Véronique Baudoux’s scientific contributions

Background: Compliance with research ethics legislation is a crucial issue for the protection of the rights of people taking part in clinical trials. Researchers have the opportunity to alert publishers when they have concerns about possible non-compliance with this legislation by other authors, and this opportunity for independent scrutiny is important. Nevertheless, some alerts may turn out to be unfounded due to a misunderstanding of the legislation or shortcomings in the analysis of the publications concerned. Methods: We present the analysis of the article by F. Franck et al. « Raising concerns on questionable ethics approvals - a case study of 456 trials from the Institut Hospitalo-Universitaire Méditerranée Infection » [1] in which the authors explain having sent numerous emails to different publishers to express their concerns about potential ethical breaches in 456 publications from a single French institution. We analyzed in detail the 248 studies that the authors present as suspect on the grounds that they all mention the same ethical number. We have also studied the texts of French law and the content of presentations given by institutions specializing in clinical research, explaining the framework of the Jardé Law. Results: Our detailed analysis of these 248 studies discovered that only seven of them report research involving humans. All the others are secondary research on existing biological samples and most of them present the results of genomic analyses of new bacteria discovered by the authors. In France, as in many other countries, research on biological samples that have already been collected for care purpose or as part of a previous study and research on micro-organisms are not research involving human subjects. The rules that apply in these specific contexts are therefore different. Conclusions: Assessing the ethical conformity of scientific publications must be based on a precise knowledge of the legislation and a rigorous analysis of these publications. When researchers who alert publishers confuse research involving the human person with research not involving the human person, the publishers are in danger of being misled by these unfounded alerts. This can have deleterious consequences for the authors but also for the entire scientific community through the distrust generated towards medical research.

Despite a peer review process prior to being published in Biomedicine & Pharmacotherapy, Pradelle et al, "Deaths induced by compassionate use of hydroxychloroquine (HCQ) during the first COVID-19 wave: an estimate" is cluttered with several scientific integrity issues, one of them being the lack of validation against real life data. Mathematical model-based research should be dealt with an even greater care by reviewers and call upon mathematicians for a review of the logic and appropriateness of the model used. This is illustrated here by the use of an odds ratio (OR) taken as the corner stone of a circular reasoning based on a simple mathematical extrapolation without cross validation against reality and without appraisal regarding the corpus of knowledge amply developed and published on the topic addressed. This resulted in a blind and oversimplistic mathematical treatment of an issue of the foremost importance. A flawed model by design has yield results that can be viewed as a pure fabrication. The OR used is not an accurate representation of the true OR associated with the treatment involved. It was based on the result of a meta-analysis that included only RCT trials with OR exceeding 1. with the two trials with an over-dominating weight having used excessively high doses of HCQ: the Recovery and Who Solidarity trials with weight of 73.7 % and 15.2%, respectively. Recovery was conducted on frail patients hospitalized at a very late disease stage and having received sub-lethal doses of HCQ, very close to the lethal dose, dramatically higher than the safe dosage established by the pharmacokinetics properties of HCQ.

Background Compliance with research ethics legislation is crucial for protecting people's rights in clinical trials. Researchers can alert publishers when they have concerns about possible non-compliance with this legislation by other authors, and this opportunity for independent scrutiny is important. Nevertheless, some alerts may turn out to be unfounded due to a misunderstanding of the legislation. Methods We present the analysis of the article by F. Franck et al., « Raising concerns on questionable ethics approvals - a case study of 456 trials from the Institut Hospitalo-Universitaire Méditerranée Infection», in which the authors explain having sent numerous emails to different publishers to express their concerns about potential ethical breaches in publications from a single French institution. We analyzed in detail the 248 studies that the authors present as suspects. We have also studied the texts of French law. Results Our detailed analysis of these 248 studies discovered that only 8 of them report research involving humans. All the others are secondary research on existing biological samples. In France, as in many other countries, secondary research on biological samples does not involve human subjects. Conclusions Assessing the ethical conformity of scientific publications must be based on a precise knowledge of the legislation and a rigorous analysis of these publications. When researchers who alert publishers confuse research involving the human person with research not involving the human person, the publishers are in danger of being misled by these unfounded alerts. This can have deleterious consequences.