A vial of hydroxychloroquine.

-

David J. Phillip / AP / SIPA

The leading medical journal The Lancet has announced that it is stepping up its proofreading process, having had to withdraw a high-profile study on hydroxychloroquine in the spring due to doubts about its reliability.

This fiasco led the journal to want to "further reduce the risk of bad behavior in research and publications," she explained. 

Proofreading by a committee of independent experts, the “peer-reviewing”, is supposed to be a guarantee of the quality of scientific publications.

For work using large databases, such as the one on hydroxychloroquine withdrawn in early June, "at least one of the reviewers (should) know the details of these data" and be able "to understand and comment on their value and their limitations. in relation to the subject ”of the study.

For larger databases, the Lancet will also have recourse to a “data science expert”.

A change in knowledge of data 

At the end of May, the Lancet published a study concluding on the ineffectiveness of hydroxychloroquine against Covid-19 and even on its dangerousness, a work based, according to its authors, on data from nearly 100,000 hospitalized patients worldwide.

The WHO had in the process stopped the trials of the drug, but doubts had quickly arisen around the data collected by the company Surgisphere (whose leader was mentioned as co-author of the study) which had refused to reveal the data. details.

Three of the four writers ultimately "retracted" it and the Lancet issued its "apologies".

For all studies, the Lancet will also require more written commitments from the authors and that all the authors of the same study engage their responsibility.

For example, “more than one author” must have had “direct” access to the raw data of the study and have “verified” them.

In addition, at least one of those who verified the integrity of the data must be a researcher, unrelated to the commercial entity (a company for example) that would have provided them if necessary.

A document should specify what data will be published, when and how it will be published.

In another text released Friday, the Lancet points out that with the pandemic, it has become "particularly complicated" to find reviewers and that "in some cases, we have received five times more manuscripts than normal."

To date, the effectiveness of hydroxychloroquine against Covid has not been scientifically demonstrated.

Society

Coronavirus: Doctors' salaries, chloroquine ... We have verified three statements by Didier Raoult in the Senate

Health

Coronavirus: European Medicines Agency warns of serious side effects of chloroquine

  • Drug

  • study

  • Covid 19

  • Coronavirus

  • Health