“We fully retract this paper from the published record”

Ouch.

Here’s the story (which Kaiser forwarded to me). The English medical journal The Lancet (according to its publisher, “the world’s leading independent general medical journal”) published an article in 1998 in support of the much-derided fringe theory that MMR vaccination causes autism. From the BBC report:

The Lancet said it now accepted claims made by the researchers were “false”.

It comes after Dr Andrew Wakefield, the lead researcher in the 1998 paper, was ruled last week to have broken research rules by the General Medical Council. . . . Dr Wakefield was in the pay of solicitors who were acting for parents who believed their children had been harmed by MMR. . . .

[The Lancet is now] accepting the research was fundamentally flawed because of a lack of ethical approval and the way the children’s illnesses were presented.

The statement added: “We fully retract this paper from the published record.” Last week, the GMC ruled that Dr Wakefield had shown a “callous disregard” for children and acted “dishonestly” while he carried out his research. It will decide later whether to strike him off the medical register.

The regulator only looked at how he acted during the research, not whether the findings were right or wrong – although they have been widely discredited by medical experts across the world in the years since publication.

They also write:

The publication caused vaccination rates to plummet, resulting in a rise in measles.

An interesting question, no? What’s the causal effect of a single published article?

P.S. I love it how they refer to the vaccine as a “three-in-one jab.” So English! They would never call it a “jab” in America. So much more evocative than “shot,” in my opinion.

22 thoughts on ““We fully retract this paper from the published record”

  1. The article didn't say much (anything?) about the relationship between MMR and autism, that all came about in a press conference that was done before (I think) the article came out – but then people think that's what the article was about, because that's what they were told in the press conference.

    Vaccination rates might have plummeted anyway – there are plenty of people who rant about them, and they don't seem to rely on evidence of any sort, so the publication of an article only helps a little.

    Another article that I can recall that had a large effect was the Kellogg's diet article, from (I'd guess) about 10 years ago, I can't find it now but at the time they said that the number of downloads was in the moderate 6 figures. I don't know if sales of Kellogg's cereals increased.

    Oh, I've just thought of another and I'm too lazy to edit what I've already written. The 1995 pill scare paper (also published in The Lancet), which said that there was an increase in risk of thrombosis in women taking the contraceptive pill. I've used this in my teaching in the past, 'cos it's a nice example of relative versus absolute risk differences (i.e. double the risk is an increase of 0.001% – or something). There are several papers I've seen that look at the repercussions – here's one: http://www.bmj.com/cgi/reprint/326/7383/254.pdf .

    One interesting feature that potentially increases the possibility of journals having an impact in the UK is that The Lancet and the BMJ are published on a Friday, so news stories have the weekend to get disseminated through the media before medical professionals (either govt or people's own doctors) can comment on them. I used to set a task for students which was to find the Friday Health Scare, which was usually pulled from one of those two journals, and said "X causes cancer: or "Y cures cancer".

    There's a Scott Adams quote I like: “Reporters are faced with the daily choice of painstakingly researching stories or writing whatever people tell them. Both approaches pay the same.”

  2. Mischief, nothing. This article, and the credibility lent to it by The Lancet, caused deaths.

    The Lancet's shameless pursuit of headlines corrupted them, and this is the result. Rather than simply retracting this article (talk about closing the barn door– 10 years too late) they should announce plans to return their focus to the validity of the work they publish.

  3. I disagree. The way that Lancet used its "star chamber" license to condemn a study on methodological grounds when the results are actually accepted by most clinicians who specialize in treating autistic kids and certainly corroborates with the experience of most autistic parents is a joke.

    As to "evidence", the best evidence is what you can touch hear and feel. Dr. Gelman is a bayesian specialist, so why is it that studies relating vaccines to autism get rejected at p values of 0.15? Is there something sacred about p=0.05? And with p values of 0.15, how does the mainstream press justify its headlines of "a new study finds no relationship between vaccines and autism"?

    You people should be ashamed of missing out one of the biggest ways bayesian inference could distinguish itself from the classical school.

  4. @mg: I have to disagree about autism-specialist clinicians' opinion of the MMR/GI/autism study published in The Lancet. I know many ASD specialists (I am married to one) who would like to see the author lose his license and Lancet editors fired.

    @Zubon: You're being a little harsh using the word "hallucination". While the alleged MMR vaccine link to autism in question has no scientific basis, there is evidence that some children may have elevated risk of autism if exposed to high doses of mercury-containing vaccine.

    Of course, vaccines with mercury preservative have not been part of the standard pediatric immunization schedule in the US for several years. Information that should get Bayesians excited.

