Mapping the Methodologies Used in Retraction Reviews

Abstract

Introduction: The impact of retracted publications has been studied across disciplines using various review methods often with limited search documentation, making replication and validation difficult. Retraction studies have been explored in some biomedical disciplines but not all. This project aims to map the methodologies used by retraction reviews in health sciences, as well as identify best practice and subject gaps. Methods: We searched Medline, Embase, and CINAHL using a mix of controlled vocabulary and keywords to capture reviews of retracted publications. We used Covidence to screen titles and abstracts through full-text review screening. Articles synthesizing or reviewing retracted publications were included while individual retraction studies were excluded. Data extraction focused on search methodology, discipline/subject, reason for retraction, and data management/analysis. Results: After removing duplicates, 3343 results remained for title/abstract screening. We conducted full-text screening of 83 articles and selected 54 studies for data extraction. Preliminary results show low use of reproducible methods. Several studies offer a minimum level of reporting, indicating only the name of databases searched and keywords used. Discussion: Librarians have expertise documenting searches in other contexts (i.e. systematic reviews) and this pre-existing knowledge will contribute to developing reproducible search strategies and methods in retraction studies. This paper maps the existing literature and methods used in reviews of retracted publications. It provides both librarians and biomedical researchers with knowledge on the gaps in methodology and insights into which biomedical disciplines are lacking retraction reviews.