Richard Smith: Is the New England Journal of Medicine anti-science?

Richard SmithAbout once a year a furious researcher writes to me complaining that the New England Journal of Medicine won’t publish a letter that strongly criticises, even demolishes, an article the journal has published. They write to me out of frustration, not because I have any influence over the Bostonian paragon, but because I’ve dared to criticise it in print a few times.

What my correspondents can’t understand is why the journal won’t publish their letter when electronic space is infinite and free. Why can’t the journal have rapid responses like the BMJ and many other journals?

I don’t know why the New England Journal of Medicine doesn’t publish electronically all the letters it receives, but I can hypothesise. The Bostonian paragon is unashamedly elitist and committed to excellence and virtue, just like their colleagues in the city teased by Henry James in his novel The Bostonians. Presumably the editors of the journal don’t want to overload their readers with what they see as ill informed criticisms, but want to present them with the quintessence of comment, beautifully edited of course.

But surely this behaviour is anti-science.

Medical journals are either explicitly or implicitly following the theory of science proposed by Karl Popper: scientists develop a falsifiable hypothesis and then test it to destruction. They never arrive at truth, but hypotheses that have survived the destructive fire serve as our best substitute for truth.

It follows that a very important part of science is giving everybody, the hoi polloi as well as the blessed, the chance to scrutinise the hypotheses, methods, data, and conclusions of studies and present their criticisms. We can be much more confident in the findings of a study that has been exposed to tens of thousands of critical eyes than we can in one that has been viewed only by the chosen few, particularly when those few are of the same mental bent as the authors of the study.

Always in this kind of discussion I’m driven back to quoting the blind poet, republican, and regicide John Milton: “Truth was never put to the worse in a free and open encounter…. It is not impossible that she [truth] may have more shapes than one…. If it comes to prohibiting, there is not ought more likely to be prohibited than truth itself, whose first appearance to our eyes bleared and dimmed with prejudice and custom is more unsightly and implausible than many errors….”

The editors of the New England Journal of Medicine must think of themselves as superior people (and they are superior to most of us, and certainly to me) capable of distinguishing truth from error, but could they be making a mistake? I urge them to follow the advice of Rudolf  Virchow, the great German doctor and intellectual, who insisted that “Everybody is free to make a fool of himself in my journal.”

Competing interest: RS admires the New England Journal of Medicine but has several times published criticisms of the journal; and he is passionately committed to open access and sees the journal as an important barrier to complete open access.

Richard Smith was the editor of the BMJ until 2004 and is director of the United Health Group’s chronic disease initiative.


    Dear Richard,
    I have a similar disappointing experience with NEJM. My letter was rejected because of “lack of space”. It referred to the MANTRA-PAF trial which results were published in the NEJM.
    This trial randomised 294 drug-naive patients to either catheter ablation or therapy with anti-arrhythmic drugs. It found that the cumulative Atrial Fibrillation (AF) burden over two years was not significantly different between the two groups. The team concluded in the NEJM: “Given the risk of complications with ablation, our data support the current guidelines recommending anti-arrhythmic drugs as first-line treatment in patients with paroxysmal AF”. Curiously, when these results were first presented in November 2011 at the American Heart Association Scientific Sessions, the team concluded exactly the opposite on the basis of the same data, saying “These data support radiofrequency ablation as a first-line treatment in patients with paroxysmal AF”.

  • AudreySilk

    All good stuff, Dr. Smith. But I especially cling to this from you: “It follows that a very important part of science is giving everybody, the hoi polio as well as the blessed, the chance to scrutinise the hypotheses, methods, data, and conclusions of studies and present their criticisms.”

    Very recently I had a discussion with someone about studies that are “hidden” behind pay walls. In other words, if you don’t shell out cash you can’t see it.

    I cannot understand how it’s (been) acceptable to only be privy to scientific research if you want to PAY for it. It’s beyond comprehension that science isn’t utterly transparent, especially (but not limited to) when its purpose or result is to influence change in society! This has to be some kind of breach of public trust.

    I’d love your thoughts if you see this, Dr. Smith.

  • Justin Coleman

    Thanks Richard

    Your Karl Popper argument is particularly cogent. It has taken my ego a while, but I now feel comfortable reading negative comments about what I write–short of abuse or trash. It expands my knowledge of what opposing views are prominent out there, and just occasionally even makes me change my mind!

    P.S. amused by the ‘hoi polio’…are they those poor commoners unlucky enough to have caught an enterovirus?

  • JDobson

    Thanks for pointing out the typo. I have amended it in the text.
    Juliet, BMJ blogs editor.

  • Ruth Macklin

    As the corresponding author of a letter to NEJM with 44 other signatories, I had to negotiate the terms of the letter with the Editor-in-Chief, Dr. Drazen. Our letter was a response to a letter signed by 46 bioethicists and pediatricians who staunchly defended the SUPPORT trial involving extremely premature infants. The NEJM allows short letters in response to items published in the journal, and longer letters not responding to published letters or articles. In order to honor the rules for a longer letter, we omitted reference to four articles in NEJM in support of SUPPORT. However, an editor inserted footnotes to our article citing the latter articles, thereby violating the journal’s own rules. When I pointed that out–saying that the rules must have changed–the footnotes were removed and inserted instead as an “Editor’s Note.” The letter that our letter criticized was published on line and later in the print journal. I was told that our article would be on line only. Dr. Drazen had earlier written an editorial in strong support of the study we criticized on grounds that the informed consent documents were seriously flawed. Apparently, some letters are more equal than others.

  • Tom Perry

    My cousin and colleagues submitted a letter to NEJM in early 2000, offering observations about pulmonary hypertension detected by echocardiography in their initial patients who took rosiglitazone (Avandia).

    NEJM considered it unworthy because they offered no hypothesis to explain the observation.

    I think it would obviously have been better to accept the letter and work constructively with the authors, given the safety implications.

    Earlier stimulation of discussion, which had been suppressed by GSK (see US Senate Finance Committee report), might have provoked better safety studies of rosiglitazone, or earlier removal from the market.

    Medical history has shown us with ample and continuing examples that much of what we believe to be true is in fact not correct. This is unlikely to change abruptly, or so reliably that we will always appreciate “truth” accurately.

    I think the most influential and widely read journals have the greatest responsibility to the global population to allow relatively unfettered search for truth. Editors could always divide e-responses into those they consider “most interesting”, “most radical”, “most establishment”, or other categories from which free readers might freely select.

    Tom Perry, M.D., FRCPC

    UBC Therapeutics Initiative
    UBC Vancouver

  • tomkindlon

    I agree with you about the value of publishing both e-letters and letters.

    However, lots of journals don’t put up e-letters – this isn’t a NEJM-specific issue. Similarly cogent letters are sometimes turned down by other journals for no good reason.