  5. @anon,

    Except for the fact that the "thimerasol-free" vaccines still contain traces of thimerasol — an organic mercury compound for which no lower amount is acceptable for human consumption (as per our EPA). Moreover, the thimerasol that was replaced was replaced by aluminum, which is toxic by itself (hello Alzheimers?) only potentiates the neurotoxicity of any trace amount of Hg.

    @Zubon,

    You're the one hallucinating, and the "null" hypothesis of genetics being the cause is getting debunked daily. Especially with the horrific rise of ASD within the US over the last 2 decades. Please don't give the idiotic "better diagnosis" canard either. It's not like parents and doctors were oblivious is 1985. Even the AAP guidelines for diagnosis have not really changed since then.

    You still haven't told me why p-values of 0.12 on such an important topic prove there is NO link. I'm waiting to hear about it.

    @everybody else,

    Look at the parent reports of the efficacy of chelation versus all other biomed interventions for autism. It has the highest rating of improvements to adverse reactions around. Do parents' reports of their own children's improvements not mean anything to you supposed "evidence-based" scientists?

    Or are you all going to use a more complex epicycle theory to explain why the earth really is at the center of the universe?

  6. @anon,

    Why would you advocate the Dr. losing his license when all the parents who had their kids treated by him stand by him? Sounds like "star chamber" material.

    Why instead don't you go after the hundreds of GPs and rheumatologists that prescribed Vioxx and Celebrex for all manors of arthritis and other pains and led to 150,000 heart attacks per year? I'll tell you why not. Because it's easier to go after weaker foes. If you do advocate all those Drs losing their licenses, at least you'll have been intellectually honest and consistent. But, I'm guessing your not, and you're just shilling for the powers-to-be.

    Clearly, the doctors prescribing Vioxx and Celebrex were responsible for horrific damage when most GPs and many others had routinely talked within the industry about the dangers of these drugs.

    First take them on, and then you'll have far more credibility speaking as a patient's advocate on who should and should not keep licenses.

  7. @anon,

    You write:

    Of course, vaccines with mercury preservative have not been part of the standard pediatric immunization schedule in the US for several years. Information that should get Bayesians excited.

    Oh really? So both the flu shot and the H1N1 shot that they were peddling on everyone like crazy a few months ago was my hallucination. What do you think is the adjuvant in those shots? Yep, it's thimerasol. And yes, the recommendation for young kids (and pregnant mothers) taking both is so routine and relentless that one might confuse it for being a part of the "standard pediatric immunization schedule".

    Please stop making grotesquely incorrect statements.

  8. @Zubon,

    If the relationship between neurological problems and vaccines are hallucinated and not real, then you won't mind serving as a guinea pig for the following experiment:

    Let me give you 40 flu shots in one day, which if you're about 175 lbs is roughly comparable to a 17 lb child getting 4 shots in one day. If you are fine after than, then I'll rest my case.

    Otherwise, keep your mouth shut because "revealed preference" shows that I'm not hallucinating.

  9. @mg: You had me at "chelation". Ultimately, all I can say is that there are many big-hearted, empathetic scientists engaged in the fight against autism (some with children on the spectrum). Their findings have pointed to a combination of genetic and, yes, environmental correlates. Their research, however, does not indicate that the answer is contained in a handful of syringes.

  10. This topic reminds me of what happened not quite a century ago in Scotland.

    A couple of lads rigged up something to a toy submarine and floated it in a nearby lake. They then took pictures of it and showed them off, albeit with some considerable prevarication.

    Who would have thought that for many decades thereafter, suddenly people were seeing "Nessie" left and rght? Yet never before the boys' little joke?

    We don't hear of Nessie much anymore. The antivaccine lobby should take a lesson from this.

    Just think, if Wakefield's et al "paper" was never published, we wouldn't be having this discussion. And all the antivaccine groups advocating irresponsible behavior would be finding a serotonin fix from faux indignation over some other topic.

  11. @anon,

    You obviously don't know anything about chelation. IV chelation is not effective anyway for a wide variety of reasons. Oral seems to be far safer and far more effective. So, your jabs are silly and ineffective.

    More importantly, you write:

    Ultimately, all I can say is that there are many big-hearted, empathetic scientists engaged in the fight against autism (some with children on the spectrum). Their findings have pointed to a combination of genetic and, yes, environmental correlates.

    Just who are these "empathetic" scientists and what environmental correlates did they find besides vaccines? Let me guess, they are NOT the scientists (equally empathetic to the plight of spectrum kids) who DID conclude that the main environmental culprit is vaccines and most likely, the toxic metals in vaccines. Thus, chelation is the natural prescription for that.

