The core information supporting science-based medicine resides in the scientific literature. There, scientists and physicians publish the results of experiments and clinical trials that seek to understand the biological mechanisms by which the human body functions and through which disease forms and to apply this understanding to test new treatments for diease. Consequently, the quality and integrity of the biomedical literature are topics of utmost importance to supporters of science-based medicine. We’ve discussed problems with the scientific literature before here, ranging from how pseudoscientific “complementary and alternative medicine” journals have insinuated themselves into the medical literature and how drug companies have managed exercise undue influence over clinical trials and journals.

One question that perhaps we have not dealt with so much is the question of the very nature of a good scientific journal, particularly what is suitable material for such a journal. For purposes of this discussion, I will focus mainly on the biomedical literature, which spans a range from basic science journals dealing with biomedical science to clinical journals, which mainly report the results of clinical trials and clinical research. Of these journals, there are in general two types, journals that primarily report original research and those that present reviews of existing research. Most journals do a mix of the two, the majority tending towards a form where most of the articles are reports of orginal research mixed in with a much smaller number of review articles.

There is one journal, however, that is different. It is a journal known as Medical Hypotheses. It is a journal that (or so it claims) exists to present radical scientific ideas, the more radical the better. Here is how the journal is described on its website:

The purpose of Medical Hypotheses is to publish interesting theoretical papers. The journal will consider radical, speculative and non-mainstream scientific ideas provided they are coherently expressed.

Medical Hypotheses is not, however, a journal for publishing workaday reviews of the literature, nor is it a journal for primary data (except when preliminary data is used to lend support to the main hypothesis presented). Many of the articles submitted do not clearly identify the hypothesis and simply read like reviews.

So far, there’s nothing inherently objectionable or anti-scientific in the concept behind MH. I can easily see a role for a journal that publishes speculative biomedical papers. However, there is a problem, and this problem has been brewing for a long time. Last month, apparently, it came to a head, and last week it was reported in Science that the publisher of MH, Elsevier, had issued the editor, Bruce Charlton, an ultimatum:

The editor of the journal Medical Hypotheses–an oddity in the world of scientific publishing because it does not practice peer review–is about to lose his job over the publication last summer of a paper that says HIV does not cause AIDS. Publishing powerhouse Elsevier today told editor Bruce Charlton that it won’t renew his contract, which expires at the end of 2010, and it asked that Charlton resign immediately or implement a series of changes in his editorial policy, including putting a system of peer review in place. Charlton, who teaches evolutionary psychology at the University of Newcastle upon Tyne in the United Kingdom, says he will do neither, and some on the editorial advisory board say they may resign in protest if he is fired.

Elsevier’s move is the latest in an 8-month battle over the journal; it comes after an anonymous panel convened by Elsevier recommended drastic changes to the journal’s course, and five scientists reviewed the controversial paper and unanimously panned it.

Before I go on, let me just say right here that I don’t necessarily disapprove of a journal devoted to highly speculative, even radical hypotheses. Such a journal can play a very important role for airing ideas at the fringes of what is known. Unfortunately. However, as I shall describe shortly, I believe that MH has been a big problem. Specifially, my problem with MH is that it blurred the line between the speculative and apparently confused “speculative” with “making stuff up.” Apparently the reviewers agreed:

Following the advice of an external panel whose membership has not been made public, Elsevier wrote Charlton on 22 January to say that Medical Hypotheses would have to become a peer-reviewed journal. Potentially controversial papers should receive especially careful scrutiny, the publisher said, and some topics–including “hypotheses that could be interpreted as supporting racism” should be off limits.

Elsevier also had its flagship medical journal, The Lancet, organize a formal review by five anonymous experts. The reviews, which have not yet been released publicly but were obtained by Science, were unanimously harsh–especially about the Duesberg paper, indicating that it is riddled with errors and misinterpretations. “It might entertain their friends and relatives on a cold winter evening, but it does not belong in a scientific journal,” one reviewer wrote. On 24 February, Elsevier wrote Duesberg that his paper–which had not yet been printed and had been taken down from the journal’s Web site in August–would be “permanently withdrawn.” Ruggiero received a similar letter 5 days later.

Why did it come to this? The reason is that a journal whose editor valued “radical” ideas far more than actual science has been polluting the medical literature in a manner that was too easily abused by cranks and quacks. Moreover, this was not a new problem; it had been longstanding and known to those of us who pay attention to such matters. Many are the times I have seen a wide variety of cranks cite MH papers as support for their positions. In essence, MH had become a vanity journal that will publish almost anything, no matter how much it goes against established science. For example, Mark Blaxill published pseudoscientific speculation that vaccines cause autism, and the anti-vaccine movement trumpeted Blaxill’s paper for the next several years as “evidence” in a “peer-reviewed journal” that vaccines cause autism. Similarly, MH published Mark and David Geier’s “hypothesis” proposing the use of the powerful anti-sex hormone drug Lupron to treat autism, an utterly ridiculous idea from a scientific standpoint that wasn’t just “radical”; it was demonstrably wrong. In the same vein, MH published a paper by a dream team of anti-vaccine activists proposing the use of spironolactone to treat autism based on its anti-androgen properties. Recently, MH published an article that appeared to echo attempts by the anti-vaccine movement to link mitochondrial diseases to “vaccine injury.”

