A public health journal has retracted an article on unintentional pesticide poisonings a year after the authors enlisted a lawyer’s help to fight the decision.
Last year, we reported BMC Public Health had decided to retract the article, “The global distribution of acute unintentional pesticide poisoning: estimations based on a systematic review,” which appeared in December 2020. The article has been cited nearly 300 times, according to Clarivate’s Web of Science, including more than 100 since the journal told the authors it would be retracted.
The authors listed affiliations with the Pesticide Action Network, a collection of organizations opposed to pesticides. In their review, they declared unintentional pesticide poisoning “a problem that warrants immediate action.”
The retraction notice cites a letter to the editor from employees of pesticide manufacturer Bayer, and the trade organization CropLife International, which criticized the analysis. The authors stood by their findings in a response, stating the critics “do not seem to have understood our estimation method.”
The authors tried to overturn the journal’s decision. In July 2023, a lawyer representing them sent the editors a letter arguing that the retraction would be “inappropriate,” as we previously reported.
In September, a journal editor wrote to the authors:
I am sorry to inform you that we will be proceeding with the retraction of your article as we have lost confidence in the conclusions presented based on the original concerns raised in the Matters Arising article by Dunn et al. and on the assessment and advice received from our Editorial Board Members.
In a rebuttal, the authors wrote:
The envisaged retraction of our paper would be an unacceptable outcome of an unacceptable process. We have explained in detail in our rebuttals that the prevalences used in our extrapolations are not higher than annual prevalences and therefore no overestimation has taken place for this reason. We have repeatedly shown that even if the criticism were correct, it would have a negligible impact on the results of our study. A retraction would therefore be contrary to the journal’s own policy, which states that retractions are not appropriate when there is inconclusive evidence to support a retraction.
The retraction notice, published Wednesday, states:
The Editor has retracted this article because concerns were raised about the use of ‘ever’ prevalence of pesticide poisoning to represent annual frequency in the extrapolations by a reader and by Dunn et al. [1]. Expert assessment has confirmed the validity of this concern and also concluded that the assumption of annual exposure for countries where the time frame is not reported is unreliable. The Editor therefore no longer has confidence in the results and conclusions presented.
All authors disagree with this retraction.
Besides citing the source of the “concerns” and registering the authors’ disagreement, the notice matches the one the journal editor proposed to the authors last May.
We asked the publisher what held up the publication of the retraction. Natalie Pafitis, senior editor of BMC Public Health, said in a statement:
While we endeavour to complete our investigations as swiftly and efficiently as possible, there are a number of reasons why these processes can, on occasion, take longer than expected.
In this instance, after readers raised concerns in January 2023, we launched an investigation into the paper in line with COPE guidelines, with support from the Springer Nature Research Integrity Group. This process included a post-publication review by several experts in the subject matter of the paper, and extensive communication with the authors. After careful consideration of both the post-publication reviews and the authors’ responses, we concluded that the most appropriate action to take was retraction of this paper.
Another group of authors has complained about the journal’s process for retractions.
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
Monsanto doing what Monsanto does.
It does. However, that doesn’t mean that the criticisms are wrong. That has to be determined from the evidence.
I have read through the methodology and I have pretty serious concerns, though I am not an epidemiologist.
I read the journal article and it certainly seems problematic to take results that ask if someone has EVER had symptoms of pesticide poisoning and assume that is also the annual number. The authors claim “the prevalences used in our extrapolations are not higher than annual prevalences” but if they had access to the annual prevalence then surely they would just use that data instead. They’re trying to estimate the annual number, so this claim doesn’t make any sense to me.
In their rebuttal (found in appendix D to the letter from their attorney) they attempt to justify their use of lifetime data as annual data by pointing out that the specific questions are capturing specific events rather than cumulative impacts. For example, they cite a study that asked “whether any of 12 listed symptoms had ever been experienced within 48 h of using such pesticides”, highlighting the “within 48 h of using such pesticides” portion. However, the fact that an event occurred within a particular year in no way indicates that it occurred within the most recent year or in every year, and thus does not justify interpreting the results as annual amounts. They appear to either not understand the criticism or not understand the data; either is concerning and raises deeper questions about the rest of their analysis and assumptions.
They state that, “even if the criticism were correct, it would have a negligible impact on the results of our study.” If that is the case, I don’t understand why they don’t just remove those studies from the dataset and recalculate.
If the “negligible impact” is nonetheless meaningful, then just give a range: “…without the ‘ever’ data, the estimate is xxx million. If we include the ‘ever’ data, the estimate is yyy million. Recognizing there is a relatively high probability the ‘ever’ data is an overestimation in at least some cases, the true number is probably somewhere between xxx and yyy million.”
I’m not in academia or a research field, so maybe there’s something that prevents them from making that sort of correction?
Ultimately, I would tend to agree with the critics that the authors made some unreasonable assumptions that call their conclusion into question.
Retraction based on scientific misconduct is good retraction practice.
Retraction based on (allegedly) bad science is however a bad retraction practice.
Good or bad science should be up to the scientific community and readers to decide, not some arbitrary panel, especially when the article has already been published.
This is a horrible precedent and a bad sign for the future of science in general
I too am skeptical of retractions only because of bad science. However, here we seem to have a case of a conflict of interest combined with questionable (maybe bad, I haven’t checked it myself) science. In such a case retraction seems appropriate.
I still tend to think this would warrant a reply article rather than a retraction.
Note those asking for the retraction ALSO have a conflict of interest. And the original conflict of interest was never omitted (which would mean much more serious ethical concerns, thus adding more reason for a retraction).