  12. @JohnnyZoom,

    Wakefield's paper is not the only one published that establishes a link between vaccines and autism. A more compelling evidence is the efficacy of chelation treatment for autistic kids. Of all the interventions that parents undergo, chelation has the most significant rate of symptom improvement with the least adverse effects, as reported by the parents.

    Instead of putting your head in the sand and only listening to the medical community star chamber, go in the field and actually talk to parents who are attempting biomedical interventions to help their kids.

    Once again, I challenge any one of you "pro-vaccine" guys to step up and truly show your courage. If vaccines are perfectly safe, let me give you 40 flu shots in one day, which is roughly equivalent of what small children receive.

    If you're not willing to man up and take the shots, then I suggest you shut up and rest your case.

  13. mg, mg, mg…

    1. P-values greater than 0.05 don't prove anything. Any statistician or epidemiologist can tell you that. What "proves" (and then only tentatively) a scientific consensus is when study after study fails to find a significant link, which is exactly the case in the vaccine-autism debate. Which leads me to…

    2. Please provide specific examples of the "numerous studies" and "numerous scientists" that have concluded that vaccines cause autism. You are misleading people who do not follow the issue. The tiny number of studies and scientists who have suggested a link have been thoroughly refuted.

    3. Please provide your calculations for why 40 flu shots is equivalent to a small child would receive. Are you referring to total liquid volume? Thimerosal? Grams of antigen? Number of antigens? These things vary by vaccine, so please specify which, exactly, you think is causing autism.

  14. Please recognize that this is a respectable and reputable blog, with many posters renowned (in a good way, Dr.W.) in their respective fields. Readers come here expecting intelligent and logic- and evidence-based discourse informed by top-of-the-line statistical thinking, and absent polemics and straw men. Please stop advocating for an intervention proven to be highly dangerous and never proven effective for the relevent malady in this thread. Autism is not a Chernobyl accident. Chelation has killed kids.

    As my son's dev ped has told me on numerous occasions regarding reports/papers on the use of chelation / hyperbaric therapy / some (not all) exotic diets, etc., "Look closely. That is not evidence. That is an advertisement."

  15. @JohnnyZoom,

    You write:

    Please recognize that this is a respectable and reputable blog, with many posters renowned (in a good way, Dr.W.) in their respective fields.

    That's precisely why I read Dr. Gelman's blog. There should be smart people reading this.

    Please stop advocating for an intervention proven to be highly dangerous and never proven effective for the relevent malady in this thread. Autism is not a Chernobyl accident. Chelation has killed kids.

    How has the intervention been "proven" to be highly dangerous? You offer an assertion with ZERO proof. I know many many parents that have gone down the road of chelation with terrific success. Almost all of them have seen significant reductions in symptoms (about 75% or so), and some of them have seen the autism diagnosis go away. So you tell me, is first hand evidence, the kind that one can touch and feel, not "scientific" enough for you?

    You're arbitrary attack on chelation is very telling. Chelation had ONE death, which was a result of the WRONG chemical being used, something that happens nearly everyday in most metropolitan hospitals. Closer to the issue, I don't see you advocating for getting rid of Ritalin, a drug that we KNOW kills 20-50 children a year because doctors prescribe it for "off-label" uses.

    This is what I mean. You guys are incredibly hypocritical when it comes to autism and the treatments that many many parents have found to be enormously successful. Hundreds of millions of dollars have gone into the supposed "genetic" smoking gun for autism in the last 10 years with NOTHING to show for it. Absolutely ZERO treatments have come out of this genetic-based research. Parents have been clamoring for a study on the efficacy of chelation, yet NIH and most medical institutions ignore it. The few studies that have been done with alternative groups have proven some degree of effectiveness.

    And NO, I won't stop putting forth assertions that are merely inconvenient to your Star Chamber opinions that are clearly not based on anything other than what you've been reading from mainstream sources.

  16. @Ryan,

    You write:

    P-values greater than 0.05 don't prove anything. Any statistician or epidemiologist can tell you that.

    I guess you failed to grasp the difference between bayesian views of probability versus the classical school, epistemology vs. ontology. If p values can be interpreted as "degrees of belief", which in many cases they rightly can, then a p value of 0.12 in a very important question DOES prove a lot. It proves that if you put a gun to someone's head and ask them if they believe in the model or not, they would rationally believe. It's not at the level of Newtonian physics, but nothing in the complex sciences (medicine, economics, psych, etc.) are.

    BTW, epidemiologists just shouldn't talk. Their whole profession has been discredited completely over and over again. Every year, they come out with idiotic statements that all contradict what was said before. Wine is bad for you. No, no, wine is good for health. Eggs are bad. No, no, eggs are good. Animal fats cause heart disease. Animal fats lead to heart disease.

    Come on. That whole profession is about as credible as voodoo chanters.

Comments are closed.