A study circulated in “pre-publication” had been widely relayed to serve disinformation on vaccines. Thanks to a proofreading by other scientists, the authors discovered an error in the calculations, which invalidates their entire study. This situation reminds us that not all publications can be relayed in the same way.
Studies whose content is not verified, but which are shared despite everything, this is one of the poisons of the pandemic. This contributes to an active circulation of disinformation about the health crisis. These problematic studies are generally “preprints” (which means prepublication). This type of work is disseminated on specific platforms, such as Arxiv and Medrxiv: these allow researchers to make their research available online and freely accessible.
But a study takes on the value of a “scientific publication” when it is then published in a journal, because the study then goes through a proofreading by an independent committee, an essential step which is called peer review (proofreading by pairs).
Disseminate studies in preprint who have not gone through the peer review, a fortiori when they are not part of a continuation of other work, therefore involves a great risk-taking. It was unfortunately common, over the course of the pandemic, for unreliable work to be widely reported. A recent story reminds us of this.
In mid-September 2021, a research team from the University of Ottawa (Heart Institute) broadcast in preprint a publication concluding that there would be a 1 in 1000 chance of contracting myocarditis after vaccination against the coronavirus. The problem is, this conclusion is wrong. The authors realized this later, after the study had already been widely publicized to support discourse of “anti-vaccine” disinformation.
When a mistake leads to misinformation
After the study was disseminated via Medrxiv, several scientists looked at the method and the results, during a process ofopen peer-review, that is to say a free re-reading. Several of them pointed to a decisive error in the calculations. To get a myocarditis risk rate of 0.1% (1 in 1,000), they divided the number of covid vaccines given in Ottawa by the incidence of heart problems.
Except that in this calculation, they greatly underestimated the number of doses of vaccine delivered. In the study, they claim that there were 32,379 in Ottawa. This is not the case: there have been 854,930 (at the time of this publication), which is 26 times more. During such a calculation, this difference changes absolutely everything: the risk rate becomes even smaller than 0.1%.
As a result, this error totally invalidates the study and its conclusion.
« In order to avoid misleading our colleagues and the general public as well as the media, we unanimously wish to withdraw this paper. »
The authors of the post themselves admitted the calculation problem. It looks like there was some confusion when reading an open source database: they got the wrong table. In a letter, they asked the Medrxiv platform to delete their study: “ In order to avoid misleading our colleagues and the general public as well as the media, we, the authors, unanimously wish to withdraw this paper due to its incorrect incidence data. The research team takes this opportunity to apologize and to thank the scientists who performed a careful review, recalling that the publication should no longer be cited as a reference. On the site of the University of Ottawa, they add to be ” sorry that this error led to misinformation ».
The presence of an error in a study preprint is not fundamentally a problem, since the process of peer review is also, precisely, for that. The study is not meant to be referenced when it is only preprint. The situation shows above all the real problem: the publication in preprint was misused, without any verification, when it did not yet have the legitimacy to be taken as proof of anything.
How to properly read a scientific study?
When you see a scientific publication being relayed, you must therefore check its status. a preprint can be quite interesting to approach, provided that it is made clear that its conclusion is not definitively recorded and that it is still subject to verification; or if there is a basic decryption with the authors and other scientists able to criticize the content of the study; or if the publication in question is part of a continuum of other work.
Finally, it should be remembered that at present, the serious studies published on an alleged link between myocarditis and mRNA vaccines have excluded a connection significant between the two. The risk appears to exist, especially in young men according to a CDC summary, but cases are extremely rare. This is why the benefits of vaccination far outweigh the risks, especially because vaccines reduce the circulation of the virus and very effectively prevent severe forms of covid – more common in infected people than are the side effects. problems that vaccinated people.
The continuation in video