A recent report by Charles Piller published in the Science Magazine highlights the inexcusable responsibility of journal editors, peer-reviewers, principal investigators, postdoctoral fellows, and graduate students, to recognize the high incidence of citing retractions in the scientific literature. This recognition first merits a bit of clarification between two similar but different concepts: a literature review (usually done before we carry out research in our laboratories) and a reference check. So, what is the difference between a literature review and a reference check (also called literature update)?
A literature review is a systematic and reproducible approach use to define a research question and taking the necessary steps to look for, select, and retrieve the appropriate type and number of papers that are relevant to the research question. On the other hand, a reference check is a somewhat easier and more precise protocol to make sure we are citing accurately all the references in a manuscript; before and after submission for publication. Literature updates are intended to make sure that we accurately cite every reference included in our manuscripts (titles; year of publication; volume, issue and page numbers; links; DOI; etc). In this regard, the Science Magazine [Science 371 (6527), 331-332] mentioned an interesting example illustrating how prevalent is the habit of citing retracted references. This is a case related to a number of papers on COVID-19 research.
According to this report, early last year the scientific community was informed of two retractions of papers published in two important (high profile) medical journals, namely The New England Journal of Medicine, and The Lancet. These retractions were broadly advertised not only in scholarly groups but also on several social media platforms like Twitter. However, despite the scandal, the scientific community at large failed to acknowledge these retractions and, regrettably, continued to use and cite them.
An examination of more than 200 COVID-19-related papers published in 2020, Science found that 52.5% of them cited at least one of the two retracted papers mentioned above. This is highly significant for two reasons: (1) scientists and peer-review systems seem to be ineffective to prevent this issue, and (2) the mistakes that caused the retractions continued to be in circulation for several more months, spreading inaccurate information and opinions in a similar way to the spread of the virus they were studying.
In the same article, Piller's report also mentioned some reports in which previous authors described the use of a clinical database produced by the company Surgisphere, in one "influential preprint" published in April 2020. This database was the same one used by one of the retracted papers (the one describing the antiparasitic drug ivermectin as beneficial in critical COVID-19 cases). Therefore, it should be clear that whenever a research group fails to identify retractions, that group is responsible for disseminating data and conclusions based on unreliable sources. Besides, the issue of preprints is a difficult one because these publications are very often kept in circulation even after the paper is published. In this regard, the report in Science also mentioned a very useful web-based resource to spot retractions called "Retraction Watch", which is co-founded by Ivan Oransky.
Nevertheless, it has become increasingly common for trainees to adopt a quick and simple copy-and-paste information from lists of papers that often include review papers (this is why I made a comparison at the beginning of this blog). Many trainees (and a few supervisors) do this uncritical use of the information without reviewing the original papers they are citing in their work, and to make matters worse, this practice is often overlooked by Journal editors (regardless of the Journal's reputation or impact factor), as well as poorly trained peer-reviewers. I reserve any comments on the potential role of bias and conflicts of interest when it comes to paper reviews.
Consequently, the scope and importance of the Science report focusing on COVID-related references go far beyond this topic, suggesting a much bigger trend affecting most (if not all) topics in research. Academic groups at least should be aware of this problem and act decisively by providing appropriate training to graduate students and postdoctoral fellows, who will later become independent researchers. One of the main responsibilities of research supervisors must be to equip their trainees with the knowledge, skills, and attitudes needed to adopt high ethical standards at every stage of their work, including the objective and unbiased evaluation of scientific papers.
Because papers retractions necessarily involve one more step in the process of publishing papers, I submit that scientific journals should have a policy in place in which they describe how they address this issue, not just when they decide to retract a paper, but also how they follow up on subsequent papers citing those retractions.
If, as scientists, we are unable to correct this concerning trend and we continue to ignore paper retractions as an issue that needs to be addressed, it is quite possible that we may erode the much needed public trust in our work. We need public trust now more than ever amid the other growing and equally dangerous epidemic, the one spreading half-baked opinions or blunt fake news about important topics all over the internet via social platforms.
Finally, the take-home messages for the readers of this blog are:
(1) To recognize the serious implications of a deficient or failed approach to identify potential retractions of references used in our work; and
(2) There are resources available to trainees to track the evolution of published papers, and to acquire or improve their writing and reviewing skills. Check them out and use them!
Comentarios