Swiss research is not immune to the growing phenomenon of withdrawals of scientific articles – rts.ch

The number of retractions of scientific papers worldwide has followed an exponential curve since the early 2000s, according to the database Retraction Watchwhich lists these articles. In 2023, nearly 10,000 publications have been withdrawn, corrected or have been the subject of concerns regarding their scientific quality. An absolute record.

The retraction of a scientific paper is never trivial. In the scientific process, the publication of an article is considered an outcome (see box) and the research that relies on it before it is questioned is not necessarily corrected afterwards. An article on stem cells published in the prestigious journal Nature in 2002 and withdrawn this year alone has been cited nearly 4,500 times in the meantime.

Swiss research is not spared from this phenomenon, according to an analysis of the Retraction Watch database by the RTS. The number of retractions of research carried out by Swiss teams – including in international collaboration – follows a similar curve to the world average, for a total of 304 retractions since 1991. More than three-quarters of the retractions took place during the ten recent years. The year 2022 was the most important, with 32 searches affected.

There are several “types” of withdrawal. First, we find the articles simply removed. This occurs in almost three quarters of Swiss cases. Next come the articles which have undergone corrections after publication. Some, for their part, are the subject of concerns regarding their scientific quality. These two types of withdrawal each represent a little more than 10% of cases. Finally, three papers were “reinstated”, after having been temporarily withdrawn for issues including notices.

Lots of articles on health

No less than 143 retractions, or almost half of the corpus, concern research related to health or medicine. This does not surprise Guillaume Levrier, political science researcher specializing in biotechnologies and member of the NanoBubbles network, which tracks scientific errors and fraud.

“Research in the biomedical field is based on methods of administering scientific proof which are historically easy to distort and difficult to invalidate, short of gross falsifications,” the researcher begins by recalling.

“We only detect the falsifications of people who falsify badly. Modifying a graph in Photoshop will be seen quite easily. Modifying the algorithms that take data from one measuring instrument to falsify it before it is captured by another system to generate the graph is almost undetectable.”

Guillaume Levrier believes that the biomedical field is in reality not necessarily more affected than the others. “A lot of studies in other types of sciences, for example in the social sciences, are probably problematic. These are fields that publish less, have different methodological and epistemological practices and potentially cheat less badly.”

Plagiarism of John Paul II

The prize for the number of Swiss retractions goes to a researcher from the University of Lugano, with six papers withdrawn, seven corrected and one raising concerns. Most cases fall under plagiarism issues. A ticket they blog from Retraction Watch tells us that the researcher in question notably plagiarized Pope John Paul II in one of his papers.

The University of Lugano sanctioned its author by suspending him for one semester in 2017. He has since regained his place. Other articles have been questioned in the meantime, but the University of Lugano did not consider it necessary to sanction him again. The last retraction of one of his papers dates back to 2019.

The pressure to publish, omnipresent in the scientific world, can lead to fraud. “Either you are a hyper-publisher, you absolutely want to publish, or you are incapable of rising to the height of the required knowledge,” explained Michelle Bergadaà, professor of management at the University of Geneva, in the RTS show 36.9°. .

>> Read also: Scientific fraud, a growing scourge that threatens the integrity of research

However, not all errors are intentional and the majority of researchers follow an honest approach. But if the fraudsters are few in number, they are generally prolific, notes Solal Prielli, doctor in computer science, graduate of EPFL, who tracks scientific fraud in his free time. The majority of errors he discovers are the work of fraudsters.

Combination of reasons

It is therefore interesting to focus on the reasons for the retractions of Swiss productions, data which is also compiled by Retraction Watch. Many criteria are taken into account, ranging from simple error to ethical problem. In particular, there are 105 papers with proven errors (in the data, in the images, in the analysis) and 90 cases where the scientists’ work raises questions. There were also 28 cases of plagiarism as well as 18 falsifications.

Most of the time, the withdrawal is motivated by a combination of reasons. In this regard, we can cite a study on the recovery of patients who received ketamine during cardiac operations, published in 2019, and in which a team from Langenthal hospital (BE) participated. She was removed the same year for ethical violations, lack of authorization and improper conduct of the investigation.

The consequences for a researcher whose article has been retracted can be very diverse. It depends on the place it occupies in the study (much research is carried out in collaboration with a multitude of teams), the intentionality or not of the error and its seriousness. There is therefore no standard sanction and universities deal with these situations on a case-by-case basis.

Prestigious journals

Among the journals with the most retractions are big names, which have significant influence in their field of research. The famous journal Science, with 14 retractions, is thus the second journal to have published the most Swiss research which was subsequently retracted.

This does not particularly concern Solal Prielli. “I would be worried if Science did not withdraw of articles. It’s normal for errors to happen.” Lonni Besançon, a researcher in information visualization at Linköping University in Sweden, and also a specialist in scientific errors, takes a more critical look. “I’m obviously disappointed to see the number of problematic articles in even prestigious journals.”

“If journals pride themselves on having an excellent selection process and being prestigious, it is abnormal for poor quality or completely fraudulent studies to be published,” continues Lonni Besançon.

Protocols in question

Does the increase in retractions of Swiss work risk harming the credibility of Swiss research? “I want to say no,” dares Andreas Mortensen, associate vice-president for research at EPFL. According to him, the number of retractions does not reach worrying values ​​given the number of articles published. “I believe that this increase above all reflects an increase in rigor in terms of scientific publication.”

It would be wrong to consider Swiss research as less and less reliable, says Katharina Froom, member of the Chamber of Universities of swissuniversities and rector of the University of Fribourg. On the contrary, the increase in retractions can, according to her, be proof of a certain scientific health.

“Retractions can be seen as an example of how the scientific system works, as dictated by the principle ‘science is self-corrective’ [la science s’auto-corrige, ndlr]. In a certain way, retractions can play a role in the visibility of negative results.” Katharina Froom notes that the increase in publications observed in recent years also automatically leads to an increase in retractions.

We can nevertheless wonder whether the protocols for verifying results before publication should not be strengthened, in order to eliminate as many errors as possible. For Andreas Mortensen, this would be counterproductive: “the cost in work and slowdown would be much greater than the gain in rigor.”

For her part, Katharina Froom explains that the issue can only be resolved through international cooperation. “Swissuniversities is part of international initiatives, such as DORA or CoARA, which aim to broaden the evaluation criteria beyond quantitative indicators to include, among others, Open Science and societal impact.”

Antoine Schaub