Continued Citation of Retracted Radiation Oncology Literature—Do We Have a Problem?

Abstract

PURPOSE: The purpose of this study was to quantify the number and explore the nature of citations of retracted articles in the radiation oncology literature occurring after publication of the retraction note. METHODS AND MATERIALS: A list of previously identified retracted articles within the field of radiation oncology was updated in June 2017. The number of publications citing retracted articles was quantified using Google Scholar, Web of Science, and Scopus. Studies citing retracted publications after publication of the retraction note were individually assessed to determine the nature of the reference. RESULTS: Forty-seven retracted articles were identified for this study. Thirty-seven studies (79%) received 1017 citations before retraction, and 34 articles (72%) received 407 English and 169 foreign-language citations afterward. The average number of citations dropped from 22 to 11 after publication of the retraction note (95% confidence interval, 0.3-21.0, P = .043). Most postretraction citations occurred during the second year after the article's retraction, originated from North America and Asia (N = 124 each, 31%) and Europe (N = 122, 30%), and featured in original articles (N = 254, 62%) and reviews (N = 73, 18%). Of the 358 individually evaluable citing studies, 92% referenced the retracted article as legitimate work. Three guidelines and 15 systematic reviews and meta-analyses were also identified that cited retracted articles as valid work. CONCLUSIONS: Postretraction citations are an avoidable phenomenon. The results of the study emphasize the need for investigators to adhere to good research practices to mitigate the influence and propagation of flawed and unethical research. Journal editors, peer reviewers, and the broader radiation oncology readership should remain diligent in ensuring that citations of retracted work are identified and removed before, during, and possibly even after publication.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 42 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.