The biggest reason for biomedical research retractions

Detection software is not enough.

18 October 2019

dane_mark/Getty

A new analysis has identified plagiarism as the most common cause of retractions for biomedical research papers in both high- and low-impact journals.

The number of retractions due to plagiarism is also on the rise, despite the use of software to detect copied work in manuscripts, the analysis found.

Isabel Campos-Varela, a hepatologist at the University of Santiago de Compostela in Spain, and her colleagues examined more than 4.3 million biomedical papers that were indexed on PubMed between 2013 and 2016.

They classified the reasons for retraction into nine categories, including plagiarism, data manipulation and conflict of interest. The team then grouped these categorized papers under misconduct, suspicion of misconduct, or no misconduct.

The results were published last month in the journal Gaceta Sanitaria.

Campos-Varela’s team found that, for every 10,000 papers, 2.5 papers were retracted, and the retractions were shared among 611 biomedical journals. The number of retractions increased over the three-year period from 70 in 2013 to 418 in 2016.

Around half of these retracted papers were published in journals with a high impact factor, and 33% were published in low-impact journals. Fifteen per cent were published in journals with no impact factor.

Close to 10% of retracted papers were in biochemistry and molecular biology – the highest proportion for all fields – followed by oncology, multidisciplinary sciences and experimental medicine.

While clear misconduct was detected in 65% of all retracted papers included in the analysis, it was more frequent in low-impact journals than higher profile titles.

Pathology journals had the highest proportion of retractions due to misconduct, followed by general medicine, oncology and experimental medicine publications.

Journals specializing in pharmacology and pharmacy research had the least misconduct-related retractions at around 48%.

Data manipulation was the second most common reason for retracting a paper due to misconduct, regardless of the journal's impact factor.

The authors said their findings highlight the need for journals to take greater measures to check for plagiarism. Such measures could include using effective detection software before an article is assigned to an editor and implementing clearer editorial guidelines.