‘Beggers’ can’t be choosers as another meta-analysis is retracted

A sample funnel plot, via Wikimedia

A group of researchers in China may be asking for a refund after, they claim, they got bad advice from a course in writing meta-analyses that led to a retraction for plagiarism and other problems. 

They may not be alone. We’re aware of at least nine articles with similar issues that have been retracted so far, part of a batch of eye-catching meta-analyses. 

Suspicions about the now-retracted 2014 paper — and dozens of others — emerged in October of that year, when Guillaume Filion, now at the Centre for Genomic Regulation in Barcelona, wrote a blog post about the articles. Filion and a colleague, Lucas Carey, of Peking University, had noticed a flurry of meta-analyses by researchers in China that had appeared in CISCOM, the articles repository for the Research Council for Complementary Medicine. As Filion wrote: 

Available since 1995, it used to be mentioned in 2 to 3 papers per year, until Feburary 2014 when the number of hits started to skyrocket. Since then, “CISCOM” surfs a tsunami of one new hit per week.

Looking deeper, Filion and Carey found 32 meta-analyses, from 28 groups, with evidence of plagiarism, such as identical passages and figures — including a tell-tale flowchart called a “Begger’s funnel plot.” 

If that doesn’t sound right, that’s because it’s in fact a “surrealistic fusion of Begg and Egger [Colin Begg and Matthias Egger both gave their name to a test for publication bias],” Filion wrote.  

Filion and Carey, both geneticists, also conducted a “geneaology” of the work to track the patterns of errors and other curious features of the articles. They concluded that the researchers could not have written the articles on their own: 

each paper derives from several templates. … it seems that the writers borrow attributes from different papers and combine them.

Most of the studies were submitted between October and November 2013, whereas the first study was published in December 2013 and indexed on PubMed in January 2014. How could the authors have plagiarized so many manuscripts before they were available?  … The only solid hypothesis is that the same ghostwriter wrote all the CISCOM meta-analyses. 

As the notice indicates, the The Scientific World Journal, a Hindawi title, accepted Filion and Carey’s analysis: 

The Scientific World Journal has retracted the article titled “VEGF Genetic Polymorphisms May Contribute to the Risk of Diabetic Nephropathy in Patients with Diabetes Mellitus: A Meta-Analysis” [1]. This article is one of a series of very similar meta-analyses written by different authors that were published in 2014 and 2015, characterized by searching the complementary and alternative medicine database CISCOM despite the topic not being about CAM [2], as is also the case with this article. The overlaps of wording with these articles are concentrated in the Materials and Methods and Results sections, and a paragraph in the Discussion. The article also mentions “Begger’s test,” which is a mistaken combination of “Begg’s test” and “Egger’s test” and is also characteristic of this series of articles. The article inappropriately mentions other cancers and genes, including gastric cancer in the abstract and “breast cancer” and “IFN-γ gene” in Figure 1 in the eligibility exclusion box, whereas the article is about diabetic nephropathy (DN) and VEGF.

Additionally, the assessment of publication bias may be incorrect. The statement that “Egger’s test also did not display strong statistical evidence for publication bias (allele mode: t = 2.92,  and dominant model: t = 2.53, , resp.)” is made despite the test result being statistically significant and the funnel plots are said to be symmetrical, though the smaller studies (with greater standard error) show greater associations with DN, which is an indication of publication bias.

The authors said they attended a course on writing meta-analyses that lead to incorrect data analysis and they apologised for the mistakes. However, they asked to retract the article (though they did not approve the content of this notice) and this was approved by the editorial board.

Filion told us: 

A few months after the blog post, one Chinese scientist contacted me and asked me to take down my blog post because it was doing a lot of damage to his career. I answered that I would keep the post online but would publish whatever he had to say along with the blog post. He did not reply. I noticed that one of the articles was retracted shortly after the blog post but most of the others remained online.

Filion sent us a list of nine already-retracted papers from the list he flagged: 

Whereas The Scientific World Journal notice provides plenty of details, the statements for most of these are more Spartan. For example, the one for “Relationships of COX2 and MMP12 genetic polymorphisms with chronic obstructive pulmonary disease risk: a meta-analysis,” which was retracted in August 2015, states: 

The Publisher and Editor retract this article in accordance with the recommendations of the Committee on Publication Ethics (COPE). After a thorough investigation we have strong reason to believe that the peer review process was compromised.

This one, from DNA and Cell Biology, does a bit better — although we have no idea what the authors mean: 

All of the authors of two articles published in DNA and Cell Biology have requested the retraction of this meta-analysis article. The reason they state is “severe conflicts of author sequences.”

Chen, Y.-Z. et al. (2014). Diagnostic performance of serum macrophage inhibitory cytokine 1 in pancreatic cancer: a meta-analysis and meta-regression analysis, DNA Cell Biol 33(6):370–377. DOI: 10.1089/dna.2013.2237. Epub 2014 Mar 4.

On PubMed.gov, the abstract is accompanied by a comment from Guillaume Filion indicating that the paper bears striking similarity to other publications.

These authors have also recently retracted three papers, one from Molecular Biology Reports and two from Tumor Biology.DNA and Cell Biology is dedicated to upholding the strictest standards of scientific publishing, and will not tolerate any improprieties.

Update, 1700 UTC, 8/26/20: Ben Dickinson, Hindawi’s research integrity manager, tells us:

We investigated a series of very similar meta-analyses based on the blog post we cited in that notice (http://blog.thegrandlocus.com/2014/10/a-flurry-of-copycats-on-pubmed). Using checks for text similarity and characteristic phrases, including the inappropriate use of CISCOM when the topic is not CAM, we identified concerns with several articles. The involvement of an editing company was confirmed in two cases and is likely for the rest. Each concerning article has been assessed internally and by the editorial boards – the following articles have already been retracted due to these concerns:

– BioMed Research International, “Retracted: Influence of Two Common Polymorphisms in the EPHX1 Gene on Warfarin Maintenance Dosage: A Meta-Analysis”, BioMed Research International, vol. 2019, Article ID 6910869, 1 page, 2019. https://doi.org/10.1155/2019/6910869
– BioMed Research International, “Retracted: Correlation between Serum Levels of High Mobility Group Box-1 Protein and Pancreatitis: A Meta-Analysis”, BioMed Research International, vol. 2019, Article ID 5190178, 1 page, 2019. https://doi.org/10.1155/2019/5190178
– The Scientific World Journal, “Retracted: VEGF Genetic Polymorphisms May Contribute to the Risk of Diabetic Nephropathy in Patients with Diabetes Mellitus: A Meta-Analysis”, The Scientific World Journal, vol. 2020, Article ID 6482307, 1 page, 2020. https://doi.org/10.1155/2020/6482307

There are a number of articles for which investigations are ongoing, and we have several additional retractions prepared. These are currently in the final stages of discussions with the relevant editorial boards and we hope to have these related cases resolved shortly. 

Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.