Abstract
Retraction of Systematic Reviews and Clinical Practice Guidelines
Ivan D. Florez,1,2 Alberto Henriquez,3,4 Andrés F. Estupinan-Bohorquez3,5
Objective
We described the influence of retracted systematic reviews and meta-analyses (SRMAs) on clinical practice guidelines (CPGs) and the characteristics of retractions in CPGs.
Design
This cross-sectional study was conducted in 2 stages based on searches focused on the Retraction Watch (RW) database and MEDLINE from inception to November 30, 2024. In the first stage, we included SRMAs. We described the reasons for retractions, and we recategorized them based on our assessment. We identified the CPGs that cited the SRMA in the Google Scholar database. In the second stage, we included the retracted CPGs, described the reasons for retractions, and recategorized them into ethical or nonethical reasons based on our assessment. Nonethical reasons were categorized as editorial or administrative or outdated guidelines, while ethical reasons were reported according to RW categories. We used descriptive statistics to summarize the findings.
Results
In the first stage, we included 377 SRMAs, of which 211 (56.0%) were retracted due to peer review or publication manipulation (eg, detected “fake” reviewers); 30 (8.0%), due to duplicate or redundant publication; and 136 (36.1%), due to intellectual or authorship disputes, plagiarism, outdated publication, retraction of included studies, methodologic or data errors, and conflicts of interests. For 49 (13.0%), specific reasons were not provided. Of the retracted SRMAs, 41 (10.9%) were cited in CPGs; 19 (46.3%) of these SRMAs were retracted due to research integrity issues and 12 (29.3%), due to data errors or being outdated. For 10 (24.4%), specific reasons were not provided. Most retractions were due to manipulation of the publication or peer review process. The median time between publication and retraction of the SRMA used in CPGs was 12.0 (IQR, 3.5-25.0) months, and the median number of SRMA citations was 40 (IQR, 22-191). In the second stage, we included 36 CPGs of the 138 potential CPGs identified. Nine CPGs (25.0%) were retracted because of ethical reasons and 22 (61.1%) for nonethical reasons; the rest had no available information. The most common ethical reasons were plagiarism, authorship or intellectual property disputes, lack of disclosure of conflicts of interest, and discrepancies between the content and the cited evidence. Among the 22 CPGs retracted for nonethical reasons, 11 were due to dual publication or incorrect citations and 9 were due to outdated recommendations. The median publication to retraction time was 10 (range, 3-96) months. All of these CPGs were cited after their retraction date, and in all cases, the citations were used to support the background of the research studies.
Conclusions
Retracted SRMAs have been informing CPGs, which provide recommendations in practice and policy. The retraction of CPGs has been neglected. The RW database should be revised according to the specificities of CPGs. The most concerning reasons are ethical. Retracted CPGs continue to be cited after their retraction, mainly to inform the background sections of articles.
1Department of Pediatrics, University of Antioquia, Medellín, Colombia, ivan.florez@udea.edu.co; 2School of Rehabilitation Science, McMaster University, Hamilton, Ontario, Canada; 3Universidad del Norte, Barranquilla, Colombia; 4Universidad Metropolitana, Barranquilla, Colombia; 5EPICLINICA SAS, Barranquilla, Colombia.
Conflict of Interest Disclosures
None reported.