In a recent research letter published in JAMA Network Open, a team of researchers compared the author characteristics and causes for manuscript retractions of coronavirus disease 2019 (COVID-19) and non-COVID-19 related research.
Background
Published papers from peer-reviewed journals are generally withdrawn or retracted because of major errors, fraudulent results, or other academic misconduct. Statistics suggest that the frequency of retractions has increased in recent times.
Since the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, COVID-19 research has also seen many retractions, possibly due to increased investigation into the subject. However, it poses questions about whether retractions of COVID-19-related research and other medical studies are similar in the reasons for withdrawal or the author’s characteristics.
About the study
In the present cross-sectional study, the researchers used the Retraction Watch Database (RWDB) to identify all retracted publications and preprints on COVID-19 and non-COVID-19-related research since February 1, 2020.
Information about the withdrawn studies, such as type of publication, author affiliations, the impact factor of the journal in 2020, and the date of retraction, was recorded. The studies were grouped according to the prominent causes of retraction. Author profiles and affiliations were verified using Scopus and Google Scholar. The years since the first publication and the H-indices of the first and last authors were compared for the two categories. The H-index is a metric indicating the productivity and citation impact of the publications from an author.
Statistical analyses to compare the various characteristics of the two retraction categories in this study included the Mann-Whitney U and Fisher Exact tests.
Results
The results indicated significant differences in the reasons, timelines, and study designs between retracted COVID-19 studies and withdrawn articles from other medical fields. The authors found 138 retracted studies on COVID-19 and 380 article retractions on non-COVID-19-related research. Of the 138 COVID-19 studies, 101 were published papers, and 37 were preprints. In comparison, all 380 non-COVID-19 retractions were published studies.
The study revealed that COVID-19-related articles were more likely to be withdrawn in the first six months from publication than non-COVID-19-related medical articles (81.9% vs 59.7%). The study design and retraction reasons also varied significantly between the two categories. COVID-19 studies were withdrawn mostly for non-misconduct reasons, with a large proportion being modeling studies. Furthermore, compared to COVID-19 studies, more non-COVID-19-related article retractions were initiated by journals, editors, or publishers.
The first and last authors of the COVID-19-related retracted articles had a longer publication history and a higher median H-index than those of withdrawn non-COVID-19 publications.
The authors noted that while most of the retracted COVID-19 studies were not associated with academic fraud or misconduct, many of the retraction notices did not provide detailed explanations for the cause of withdrawal. Moreover, the retraction notices were also not uniform across publishers and journals, making it difficult to decipher the reasons for the withdrawal.
Conclusions
To summarize, this cross-sectional study compared retracted COVID-19 and non-COVID-19 related studies to understand similarities or differences in reasons for retraction and author characteristics. The findings suggested that many COVID-19-related studies were withdrawn for reasons unrelated to academic misconduct, but not all retractions were accompanied by detailed explanations on the RWDB.
Furthermore, a large portion of the examined COVID-19-related article withdrawals occurred in the six months following publication, indicating that research on the SARS-CoV-2 pandemic was under increased scrutiny.
Notably, most of the non-COVID-19 related medical studies were retracted by journals, publishers, or editors, and not the authors.
Journal reference:
- Shi, X., Abritis, A., Patel, R. P., Grewal, M., Oransky, I., Ross, J. S., & Wallach, J. D. (2022). Characteristics of Retracted Research Articles About COVID-19 vs Other Topics. JAMA Network Open. doi: https://doi.org/10.1001/jamanetworkopen.2022.34585 https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2797029