Retracted articles in the obstetrics literature: lessons from the past to change the future

BACKGROUND: The publication of invalid scientific findings may have profound implications on medical practice. As the incidence of article retractions has increased over the last 2 decades, organizations have formed, including Retraction Watch, to improve the transparency of scientific publishing. At present, the incidence of article retraction in the obstetrics and maternal-fetal medicine literature is unclear. OBJECTIVE: This study aimed to determine the number of retracted articles within the obstetrics and maternal-fetal medicine literature from the PubMed and Retraction Watch databases and examine reasons for retraction. STUDY DESIGN: A retrospective review of the PubMed and Retraction Watch databases was performed to identify retracted articles in the obstetrics and maternal-fetal medicine literature from indexation through December 31, 2019. The primary outcome was defined as the number of identified articles and reason for retraction. Within PubMed, articles were identified using a medical subheading search for articles categorized as withdrawn or retracted. In addition, the Retraction Watch database was queried and non obstetrical articles were excluded. The reason for retraction was classified according to the categories listed in Retraction Watch. The subject matter was classified on the basis of the Society for Maternal-Fetal Medicine criteria. Data were collected from retracted articles for author name, country, journal name and impact factor, year of publication and retraction, study type, and response of the publishing journal. Descriptive statistics were performed. RESULTS: Of the 519 obstetrics and gynecology articles in Retraction Watch, 122 (23.5%) were specific to the obstetrics and maternal-fetal medicine specialties. In addition, 39 (32.0%) were identified from PubMed, all of which were included in Retraction Watch. There was a median time to retraction of 1 (range, 0-17) year, with a median of 3 citations per article (range, 0-145). In addition, the median journal impact factor was 2.2 (range, 0.1-27.6), with median first and senior author Hirsch index values of 6.0 and 13.5, respectively. Most articles were original research (n=80; 65.6%), specifically retrospective studies (n=11; 9.0%), case reports (n=19; 15.6%), prospective studies (n=18; 14.8%), randomized controlled trials (n=11; 9%), basic science (n=18; 14.8%), and systematic review or meta-analysis (n=3; 2.5%). Of eligible articles, 32 (26.2%) were published in journals with an impact factor >= 4, and 21 articles (17.2%) were published in the top 10 leading impact factor obstetrics and gynecology journals. Most retractions were for content-related issues (n=87; 71.3%), including 21.3% (n=26) for article duplication, 18.9% (n=23) for plagiarism, and 16.4% (n=20) for errors in results or methods. Additional reasons included author misconduct (n=12; 9.8%), nonreproducible results (n=11; 9.0%), and falsification (n=8; 6.6%). The most common journal response was an issued statement of retraction (n=82; 67.2%). Lack of retraction notice and limited to no information provided by the publishing journal occurred in 19 retracted articles (15.6%). CONCLUSION: In the obstetrics and maternal-fetal medicine literature, retraction of scientific articles is increasing and is most often related to scientific misconduct, including article duplication and plagiarism. Improved prevention and detection are warranted by journals and healthcare institutions to ensure that invalid findings are not perpetuated in the medical literature, thereby avoiding adverse consequences for maternal and perinatal care.