An analysis of 20 years of retracted articles in psychiatry

AIMS AND METHOD: To characterise recent trends in retractions of publications in the psychiatric literature. We examined the Retraction Watch Database for publications with the subject term 'psychiatry' between 2003 and 2023. We then characterised publication type, subject, country of origin, reason for retraction and time between publication and retraction. RESULTS: A total of 315 psychiatry publications were retracted; the annual number and rate (relative to publications) rose steadily across the study period. The median number of days from publication to retraction was 643 (interquartile range 223.0-2966.0). Reasons for retraction were unclear in 20 (6.3%) cases only; most commonly they were attributed to data error (42.5%) or problems relating to peer review (11.4%) - the period between publication and retraction was particularly long for the latter studies (median 3082 days, interquartile range 575-3364). A total of 4.4% were retracted because of falsification, and 6.32% because of plagiarism. Retracted items frequently focused on mood disorders (24.4%), psychosis (14.9%) and neurodevelopmental disorders (13.7%). CLINICAL IMPLICATIONS: Clinicians and policy makers rely on academic literature, but should be aware of the ethical, legal, business and professional dilemmas often encountered, yet not reported, in the process of considering retractions of unreliable data. Contested and unreliable findings may persist in the scientific record, with potential negative consequences for evidence based-care and policy, including the erosion of public trust.