There’s good news and bad news in radiology research, according to a new study: The number of retractions is increasing in radiology journals, but the rate of retraction remains lower than that seen in biomedical journals outside the field of radiology.
According to the study in the American Journal of Roentgenology, between 1986 and 2001, radiology journals retracted — at the most — one paper per year, but from 2002 to 2013, at least two papers were pulled each year. Overall, roughly 11 articles are retracted out of every 100,000 articles published in radiology journals — compared to 15 out of 100,000 for biomedical journals outside radiology.
Still, writes author Andrew Rosenkrantz in “Retracted Publications Within Radiology Journals:”
every instance of a retracted publication represents an important concern, and the number of retractions annually for radiology journals has increased over the past decade.
Retractions are indeed increasing in all areas of science publishing. In 2013, Daniele Fanelli proposed that the increase may be attributable to more journals retracting articles, rather than just the same journals retracting more articles — a sign that more journals may be getting better at finding misconduct, rather than an increase in misconduct itself.
We asked Rosenkrantz, based at New York University, why he believes retractions are increasing in radiology journals, specifically. He told us:
There has been an increase in the frequency of retractions within the biomedical literature in general, I do not sense that this is unique to radiology. So what is observed may reflect this broader overall trend. Various reasons have been suggested by past authors to possibly explain this increase. This could in part relate to greater detection through use of the Internet, electronic search tools, etc.
The number of retractions during the time period in radiology journals was quite low — only 48 — but it’s still interesting to compare the retraction rate to that in other fields.
For instance, retraction rate in radiology journals — 0.01% — is lower than some other fields that have been considered individually, such as stem cell research, oncology or cell biology. However, higher rates in other disciplines may be deceptive, as in the case of orthopedics and anesthesiology, where a disproportionate number of retractions resulted from the misconduct of only a few individuals.
Rosenkrantz told us:
The intent is to compare the rate of retractions in radiology generals vs. the rate in the biomedical literature in general. It is likely that the rate in other specialties will be variable, with some being higher and others being lower. Indeed, such variation among other disciplines outside radiology remains unknown from this particular assessment. Some readers may still find the provided calculation helpful for having a sense of the comparison with the literature in general.
It took equally long to retract one of the 48 papers from radiology journals as (mean time 2.7 ± 2.8 years) it took the over 2000 non-radiology retractions from the PubMed database, suggesting that radiology journals face the same barriers to retracting a paper as other science journals. However, little other consistency could be found, leading Rosenkrantz to conclude in the paper that:
These concerns indicate the need for greater awareness and training in proper biomedical research conduct as well as for standardized policies and procedures among publishers for handling retractions.
Radiology journals as a subset tend to have comparatively lower impact factors (<7.2), and many studies have suggested a relationship between higher retraction rates and higher impact journals.
Rosenkrantz has published nearly 200 papers, on topics spanning from imaging lymphomas to the use of twitter by academic radiology departments. For this study, he chose a single day’s search on retractions found on PubMed, filtered by correlation with radiology journals categorized and indexed in Reuters’ Journal of Citation Reports®. The resulting study sample was small, although not unexpectedly so, as radiology is a narrow field of publishing.
When we asked Rosenkrantz about his sample size, he told us:
While the 48 retractions is not a large number, this does represent all retracted articles within radiology journals using the outlined approach.
Hat tip: Rolf Degen
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our new daily digest. Click here to review our Comments Policy.