After a year and a half of a pandemic during which conflicting scientific opinions and controversial media figures overwhelmed the public and dented their confidence, the question of scientific integrity is central. While a new Australian study raises the question of the external pressures received by researchers from funders, the president of the ethics committee of Inserm Hervé Chneiweiss takes stock of the real sticking points of integrity scientist, at the dawn of a “new world“.
One in five researchers report external pressures
“One in five researchers said they were pressured to delay, change or not publish the results of trials of health behavior interventions“, reports an Australian publication in the newspaper PLoS ONE. To achieve this result, a survey received by pread by 200 scientists around the world. All of them participated in the work of the Cochrane review (which is the benchmark in medicine for its reviews of the literature) on nutrition, physical activity, sexual health, smoking and substance use.
These scientists had to indicate whether or not they had already been pressured by the organizations or companies funding their work, including the brake on the publication of unfavorable results, requests to alter conclusions or suppress certain results. As a result, of the 104 researchers who responded, 18% reported pressure received. “The most frequently reported deletion event was the reluctance of the funder to publish the ‘unfavorable’ results“, report the authors of this work.”Our overall estimate of 18% is similar to estimates from previous studies conducted in Australia (21% to 34%) and Canada (24%)“, they specify, adding that their result is probably underestimated.
–
Behind the pressures, government agencies rather than industries
Apparently surprisingly, manufacturers did not emerge as the major contributors to these pressures. “In comparison (government agencies, independent sources and multiple funding sources), researchers with industrial or philanthropic funding did not report a single case of suppression“, indeed specify the researchers.”It’s not surprising“, points out the researcher and president of the ethics committee of Inserm Hervé Chneiweiss.”Cochrane articles are scientific consensus reviews on public health issues. It is not very surprising not to find the industry there, which is more interested in original work.On the contrary, government authorities rely on these consensus reviews to support or justify policy measures.
What if the criticisms were valid?
Are government agencies therefore sometimes a threat to scientific integrity? Funders “may employ strategies that may undermine the integrity of science and government“, the Australian authors conclude. Yet there may not be much wrong with this situation.”This article does not say anything about the reasons why these deletions would have been requested: was it justified or not? Sometimes the criticisms are valid“, remarks Hervé Chneiweiss.
Because if in France the commissioning entity has no right of scrutiny over the works financed when it comes to a call for tenders, it may be otherwise for the requests for expertise. “At Inserm, we have a collective expertise structure, with various sponsors, such as the CNAM for example“, illustrates Hervé Chneiweiss. If the expert structure is completely independent, the report is”returned to the sponsor before being published, and it is possible that the sponsor will then make comments, not necessarily to exert pressure, but to point out unclear or poorly substantiated things“. Cochrane reviews are indeed meta-analyzes, that is to say they bring together all the work deemed relevant on the same subject to make a global analysis. However, this type of analysis is often “controversial over its methodology and results“, because it requires comparing clinical trials which have methods, durations, rigors or even numbers of patients that are sometimes very different, observes Hervé Chneiweiss.
Target the breaches of researchers rather than acts of fraud
For the chairman of Inserm’s ethics committee, maintaining and improving scientific integrity does not primarily lie in cases of fraud or external pressure. “There will always be crooks and frauds, even the research community remains human“, slice Hervé Chneiweiss. The crux of the problem seems much more to be a matter of training and method than of ill will. falsified day of data, against 34% (more than 2 million scientists) admitting to having used “questionable research practices“or”less serious breaches, which are rather bad research habits“, as defined a Senate report from March 2021. These bad practices relate in particular to “biased or erroneous methods, statistical tests applied despite common sense or data not fully identified or provided“, lists Hervé Chneiweiss.
Open science, the future of scientific integrity
For him, these shortcomings underline the importance of open science: the possibility for each member of the research community to be able to read and validate the work of his peers, even outside the traditional system of proofreading by the journal reading committee. scientists. “It is necessary to create training for researchers and engineering technicians, and to align with international FAIR standards, that is to say to make the data accessible and interoperable in terms of publication format to make them reusable. This is the great principle of open science.”
Open science is the new paradigm of the scientific world: while scientific journals (Nature, Science, The Lancet…) Were the main guarantors of the quality of the work available online, many pre-publication sites such as MedXriv or BioArchive now allow researchers to submit their work more quickly to their peers for comments. The health crisis has accelerated the movement, with thousands of pre-published publications validated or criticized collectively, faced with the urgent need to share work without waiting several months for them to be confidentially reread by established journals.
“For a long time, the quality of a researcher’s work was judged by the prestige of the journals in which he publishes, such as Nature, and its impact factor “, a quantified indicator of the number of times the article has been cited by other works, explains Hervé Chneiweiss. “In the open science model, we don’t have a ready-made solution to guarantee scientific integrity: biometric analyzes of the number of citations? Trip Advisor-style likes as readers reread the publication? There will have to be a new system. We are at the start of a new story.”
–