It worked, too, because most lay people can’t distinguish between a highly speculative scientific article and a scientific report based on sound data from well-designed experiments and/or clinical trials, with solid scientific reasoning leading to its conclusions. Nor do most people–even scientists– have any idea of some of the other amusingly (and not-so-amusingly) wacky “hypotheses” published in MH, such as ideas that masturbation is a treatment for nasal congestion, a paper linking high heeled shoes to schizophrenia, a meditation on the nature of navel fluff, and truly offensive speculations about “mongoloids.” And there was support for the most blatant pseudoscience as well, even above and beyond support of HIV/AIDS denialism. For example, over the years MH has published several articles arguing for the “plausibility” of homeopathy, such as this, this, and this.

Perhaps the worst debacle suffered by MH came to pass last summer, when it published an article by HIV/AIDS denialist Peter Duesberg that was outrageously wrong and even downright racist. So bad was the article, that Elsevier apparently felt compelled to act and in doing so administered one of the worst indignities imaginable to a scientist. It withdrew Duesberg’s article from MH last year. Now I know what happened since then. Elsevier ordered an external review, the results of which were reported in the news article in Science cited above. Given the history of MH publishing anti-vaccine pseudoscience and other highly dubious papers, I was curious why this particular manuscript would have been withdrawn from Medical Hypotheses, given the usual low standards demonstrated by MH and that the article expresses the usual nonsensical “scientific opinions” that HIV isn’t enough to cause AIDS, that HIV doesn’t kill as many in Africa as estimated, and that anti-HIV drugs don’t work. What changed between the time that Medical Hypotheses (MH) accepted this article and decided to withdraw it? Why would this one article be worse than all the other pseudoscience routinely published by MH? Quite frankly, I was puzzled at the time.

My search led me to Ben Goldacre’s column entitled Medical Hypotheses fails the Aids test, which in turn led me to this detailed explanation on AIDSTruth.org entitled Elsevier retracts Duesberg’s AIDS Denialist article. After that, it all became clear. Basically, the slapdown administered to Duesberg and other HIV/AIDS denialists through his article being retracted by MH had its genesis in a study by Pride Chigwedere and coinvestigators at Harvard University, who estimated that delays in providing antiretroviral drugs in South Africa because of state-supported AIDS denialism (in which Duesberg played a prominent role in promoting) had caused over 300,000 deaths. In fact, the article mentioned Duesberg’s role in promoting HIV/AIDS denialism in South Africa. This is what happened:

AIDS denialist Peter Duesberg, whose influence on the disastrous South African government policies was mentioned in Chigwedere’s article, submitted a response to JAIDS that was co-authored by four others including Rasnick. After this article was rejected because of its poor academic quality, Duesberg et al. submitted it to a different journal, Medical Hypotheses. Two days later, the editor accepted the paper. Medical Hypotheses does not practice peer review, a process in which several scientists check a submitted academic paper for quality and suggest needed improvements over a period of weeks or months. The Duesberg et al. paper was accepted without such a review process, after inspection only by the editor of Medical Hypotheses.

This appears to be a relatively common occurrence, with articles rejected by real journals somehow having a way of finding their way into Medical Hypotheses. Duesberg was particularly crass, too, writing in the article:

A precursor of this paper was rejected by the Journal of AIDS, which published the Chigwedere et al. article, with political and ad hominem arguments but without offering even one reference for an incorrect number or statement of our paper (available on request).

Apparently I missed it at the time, but HIV/AIDS denialists were crowing about this, just as anti-vaccinationists crowed about papers by Mark Blaxill and Mark Geier making it into Medical Hypotheses and held them up as “evidence” that their ideas were making it into mainstream scientific journals.

The publication of Duesberg’s HIV/AIDS denialist paper also led an effort by scientists to do something that should have been done a long time ago. For some reason that has always eluded me, ever since I first discovered the land of woo that is Medical Hypotheses, this journal is indexed with MEDLINE and shows up on PubMed searches. It goes against a lot of what MEDLINE claims to be its standards for indexing a journal or even the very functions of MEDLINE. Based on the co-optation of Medical Hypotheses by HIV/AIDS denialists to their cause, a number of academics, authors, and researchers wrote a letter to Donald A. B. Lindberg, M.D. Director, National Library of Medicine Betsy L. Humphreys Deputy Director, National Library of Medicine Sheldon Kotzin Associate Director, Library Operations Library Selection Technical Review Committee calling for Medical Hypotheses to be delisted from MEDLINE.

