E. Petersen /Science
By Charles Piller
This story was supported by Science Fund for Investigative Reporting. Please help Science pursue ambitious journalism projects.
In June 2020, in the biggest pandemic research scandal so far, two of the most important medical journals, each portrayed a high-profile study of COVID-19 patients. Thousands of news articles, tweets and academic comments highlighted the scandal, but many researchers apparently did not notice. In an examination of the 200 most recent academic articles published in 2020 that cite these articles, Science found that more than half – including many in major newspapers – used the wretched articles to support scientific discoveries and did not notice the retractions.
COVID-19 “is such a hot topic that publishers are willing to publish without due verification,” even in the face of retractions that made headlines globally, says Elizabeth Suelzer, a reference librarian at the Medical College of Wisconsin who wrote about problematic quotes to a study portrayed in 1998 in The Lancet falsely linking vaccination to autism.
Both papers collected from COVID-19, one in The new English medical journal (NEJM) and the other in The Lancet, were based on what appeared to be a huge database of patient records compiled from hospitals around the world by Surgisphere, a small company operated by vascular surgeon Sapan Desai, who coauthored each article. May 22, 2020 Lancet The paper ostensibly showed that hydroxychloroquine, an antimalarial drug promoted by President Donald Trump and others, could harm rather than help COVID-19 patients. Its publication led to the temporary suspension of a large clinical trial and sparked an already controversial debate about the drug, which proved to be no help against COVID-19. The 1 of May NEJM The article corroborated other evidence that people who already take certain blood pressure medications were not at greater risk of death if they developed COVID-19.
However, doubts soon arose about the validity, and even the existence, of the Surgisphere database, and the retractions followed on June 4. But, of the 200 articles examined by Science—All published after retractions – 105 cited one of the unfortunate studies inappropriately. In several cases, it was a primary source for a meta-analysis combining several studies to draw comprehensive conclusions. Most studies were cited as scientific support or context. Science they also found a handful of articles that uncritically cited an influential April preprint based on the same data set from Surgisphere, which described the anti-parasitic drug ivermectin as beneficial in critical cases of COVID-19. (There is no standard way to shrink preprints, however.)
Ivan Oransky, co-founder of Retraction Watch, says that such errors occur because “people do not check references intentionally or negligently”. Many authors copy and paste lists of seemingly relevant citations from similar articles without actually reading them, he says. “That’s scary. It is terrible, but common. “
Many of the contaminated citations appeared in articles published by little-known journals, but at least a dozen were found in major publications. For example, three articles in PLOS ONE, the prominent open access newspaper, cited articles portrayed in discussions about pandemic conditions in Europe. A December 28 newspaper in the Proceedings of the National Academy of Sciences (PNAS), One of the most influential magazines, discussed the risks and benefits of drugs to treat COVID-19 and noted the Lancet retraction in its citations, but the text only noted the newspaper’s hydroxychloroquine findings as “controversial”.
Related
The editors of these two publications said they would correct the references and take steps to avoid such problems in the future. Renee Hoch, a PLOS ONE editor and ethics manager of the publication, wrote in an email that the publication depends on authors and their voluntary editors to verify citations, and she was taken aback when contacted by Science. “We are currently following this issue with high priority, in light of the public health implications and the ongoing research of COVID-19,” she wrote.
Hoch added that relying on portrayed works, “either directly or in the form of supporting references,” can be harmful. “[W]here, the withdrawal work has clinical implications, this can result in direct risks for patients ”.
In a written response to questions about the quote from the Lancet May Berenbaum, editor in chief of the PNAS, said: “The authors really should have removed the quote, added more text on why it was included, or quoted the disclaimer.” Since no editor or reviewer detected the problem, she said: “I intend to discuss with the team the incorporation of this screening in the processing of the manuscript”. A co-author of the article, biostatistics Clelia Di Serio, from the Vita-Salute San Raffaele University, later said the reference to Lancet the paper would be eliminated.
Some journal editors have noted mitigating factors. Inside Leakage, an important medical journal, a December article on ischemic stroke in referenced COVID-19 patients NEJM not to mention shrinkage. Leakage Editor-in-chief Ralph Sacco wrote in an email that the retraction occurred after the initial receipt of the article. However, a review of the article was resubmitted months after the event, according to the newspaper itself. Sacco said he would not issue any corrections because the retraction “is not relevant to the conclusions”.
A December 16 article on SARS-CoV-2 genetics in Nature Communications, another high-profile newspaper, also cited the NEJM article without reference to retraction. Elisa De Ranieri, editor in chief of the magazine, said Science his diary does not routinely check for “retractions or other post-publication updates”. One of the lead authors of the article, biomathematician Maik Pietzner of the University of Cambridge, said that although the article was submitted after the retractions occurred, it was written in advance and “the current pandemic requires an immediate response”. However, the article was published 4 months after its submission.
Suelzer says that inappropriate quotes from portrayed articles are difficult to excuse. Retraction Watch publishes a free retraction database that has been integrated with a series of automated services to verify quotes, including scite.ai, Zotero and RedacTek. Failure to use such tools “is a disservice to readers and researchers,” says Suelzer. “They are very low bars.”
Still, Oransky estimates that, in biomedicine, up to 90% of the citations of articles portrayed do not mention their fall from grace. “Half of the time [as seen with the Surgisphere papers] it’s an improvement. This is what is shocking. “