Richard Smith on editors’ conflicts of interest

Richard SmithWe are all more interested in the conflicts of interests of others than we are in our own, and editors are no exception. Having preached to authors and reviewers on conflict of interest, editors have largely neglected their own, but an important and fascinating paper in PloS Medicine shows how editors can be exposed to dramatic conflicts of interest. 

The paper is suitably po faced, as is the accompanying editorial, but, as a blogger and ex-editor, I can spell out one of the conflicts of editors in stark terms. It arises when considering a large clinical trial funded by a drug company, and, for example, a third of the trials in the New England Journal of Medicine are funded by industry with almost another half having mixed funding that includes a drug company. Editors know well that they may be able to sell a million dollars worth of reprints of such an article, with a profit margin of perhaps 70%. In other words publishing that one paper will lead to $700 000 on the bottom line. Very few actions in business provide such a substantial profit from so little.

Deciding whether to publish such a paper provides a stark conflict of interest because editors have to think a lot about money. As income from advertising falls, as it is for many journals just as with newspapers, some journals slip into the red. The minute that a journal loses rather than makes money the owners begin to wonder if they need it. Then even if the journal makes a great deal of money, as do several of the bigger journals, the owners would always like it to make more—or at least maintain the profits because they may well be dependent on them.

So how else might editors put $700 000 onto the bottom line? They have to sell about 10 000 extra subscriptions—because subscriptions have a much lower profit margin. To sell so many subscriptions is virtually impossible—and remember that this is to replace the profit from publishing one drug company sponsored trial. As the paper in PloS Medicine shows, the New England Journal of Medicine in 2005-6 published 66 trials supported solely by industry and another 95 with some industry funding.

Another option would be to cut costs, which probably means staff.  The editor would have to fire at least five staff (and probably many more) to add $700 000 to the journal’s profits, and clearly such a strategy could not be followed more than a couple of times.

It’s thus very tempting to publish that drug company sponsored trial, and the temptation is increased further by such trials boosting impact factors, as the PloS Medicine paper shows. Such trials are well cited partly because they are important and partly because drug companies have considerable resources to promote the papers, not least by distributing hundreds of thousands of reprints. The PloS Medicine authors calculate that the impact factor of the New England Journal of Medicine would be reduced by about 15% if it declined to publish drug company sponsored trials.

And editors care very much about impact factors. A high score means not only prestige but also more important papers, more subscriptions, and so more money. Publishing a drug company sponsored trial rather than, say, a study of changing the built environment to increase physical activity will bring both profit and an increased impact factor. How tempting.

The PloS Medicine authors show, as have others, that the proportion of trials funded solely by industry ranges from 7% in the BMJ through 26% for JAMA to 32% for the New England Journal of Medicine. The authors asked the owners of the six journals they studied to give them data on sources of income, but only the BMJ and the Lancet agreed. In 2005-6 the BMJ made 16% of its income from display advertising and 3% from the sale of reprints, while the Lancet made 1% from display advertising and 41% from reprints. The Lancet sold 11,5 million reprints, and the BMJ 968 000.

The authors tried to work out the income of the American journals from their tax returns, but it wasn’t easy because their owners publish more than one journal. They asked the owners to confirm their calculations but were unable to elicit any response from the editors of JAMA and the New England Journal of Medicine.

The conclusion of the authors is that journals should disclose financial conflicts of interest in the way that they require authors and reviewers to do. They will, I suspect, have a long wait.
 I must confess that nothing surprised me in the paper, but I do think that it’s am important step to begin to provide some data, albeit tentative and incomplete, on the financial conflicts of interest of editors and journals.

But what did surprise me was the quote from Marcia Angell with which Harvey Marcovitch ended his editorial in PloS Medicine. Angell was an editor on the New England Journal of Medicine for some 20 years, and I knew her a little. She is brilliant and feisty, has been voted one of the most influential women in the US, and fought battles with the Massachusetts Medical Society, the owners of the  New England Journal of Medicine. She wrote a very readable and influential book on medicine and the drug industry, and she writes excellent pieces in the New York Review of Books.

I’ve read several of her articles, but I hadn’t read the one that Marcovitch quoted in which she wrote: “it is simply no longer possible to believe much of the clinical research that is published, or to rely on the judgment of trusted physicians or authoritative medical guidelines. I take no pleasure in this conclusion, which I reached slowly and reluctantly over my two decades as an editor of the New England Journal of Medicine.”

Sadly I followed the same path and spelt out my disillusionment in my “j’accuse” book The Trouble with Medical Journals. I wrote it in 2004, and since then my pessimism has deepened.

Competing interest: RS was editor of the BMJ and chief executive of the BMJ Publishing Group from 1991 to 2004. He is also a member of the board of the Public Library of Science.

  • Bill

    “how else might editors put $700 000 onto the bottom line? They have to sell about 10 000 extra subscriptions” — so a subscription represents $70/year in profit? That seems rather low, given that they can cost the subscriber tens of thousands.

  • Richard Smith

    Actually, Bill, it's probably too high.

    There are two sorts of subscriptions–institutional and personal. The profit on an institutional subscription is higher than $70, but the profit on a personal paper subscription is much lower, more like $10. Journals have zero chance of selling 10 000 more institutional subscriptions and very little chance of selling 10 000 personal subscriptions.

    So the basic point–that reprints provide easy profits and subscriptions impossibly hard profits–is very true.

  • Karen Weintraub

    So, what's the answer? Stay healthy? Get the government to pay for more scientific research? Alternative therapies for which there's no data at all? I'm not trying to be snarky, I'm really trying to understand how to cope with this info. Is the medical publishing world like democracy – an imperfect system, but the best we've got? Or is the system so biased as to be useless, or worse?

  • Wouter

    One solution could be to go all digital/paperless to reduce costs and have the universities/hospitals publish the journals as they already pay either directly or indirectly for most of the subscriptions anyway. Preferably all journals would be free/open access as well. Probably won't happen soon though…

  • Erick Turner

    A new paper related to this topic is reported on here:

    Money is one commodity, and a closely related one is the splashy positive finding–the kind that make headlines and gets the attention of other researchers, which gets lots of citations for the journal and helps out their impact factor. And a higher impact factor means more 'eyeballs' for the journal and probably the ability to command higher prices for reprints.

    The selective publication of positive findings is now quite well-established, but it's been hard to nail down who the culprits are. This new study shows that journal reviewers are part of the problem. This study didn't discuss journal editors, but I see no reason why they should be any different.

    For the past few years, the ICMJE (International Committee of Journal Editors) has announced on its website that its member journals have an “Obligation to Publish Negative Studies” ( My sense is that, with the exception of a very few journals, this mandate is being ignored. It is not in the interest of journals to publish negative studies; otherwise, they wouldn't get quoted as much in the media, and their impact factors might be sullied.

    The article will be appearing in the Annals of Internal Medicine soon, but for some reason, it doesn't show up among their Early View articles.

    (In case you're wondering, I am not an author of this study or an agent, but I did hear an early version of it presented at a meeting last year.)

  • Erick Turner

    A correction to my earlier post: the journal name is Archives (not Annals) of Internal Medicine. Here's the full reference:
    Emerson et al. Testing for the presence of positive-outcome bias in peer review: a randomized controlled trial. Arch Intern Med (2010) vol. 170 (21) pp. 1934-9