My guess is that the serious threat of having MH removed from list of journals indexed by MEDLINE is what led Elsevier to act. After all, if a journal isn’t listed in MEDLINE, its articles won’t show up on PUBMED searches, and that is the kiss of death for any biomedical journal that hopes to maintain any sort of reputation at all. But did Elsevier go too far? Although I view Elsevier’s action as a long overdue effort at finally exercising some quality control, there are those who vigorously defend MH and its editor Bruce Charlton. In fact, Charlton published some of these defenses on his blog, and equally unfortunately some of them were quite poorly reasoned, such as this defense by Professors Lola J. Cuddy and Jacalyn M. Duffin:

If it emerges that Duesberg’s paper erred beyond his minority viewpoint to actual errors–be they deliberate or accidental, a signal comparison can be made to two leading medical journals. Medical Hypotheses would have been no less a victim or a wrongdoer than the distinguished entities The Lancet and the New England Journal of Medicine.

Earlier this month, Elsevier’s flagship journal The Lancet withdrew a 1998 paper by Andrew Wakefield et al. that helped foster the now discredited theory linking autism and MMR vaccines. No one has called for the alteration of Lancet. Indeed, the issue has drawn attention to the preeminent leadership role that The Lancet plays in the dissemination of knowledge and ideas.

Similarly, in 2000 the New England Journal of Medicine published a peer-reviewed paper that strongly supported the use of rofecoxib (known as Vioxx®). Later it emerged that the paper had suffered improper industry interference and failed to declare a treatment-related death. The drug was taken off the market in 2004. Considerable discussion surrounded the editorial responsibilites for the 2000 article when the flaws came to light in 2005. But no one called for the New England Journal of Medicine to be altered in any way. Jeffrey Drazen is still its editor-in-chief.

This is, of course, comparing apples and oranges. It’s been well known that fraud is very difficult to detect through a standard peer review of a scientific paper, and most journals do not have good systems in place to detect undisclosed conflicts of interest. In any case, in their defense of MH Cuddy and Duffin are demonstrating extreme ignorance at best or extreme disingenuousness at worst. There is a profound difference between a journal’s peer reviewers missing examples of scientific fraud, which peer review tends to be ill-equipped to detect, and a journal editor just taking any ridiculous “speculative” paper that comes along and calling it science. Although an argument can be made that Wakefield’s paper should never have been accepted for publication because it was such thing gruel, the Vioxx paper at the time it was published looked like a perfectly legitimate and reasonable randomized clinical trial. It took years to discover the problems with both papers. Cuddy and Duffin are in essence invoking the tu quoque fallacy by pointing to failures of peer review as a justification for the failure of MH not only to maintain a reasonable level of quality control but even to exercise any peer review at all. Remember, MH is not peer reviewed. It is, as Charlton has described it, “editorially reviewed,” which means basically that Charlton picks what is published. Unfortunately, he has chosen poorly so often that he finally got burned.

Another issue illustrated by this controversy is the role of editors, their editorial independence, and how far publishers should go in influencing the content of scientific journals that they publish. Editorial independence is indeed very important for preventing external forces, such as advertisers (i.e., drug companies) from having undue influence over editorial decisions over scientific content. Some argue that editorial independence is, for all intents and purposes, paramount. On the other hand, it is the responsibility of the publisher to exercise quality control, and that’s just what Elsevier’s action appears to be in this case. I’ve complained about Elsevier before in the context of its allowing a pharmaceutical company to pay it to publish what was in essence fake medical journal, but in this case it appears to be doing (mostly) the right thing. I say “mostly” because I do agree that one aspect of its actions does make me uncomfortable, specifically the part about saying that some controversial topics should be off-limits. That drifts uncomfortably close to the realm of pre-emptive censorship. Science depends on the unfettered dissemination and discussion of ideas, even ideas that many might find offensive.

Science is also served by fearlessness in discussing radical ideas. Howver, care has to be taken to make sure that there is a clear line between what is speculation and what is not, and it has to be made very clear when such a journal is not peer-reviewed. MH under Charlton’s leadership has failed on both counts. Whatever the value of the concept behind a journal like MH, in the case of MH that value has not only failed to be realized, but has in fact been degraded and brought into serious disrepute. Charlton has, through his carelessness, arrogance, and his ideologically blind refusal to enforce even the most minimal minimal scientific standards on articles submitted to MH, has resulted in embarrassment after embarrassment falling upon his journal, from its abuse by the anti-vaccine movement to the latest debacle. The Peter Duesberg HIV/AIDS denialist paper retracted by Elsevier was merely the last straw. As a result, Elsevier decided that it had little choice but to order Charlton to impose peer review or to resign. By refusing to adhere to even a modicum of scientific rigor, Charlton has destroyed the aspect of MH that he apparently most values.

A journal devoted to cutting edge, even fringe scientific hypotheses might indeed be valuable, but because of his carelessness, Charlton guaranteed that Medical Hypotheses was not that journal and that it never will be. If there is to be a journal devoted to highly speculative scientific articles, it’s clear that Medical Hypotheses isn’t it and can no longer even attempt to be it. Bruce Charlton saw to that. Maybe MH under new editorial leadership can claim that role, or maybe a new journal will arise to fill such a niche. Either way, the old MH had to change or risk being destroyed. “Radical ideas” are all well and good, and bouncing them around may make for a fun drinking game among scientists. However, if these radical ideas are not supported by known facts, sound scientific studies, and strong reasoning, they are nothing more than that and can serve to mislead more than to enlighten.