New York Times, Dr. José Baselga — who was chief medical officer of the venerable cancer clinic until resigning Thursday — has what can charitably be described as an inconsistent personal policy on revealing companies that have given him cash or other potentially lucrative fillips. Baselga also has stayed mum about his conflicts of interest — which also involve research funding and seats on advisory boards — in many of his publications, including those in high-rent titles like the New England Journal of Medicine, and despite policies from the journals demanding that authors reveal such relationships.

To be fair, some of the reason we know that Baselga wasn’t fessing up was because some journals — including the NEJM — publish conflict-of-interest disclosure forms that authors submit. That’s not a uniform practice; the Journal of the American Medical Association doesn’t, and in at least one case declined to release the form when pressed by a reporter.

Journals that don’t publish such forms are telling readers that the only judgment of whether a conflict is relevant is theirs. They’re also leaving it unclear whether — when a conflict such as Baselga’s is found — it was the author who didn’t declare it, or the journal that didn’t find it important enough to mention.

So as a first step, journals should publish those forms. Second, just as nearly every publisher now screens submissions for evidence of plagiarism, it’s time the journals do the same for conflicts of interest.

Journals already have some of the tools they need to look for conflicts. A 2010 law, the Physician Payments Sunshine Act, established a government database called Open Payments to disclose industry payments to physicians. Dollars for Docs, from ProPublica, also tracks these ties. However, neither database, as Matthew Herper at Forbes has noted, catches participation on corporate boards or work on drugs or devices that the Food and Drug Administration hasn’t yet approved. Nor do they include patents for relevant discoveries, which a comprehensive repository should house.

The Baselga affair and others illustrate how flawed the system is, but even when journals do learn of undisclosed conflicts of interest, they tend to respond weakly. Although the Committee on Publication Ethics says failure to disclose a major conflict of interest can be grounds for retraction, journals almost always choose to issue corrections instead. At the time of this writing, our database of retractions shows 114 removals of papers for undeclared conflicts of interest, less than 1 percent of the more than 18,000 total retractions, and most of those include other reasons without which the article more than likely wouldn’t have been retracted. That sends the message to authors and readers alike that journals don’t take these issues seriously.

It seems clear, then, that it is time to take this responsibility away from journals, or at least add another layer. What’s needed is a single pool of conflicts data — one that captures cash payments, patents, stock holdings, board seats (both corporate and foundations), funding sources for research — against which editors can screen authors. Given that unscrupulous or indifferent authors are likely to omit potential conflicts, the burden of tracking them on an individual level must be as automated as possible. The names of authors would be linked to their entries in the database, so any reader could click and judge what’s relevant and what isn’t.

We note that this will not be a simple task, and would probably be best tried as a pilot program first, in a small area of research, with some government or foundation funding. The system also needs some sticks. Journals could let authors know that if they find out about relevant conflicts later, they won’t publish work by that author until the database is complete. A version of that approach has worked when it comes to pre-registration of clinical trials, although it took an act of Congress to really cement that incentive — and as STAT has shown, sharing data itself on that platform isn’t quite there yet.

Of course, and this point is critical, a disclosed — or even an undisclosed — potential conflict of interest does not necessarily mean that a study, guideline, or bit of clinical advice is unreliable. Indeed, scientists often are highly sought after by industry because they have a reputation for integrity, fairness, and expertise in their field. Yet evidence shows that physicians who take money from industry are more likely to prescribe brand-name (and therefore more expensive, generally) medications. And, as ProPublica found, the more money, the greater the effect on prescribing habits.

But as the ethicist Art Caplan of New York University tweeted in response to the Baselga news: “We have yet to figure out what [conflict of interest] means or how to manage it in a health care world where industry ties are everywhere.” The trick, to paraphrase Bruce Springsteen, is not to be blind to those ties that bind.