Print this page

Published scientific articles should be read nowadays with skepticism, and the preprints even more...

 The COVID-19 pandemic has affected human activities on a scale never seen before. In the scientific world it has been particularly hard: people involved in the activity of generating new knowledge have been under both huge demand and maximum expectation of proposing solutions to stop the spread of SARS-CoV-2. Under these exceptional circumstances, the research community has sought alternatives for the rapid dissemination of their work. An example of such alternative is the “fast track publication”, which was proposed by WHO during the epidemics caused by the Ebola and Zika viruses in 2015-16 (1). Many funding agencies and journals, including Memórias do Instituto Oswaldo Cruz (2), joined this initiative to offer a rapid dissemination of scientific evidence on these infectious agents (3).

Regardless of the submission system, e. g. read first - peer review later (fast track) or peer review first - read later (conventional publication), all articles submitted to a journal must get the approval of two or more external peers to reach the status of "published article" and enter the scientific record. It is up to the journal editor to decide whether the reviewing procedure has been adequate or not. Currently, the public has no means to assess the editor's judgment for most of the published articles. Disclosure of the conversation between editor and author(s) does not happen even when someone reports (or whistle-blowing) on problems in an article on public domain!

Although the ''fast track'' remains an interesting alternative to quickly disseminate scientific results, the covid-19 pandemic has brought the preprint repositories into the spotlight. The urgent need for information about the Sars-Cov-2 has consolidated the role of preprint repositories as an adequate vehicle for the immediate release of recently obtained research data. The pandemic also proved to be a testing ground for preprint repositories, as some of the uploaded articles have been shown to contain serious infringement of publication ethics . These articles have been withdrawn (4) (https://retractionwatch.com/retracted-coronavirus-covid-19-papers/).

bioRxiv, the first preprint repository for the biological and biomedical sciences, was launched in 2013 but the academic community did not immediately embrace it. This lack of enthusiasm notwithstanding, new preprint repositories have been created in the recent years, e. g., medRxiv, agriRxiv, chemRxiv. In 2020, we witnessed the launch of the first preprint repository in Brazil, the Scielo Preprints (http://preprints.scielo.org).

In the short term the communication through preprints will likely be the first option in case a scientist wishes to receive a quick feedback on her/his new scientific data. Yet this initiative poses two fundamental questions:
1) Is the content of the preprint articles, e. g. methods, data, analyses, images, graphics, conclusions and citations, reliable?
2) How should we absorb this information if we no longer have the “gates” of editors and reviewers (as long as they remain in the repository without being sent to publication with peer reviewing)?

Answering these questions is not trivial! They deal with the essence of the collective endeavor we call "Science": the gathering of evidence with rigor, quality and creativity, the credibility of results and conclusions, and the proper communication to the public. Possible answers should include the evaluation system that scientific journals currently use, e. g. the external peer review, which should be modified to work in an iterative and continuously updated environment. It cannot simply be the use of peer review as it is currently practiced, that is, a kind of consulting mechanism to say "yes" or "no" to authors. According to the news of "retraction notice" released every week by scientific journals (see the retraction watch reports https://retractionwatch.com/) the current peer review system appears to be ineffective.

Whatever the mechanism to assess the quality and veracity of articles in preprint, it must make use of permanent skepticism, the objective questioning whether the data/analysis/conclusions are exactly as described in the article. In other words, the reader of a preprint should always be skeptical, and ideally he/her should do at least one of following actions: a) to replicate the results (if possible); b) to reanalyze them (if the original data are publicly available); c) to consult and discuss with other experts in the field. In the biological/biomedical sciences, replication of published research work is, in most cases, difficult to implement because it is expensive, time-consuming and without any recognition. There are no formal incentives for researcher groups to replicate the content of a published article or to verify the validity of its conclusions. On an individual basis, however, the act of "critical reading" can be done: maximize the skepticism and always question the authors about inconsistencies contained in methods, data and analysis described in the article.

If all the people who read a preprint were able to elaborate technical questions about the described data and their analysis we would have something very close to an ideal peer review. This ideal presumes that many people are questioning and demanding answers from the authors about the research work, in an “iterative” process of suggestions, error detection, corrections, open debate, and quick update. This is quite different from the current publishing practice: an editor and two or three reviewers anonymously (in most journals) decide which scientific evidence is valid (i. e., is publishable) or not.

For the time being the landscape of preprint repositories is far from including the 'ideal peer review', or any peer review at all. Most notably many people in the scientific community have not yet realized the disruptive potential of publishing their scientific work in a repository under the assumption of an "open, immediate and iterative" peer review!

The immediate release of scientific information has opened a new chapter on the relationship among scientists, funding agencies, publishers and the general public. There is an expectation that it will strengthen some values of the contemporaneous science such as open access, transparency, quality, efficiency, integrity, rigor and credibility.

This landscape is under construction...and it is only the beginning!

Adeilton Brandão

REFERENCES

1. https://wellcome.org/news/sharing-data-during-zika-and-other-global-health-emergencies. Accessed November 27, 2021
2. Pirmez, Claude, Brandão, Adeilton Alves and Momen, Hooman. Emerging infectious disease and fast-track publication: when public health gets priority over the formality of scholarly publishing. Memórias do Instituto Oswaldo Cruz. 2016, v. 111, n. 5. https://doi.org/10.1590/0074-02760160001.
3. Johansson MA, Reich NG, Meyers LA, Lipsitch M (2018) Preprints: An underutilized mechanism to accelerate outbreak science. PLoS Med 15(4): e1002549. https://doi.org/10.1371/journal.pmed.1002549
4. https://retractionwatch.com/retracted-coronavirus-covid-19-papers/. Accessed November 27, 2021