After reading a paper published in The Annals of the Royal College of Surgeons of England last March, Andrew Thomas, an orthopedic surgeon in the UK, noticed that it was very similar to an article published the previous December in another journal.
He wrote a letter to the editor of Annals, notifying the journal of the similarity between its paper, “The possible effect of different types of ventilation on reducing operation theatre infections: a meta-analysis,” and “The effect of type of ventilation used in the operating room and surgical site infection: A meta-analysis,” published in the journal Infection Control and Hospital Epidemiology.
In his letter, Thomas pointed out several apparent mistakes in the analyses, and also noted similarities between both papers and a 2017 article published in The Lancet Infectious Diseases , which has been cited 93 times, according to Clarivate’s Web of Science.
Despite the similarities Thomas noted, which another surgeon verified with anti-plagiarism software, the journal has not retracted the paper, but flagged it with an expression of concern so readers can “draw their own conclusions.” At the same time, the journal retracted an unrelated article that was also found to be similar to one published elsewhere, then the retraction was changed to an expression of concern, and now neither notice appears online.
Though it is unclear when the March 2021 Annals paper Thomas called out was originally submitted for publication, it was accepted in July 2020. The article in the other journal does not indicate when it was submitted or accepted, only that it was published in December 2020. There is no obvious overlap between either the author list or the authors’ institutions, which include a few different hospitals in China. Both papers have been cited fewer than 15 times put together.
My main concern in sending the letter was the bad science involved. The paper appeared to be one of a number of rather similar papers purporting to be a meta analysis of registry data on the role of Ultra Clean Air in reducing the risk of deep infection in joint replacement surgery. The plain fact is that registry data is never as good as a randomised trial, and meta analysis does not improve it. Frankly I thought that the paper was total trash, for the multiple reasons outlined in my letter, so I decided to check what the source papers were for the meta analysis by a simple PubMed search, which is when I found the other Chinese paper.
My concern is that the paper is dangerous because it may make surgeons think that infection in joint replacements is not a problem that needs Ultra clean conditions to reduce infection risks to patients. My letter was designed mainly to tell the readers not to take any notice of the paper. I decided that the paper was a clear duplicate, but who had copied from whom was not a matter for me, but for the editor.
In an editorial published in the September 2021 issue of Royal College of Surgeons of England Annals, editor-in-chief Benedict Rogers highlighted that the journal had issued an expression of concern for the article. Rogers wrote that the journal had decided not to retract the article so readers could “draw their own conclusions after reading the article.”
The expression of concern reads:
It has been brought to our attention that this article is very similar to a publication in the journal, Infection Control & Hospital Epidemiology. Due to the timing of publication and difference in authorship between papers, we have decided to not retract this publication in this case, but would like to highlight the nature of this duplicate publication.
After this editorial was published, Chris Roche, a cardiothoracic surgeon in the UK, ran the two papers through anti-plagiarism software. He wrote a letter to the editor about his findings:
Although it might seem obvious that these two papers are the same article, I still ran them through my institution’s anti-plagiarism software and, as expected, the software flagged duplication. Do you run articles through anti-plagiarism software at the time of submission (as all journals from the publisher Frontiers do automatically and transparently now at the time of submission using their Artificial Intelligence Review Assistant, AIRA)? This would have avoided the problem from the outset or at least provided evidence of who submitted first (although this still does not tell you anything about true authorship). If you had done this, then you might also have found a higher than usual level of textual crossover with a four-year old article in the Lancet Infectious Diseases (Bischoff et al. 2017).
Of course, this is the same 2017 article that Thomas had identified.
The paper in Infection Control and Hospital Epidemiology has not been retracted either, but received this corrigendum in September:
The editors of the journal would like to alert readers of this article that the reported systematic review is an extension of a previously published systematic review of the same clinical questions by Bischoff et al. The authors have extended their systematic review to include two studies published after the cutoff point of the previous systematic review. Although cited in their article, the authors did not make the connection with the previous publication transparent to readers of the article.
In the same September 2021 editorial, Rogers said the journal had retracted another article with a twin, “Tumour-to-tumour metastasis: breast carcinoma to an olfactory neuroblastoma,” published online in May 2020 and cited twice. The paper is a case report describing a patient seen at Royal Surrey Country Hospital in the UK, where the authors are doctors. The paper was accepted for publication in the journal in January 2020. Corresponding author Anthony Bashyam declined to comment.
The similar paper, “Lobular to Lobule: Metastatic Breast Carcinoma to Olfactory Neuroblastoma,” was published in Head and Neck Pathology in July 2020. The articles share one author: Silvana Di Palma, a pathologist at the Royal Surrey Country Hospital.
Of the authors of the study in the Head and Neck Pathology, only Di Palma is affiliated with the hospital where the patient was seen, while the two other authors are affiliated with U.S. medical centers in Texas and California. Lester D. R. Thompson, the corresponding author of the paper in Head and Neck Pathology, is also the founding editor and current social media editor of the journal. The paper was received by the journal in June 2020 and accepted the following July. Thompson did not reply to a request for comment from Retraction Watch.
Following the retraction of the Annals article, the editors apparently changed their minds and took back the retraction. A note detailing the change in the paper’s status, which no longer appears online, stated:
It has come to light that the following paper was retracted in error. The publisher should instead have issued an ‘expression of concern’. The concern is currently being investigated and an update will be provided here when a decision has been reached. In the meantime, we apologise to the authors for the error and any associated distress.
The expression of concern, which is also no longer online, read:
An expression of concern has been articulated at the request of the Journal Editor(s) and the Publisher, for the following article: Bashyam A, Grammatopoulou V, Crook T, Di Palma S, Sunkaraneni VS. Tumour-to-tumour metastasis: breast carcinoma to an olfactory neuroblastoma. Ann R Coll Surg Engl. 2020; 102(6): e118–e121. doi: 10.1308/rcsann.2020.0038.
This article was brought to our attention, as it is similar to an article published in the journal, Head and Neck Pathology: Swimley, KM, Di Palma, S and Thompson, LDR. Lobular to Lobule: Metastatic Breast Carcinoma to Olfactory Neuroblastoma. Head and Neck Pathol 2021; 15: 642–648. doi: 10.1007/s12105-020-01199-z.
We have chosen to highlight the similarity of the two articles so that readers can make their own informed decision while we investigate the concern further.
The online article in Head and Neck Pathology has not been retracted, has no message of concern, and does not have any indication that the article is similar to one in another journal.
Rogers did not respond to an email from Retraction Watch requesting further comments on the two articles. When he wrote the editorial flagging the duplications, he announced that the journal had changed its policy on the use of anti-plagiarism software: While it was used in some cases before, the software is now used with every paper submitted to the journal.
Roche said he was struck by the ease with which he, a random reader, was able to identify the plagiarism from the 2017 Lancet paper, even when the journal had missed it. At the time he ran the articles through anti-plagiarism software, he was a PhD student studying cardiac bioengineering at The University of Sydney Northern Clinical School in Australia. He said:
I think the most interesting thing about this story is that I was just some eccentric guy in Australia with a laptop and some time on my hands. If you plagiarise these days someone’s going to notice.
Adam Marcus, a cofounder of Retraction Watch, is an editor at Medscape.