E. G. King1, I. Oransky2, T. E. Sachs1, D. B. Flynn3, A. Farber1, J. A. Kalish1, B. M. Krafcik1, J. J. Siracuse1 1Boston University School Of Medicine,Department Of Surgery,Boston, MA, USA 2The Center For Scientific Integrity And New York University Arthur L. Carter Institute,,New York, NY, USA 3Boston University,Alumni Medical Library,Boston, MA, USA
Introduction: Retractions of scientific articles represent attempts to correct the literature. Our goal was to examine characteristics of retracted surgery-related papers.
Methods: NCBI Pub-Med database was queried using ‘surgery,’ ‘surg,’ or ‘surgical’ and ‘retracted’ or ‘retraction.’ Unrelated and non-retracted articles were excluded. Article, journal, and retraction details were recorded.
Results: There were 186 retracted surgical papers identified from 1991-2015. Average time to retraction was 3.6 years. Retracted papers were most often seen in general (25.8%), cardiac (22%), and orthopedic (10.2%) surgery. Articles were related to anesthesia and basic science in 32% and 11% cases, respectively. Authors were most commonly from the United States (21.5%), Germany (19%) and Japan (16%). Median impact factor (IF) of retracted papers was 2.4 (0.27-12.48).
Retraction reasons were 1) content related: duplication (34%), falsified data (15%), data errors (9%), plagiarism (8%), and 2) administrative: IRB concerns (19%), author dispute (9%), copyright (2%). No reason was given in 8% and 24% were behind a paywall (requiring payment to view notice). Median IF was higher for administrative than content-related reasons (3.42 vs. 2.12, P<.01). Plastic surgery, transplant surgery, basic science, non-anesthesia content, and publisher were predictive of a paywall for the retraction (P<.01). IF, content, and publisher did not predict whether a reason for the retraction was given.
Conclusion: The majority of surgical retraction notices have a reason listed and do not require payment to read. These findings provide a baseline analysis for surgeons to identify reasons for retraction and to focus on areas for improvement in transparency.