When selecting and assessing studies to include in a review, be very aware of the possibility of retracted articles being included in your results.
"Retraction is a mechanism for correcting the literature and alerting readers to articles that contain such seriously flawed or erroneous content or data that their findings and conclusions cannot be relied upon." (COPE Council, 2019, p. 3).
Articles can be retracted for a number of reasons, such as:
According to Retraction Watch founder Ivan Oransky (2022), the number of retracted articles has increased from 45 a month in 2010 to around 300 a month in 2021, although given the increased number of articles published, this still only equates to under 0.1% of articles published.
So if it is only a relatively small portion of articles that are retracted in the end, why do we need to worry about it when conducting reviews?
Evidence shows that retracted studies are still cited by other researchers, with only a very small proportion of the post-retraction citations noting the retracted status of the original article (Hsiao & Schneider, 2021; Theis-Mahon & Bakker, 2020). One study even found a clinical trial which used false data was continuing to be cited eleven years later in order to support a medical intervention (Schneider et al., 2020).
When retracted articles are included in evidence syntheses such as systematic reviews, meta-analysis or practice guidelines, the results and subsequent recommendations may be inaccurate. In a recent example, a meta-analysis found ivermectin was beneficial in treating COVID-19, however after several of the included studies were retracted due to plagiarism, ethical concerns and inaccurate data collection, a reanalysis of the data that excluded those studies found ivermectin had no benefit for treating COVID-19 (Hill et al., 2022).
It can be challenging to identify when an article has been retracted, owing to inconsistent practices across databases. A recent study by Suelzer et al. (2021) found one database had no retraction label or links to a retraction notice applied to 23 retracted journal articles, and a quarter of retracted articles on publisher websites followed five or less of the recommendations for retracted articles by the International Committee of Medical Journal Editors.
Some suggested methods for avoiding including retracted articles in reviews are: