Evaluation of a new method for librarian‐mediated literature searches for systematic reviews

Abstract

ObjectiveTo evaluate and validate the time of completion and results of a new method of searching for systematic reviews, the exhaustive search method (ESM), using a pragmatic comparison.Methods: Single-line search strategies were prepared in a text document. Term completeness was ensured with a novel optimization technique. Macros in MS Word converted the syntaxes between databases and interfaces almost automatically. We compared search characteristics, such as number of search terms and databases, and outcomes, such as number of included and retrieved references and precision, from ESM searches and other Dutch academic hospitals identified by searching PubMed for systematic reviews published between 2014 and 2016. We compared time to perform the ESM with a secondary comparator of recorded search times from published literature and contact with authors to acquire unpublished data.Results: We identified 73 published Erasmus MC systematic reviews and 258 published by other Dutch academic hospitals meeting our criteria. We pooled search time data from 204 other systematic reviews. The ESM searches differed by using 2 times more databases, retrieving 44% more references, including 20% more studies in the final systematic review, but the time needed for the search was 8% of that of the control group. Similarities between methods include precision and the number of search terms. Conclusions:The evaluated similarities and differences suggest that the ESM is a highly efficient way to locate more references meeting the specified selection criteria in systematic reviews than traditional search methods.Further prospective research is required.KEYWORDS bibliographic databases, information storage and retrieval, quality improvement, review literature asThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Translating the Cochrane EMBASE RCT filter from the Ovid interface to Embase.com: a case study

“…Research has shown that translators should take care because interface differences can impact on retrieval (Damarell et al, 2013;Neyt & Chalon, 2013). Authors have also provided algorithms (macros) to automate the conversion of search strategies between interfaces (Bramer, Rethlefsen, Mast & Kleijnen, 2017). However, despite this information, we suspect that the impact and performance of such translations are rarely investigated in-depth because simple conversion is already a time-consuming process (Bramer et al, 2017).…”

State of the art, state of the science?

“…While not yet at industrial levels of production, the papers in this Special Issue, dominated by multi‐author and multidisciplinary work, attest to the demise of the information retrieval “cottage industry”. Collaborations span institutions and the academic‐commercial divide . Academic review organizations (eg, ScHARR, PenTag, and YHEC) and health technology agencies (eg, CADTH) and networks (such as Cochrane , and the Medical Library Association collaboration) continue to play a critical part.…”

“…This accurately captures the prevailing zeitgeist for systematic review efficiency and rapid review methods. Efficiencies are claimed for search strategy construction, randomized controlled trial (RCT) screening, and for searching for protocols, and all three authorial teams provide empirical evidence to support their claims. Such efficiencies must be accompanied by robust quality assurance procedures; for one of which, peer review of search strategies, we now have important supplemental evidence of benefit …”

“…A secondary theme is “working smarter”, whether by sharpening the efficiency of search protocols or by harnessing the potential offered by technology . As the guide to machine learning illustrates, technologies may be harnessed in the screening of specific study types—initially, RCTs but ultimately a wider range of study types.…”

Innovation in information retrieval methods for evidence synthesis studies

“…Two contributions address two of the most important overarching issues for information retrieval in evidence synthesis, namely, the efficiency and quality of our practice. Bramer and colleagues consider innovative ways of interrogating traditional bibliographic databases by evaluating librarian‐mediated single‐line search strategies in terms of precision and time taken to search …”