Reconsidering the Need for De-Duplication Prior to Screening in Systematic Reviews

Abstract

Nowadays, systematic reviews play a crucial role in scientific research as an efficient method for summarizing the literature related to a specific topic. Systematic reviews rely mainly on searching several databases with a suitable search strategy to make sure all relevant articles are covered. However, due to the overlapping of databases and the fact that many journals may be indexed in several databases, many retrieved articles from the search may be duplicated. As a result, removing duplicates before the screening stage can save time and effort that may result from screening the same article multiple times. Although automated tools are widely available to detect duplicates, they typically do not achieve 100% specificity, which may result in relevant references being mistakenly removed as duplicates