Grace C. Bellinger,1 Abigail S. Baldridge,1 Luke V. Rasmussen,1 Oriana M. Fleming,1 Eric W. Whitley,1 Leah J. Welty1
To characterize clinical and translational science publications retracted for reasons related to the capture, management, or analysis of data to better understand errors that may occur in the research pipeline.
This scoping review complied with Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines and followed a preregistered protocol. The Retraction Watch database was queried through March 12, 2020.1,2 Records were eligible for abstract review if they were published between 2010 and 2020, had a subject list containing terms related to clinical and translational science, and were retracted because of concerns with the capture, management, or analysis of data. Abstracts were reviewed in duplicate. A retracted article was eligible for full-text review if abstract review determined it was published in English and related to clinical and translational sciences. During full-text review, the study team extracted information on number of authors, author attributions, data types and sources, study design, statistical analysis plan, software, and data availability. Research electronic data capture (REDCap) was used for import of publication information from Retraction Watch and data entry throughout the abstract and full-text review processes.3 A random sample of 5% of articles were reviewed in duplicate. Descriptive analyses were performed using R, version 4.0.1 (https://www.R-project.org/).
Of 21,252 records retrieved from Retraction Watch, 1266 (6%) were eligible for abstract review and 884 (4%) were eligible for full-text review. Of the 884 publications eligible for full-text review, 786 (89%) were available online through Northwestern University’s library system and included in the final analyses. The analytic set included 571 articles (73%) involving human research, 213 reports (27%) of animal research, and 47 systematic reviews or meta-analyses (6%). Few retracted articles described data that were publicly available (67 [9%]) or stated that data were available on request (21 [3%]); most articles contained no statements related to data availability or methods for data capture and management. More than one-third of the retracted articles (300 [38%]) did not specify the statistical analysis software used. The most-used programs were SPSS/PASW (229 [29%]) and GraphPad PRISM (86 [11%]). Statistical software such as Stata (43 [5%]), SAS (38 [5%]), or R/R Studio (29 [4%]) were infrequently reported.
This scoping review identified more than 800 articles in clinical and translational sciences retracted over 10 years for concerns related to data capture, management, or analysis. The results describe this cohort of retracted articles. Future work will include comparisons with a set of randomly selected publications that have not been retracted but are matched on journal and time frame. Authors can improve the rigor of scientific research by reporting software used and data availability. Publishers, editors, and peer reviewers can contribute to these improvements by advocating for widespread adoption of transparent documentation.
1. Baldridge AS, Welty LJ, Rasmussen L, Whitley E. Analysis of Retraction Watch Database for Retractions Due to Errors in the Clinical Science Reproducible Research Pipeline: A Scoping Review Protocol. DigitalHub, Galter Health Sciences Library & Learning Center; 2020. doi:10.18131/g3-pep3-ky27
2. Retraction Watch database. Center for Scientific Integrity. 2018. Accessed March 12, 2020. http://retractiondatabase.org/
3. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377-381. doi:10.1016/j.jbi.2008.08.010
1Feinberg School of Medicine, Northwestern University, Chicago, IL, USA, firstname.lastname@example.org
CONFLICT OF INTEREST DISCLOSURES
Research reported in this abstract was supported, in part, by the National Institutes of Health’s National Center for Advancing Translational Sciences (UL1TR001422).
Role of the Funder/Sponsor
The National Institutes of Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the abstract; and decision to submit the abstract for presentation.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
We thank the Retraction Watch team as well as Q. Eileen Wafford (Galter Health Sciences Library, Northwestern University).