Neurology Journal Fixes False Claim in MUS Paper–But Fails to Publish a Correction Notice (Ironically, I Have Added a Correction to the Post!)

By David Tuller, DrPH

UPDATE: August 4, 2021

Dr Villemarette-Pittman, the managing editor of Journal of the Neurological Sciences, has informed me that she has learned from Elsevier that a corrigendum has in fact been written and will be published in the near future. She also informed me that she plays no role in deciding on or setting policy about corrections and corregendi, and when or how they are published. I am very glad to hear about the corrigendum. Had I been informed earlier in the cycle of e-mail exchanges that the lack of a corrigendum did not mean that none was forthcoming and that a response from Elsevier was being sought, I would have written the blog post differently and/or updated it sooner. Dr Villemarette-Pittman had responded on behalf of the journal to my initial e-mail about the error, so I assumed she understood the practices and policies around corrections and corregendi. I apologize to Dr Villemarette-Pittman for having made that mistaken assumption.

**********

*Correction: I initially referred to Dr Villemarette-Pittman as a “neurologist.” I did not mean to indicate that she is a clinician but rather an academic specialist. She is an associate professor of research in neurology and oversees neurological research at her university. I apologize for any confusion.

*Note: Also, after reading this blog post, Dr Villemarette-Pittman noted that she is not the editor-in-chief nor the publisher of the journal, and clarified that she does not make decisions about corrigenda. (For the record, I have not referred to her as the editor-in-chief or the publisher, nor have I stated that she makes decisions about corrigenda.) I wrote to two editors initially about the error, and she responded. That is why we have been in correspondence over this issue.

I’ve recently bugged two separate journals to correct the same error—a mis-citation of a study concerning the estimated costs to the UK’s National Health Service of so-called “medically unexplained symptoms” (MUS). One of the journals, Psychological Medicine, has corrected the mistake and has also published a standard notice called a corrigendum that alerts readers to the fact that a change has been made.

In contrast, Journal of the Neurological Sciences has corrected the error in the text of the paper in question but has not published a corresponding corrigendum—or at least not one that I can find on the site of this Elsevier publication. Perhaps one is coming, although the offending sentence was fixed weeks ago, according to my communications with the journal. It is not acceptable for a journal to make a substantive correction and not indicate that this has happened.

I’ve got a real thing about this particular error because it has been repeatedly committed in multiple articles since the paper containing the original finding, Bermingham et al, was published in 2010. Bermingham et al estimated that the cost of treating working-age people with MUS accounted for 10% of the NHS budget for that demographic group. (The study involved England only.) That statistic has routinely been mis-rendered as 10% of total NHS costs—that is, for all ages. In fact, the study estimated that the total spent on working-age people with MUS accounted for around 3% of the NHS budget, not 10%. (In making this point, I am not endorsing Bermingham et al’s methodology or its estimates.)

The same time period has also witnessed a forceful push to expand NHS mental health care for patients diagnosed as experiencing MUS—an expansion implemented through a program called Improving Access to Psychological Therapies (IAPT). Medical experts in the MUS field have been among those advising the government on this IAPT expansion. They have also been among those mis-citing Bermingham et al in a way that dramatically increases the apparent health care burden of MUS. It is self-evident that demonstrating the high cost of MUS to the NHS would be likely to generate or bolster support for expanding services for these patients—as advocated by some of those doing the mis-citing.

I’m not suggesting that these mis-citations have been intentional. In fact, I assume they were all unintentional. This episode demonstrates how a false data point or meme takes hold and gains currency because it fits the preconceptions and the interests of those perpetuating it. Some of those mis-citing the claim likely picked it up from a previous mistaken reference without actually checking the original source themselves. After all, it made sense to them that MUS was such an enormous problem that it could account for 10% of the total NHS budget.

Some investigators have been repeat mis-citers of Bermingham et al. That is the case with Professor Anthony David, a neurologist at University College London and the senior author of the article published by Journal of the Neurological Sciences, called “Medically unexplained visual loss in a specialist clinic: a retrospective case-control comparison.” He also was senior author of the paper in Psychological Medicine that was corrected due to my intervention. And he is a co-author of the published protocol for a major clinical trial, which included the same mistake and was corrected in 2019 after I alerted the journal.

This was the sentence in the first version of the paper published by Journal of the Neurological Sciences (with the original misspelling of the first author of the cited study retained): “In estimation of the associated health costs of medically unexplained symptoms, Birmingham [sic] & colleagues propose that healthcare utilisation amounts to £3billion per annum, or 10% of total NHS expenditure.” In the current version of the paper, the sentence now reads: “In estimation of the associated health costs of medically unexplained symptoms, Bermingham & colleagues propose that healthcare utilisation amounted to £3billion in the year 2008-2009, or approximately 10% of total NHS expenditure on these services for the working-age population during this period.” (I’m glad they also corrected the spelling of Bermingham et al.)

Professor David* [I initially wrote Professor Anthony here–I apologize for the error], his co-authors and Journal of the Neurological Sciences have an obligation to make corrections in a transparent manner. Fixing sentences but giving the appearance that nothing  has been done does not meet this basic academic standard and cannot be justified. Interestingly, the same journal published a formal corrigendum last summer when a misspelt name was corrected. Certainly a mistake that has the effect of more than tripling the apparent costs of MUS to the NHS is of sufficient gravitas that the journal should be willing to accord it the same treatment.

**********

Letter to editor of Journal of the Neurological Sciences

I have sent the following letter to Dr Nicole Villemarette-Pittman, an assistant professor of research in neurology* [this sentence has been corrected] at Louisiana State University and the journal editor with whom I had been corresponding about the error:

[After reading this blog post, Dr Villemarette-Pittman sent an e-mail noting that she is not the editor-in-chief nor the publisher of the journal, and does not make decision about corrigenda. I have not called her the editor-in-chief or the publisher, nor have I claimed she makes decisions abou corrigenda. I wrote to two editors initially about the correction, and she responded. That’s why I have been in correspondence with her.]

Dear Dr Villemarette-Pitman–

I am writing a post for Virology Blog about the apparent decision by the Journal of the Neurological Sciences not to indicate or flag the correction of O’Leary et al with an actual corrigendum. (If there is a corrigendum and I can’t find it for some reason, please advise!) Do you want to comment on why the journal has chosen not to issue a corrigendum for this correction? It does seem rather unusual to correct a substantive factual error and not alert readers that anything has been changed. 

This was not a typo or a grammatical error. The mistake had the effect of more than tripling the presumed impact of so-called “medically unexplained symptoms” on the National Health Service’s budget. According to the corrections policies of Elsevier, your publisher, a corrigendum is appropriate when “a small portion of an otherwise reliable publication proves to be misleading (especially because of honest error).” That certainly seems to be the appropriate approach here–or should have been. The claim in the paper was highly misleading and presumably unintentional.

The International Committee of Medical Journal Editors has also highlighted the importance of identifying a correction. When a correction is made, according to the ICMJE guidelines (as cited in Elsevier’s website), “the journal should publish a correction notice as soon as possible detailing changes from and citing the original publication.” 

In other words, simply correcting a sentence that includes a factual error–as you have done in this instance–is insufficient. By any standard, transparency requires the publication of an explanatory corrigendum as well.

I will likely post this comment on the journal’s actions tomorrow (Friday). If I receive a response from you after I have already posted, I will be happy to update the blog.

Best–David

David Tuller, DrPH
Senior Fellow in Public Health and Journalism
Center for Global Public Health
School of Public Health
University of California, Berkeley