Frequency and Effects of Search Strategy Characteristics on Relevant Article Retrieval in Systematic Reviews

Abstract

Objectives To identify common characteristics of highly effective search strategies for a clinical systematic review topic.

Methods Before attending an in-person systematic review workshop, participants asked to draft a reproducible search strategy based on a brief scenario and a research question from a published systematic review related to blood transfusion and radical prostatectomy. Participants are provided with three studies that were included in the published systematic review, but are not given the systematic review itself. The scenario proposes three commonly-requested limits: date range, inclusion of specific outcome, and human studies that participants can choose to apply or not. The submitted strategies are evaluated for reproducibility and effectiveness of retrieval of the 10 studies included in the published systematic review. Strategies were considered ”highly successful” if they returned all 10 included studies. We conducted a thematic analysis on the 14 highly successful strategies to identify common characteristics between them that could guide future searchers. Two studies were disproportionately missed by the other 98 search strategies, and their PubMed records were analyzed to identify what made them particularly challenging to find.

We thank the MLA Research Training Institute for its training, support and encouragement to carry out this research. This project was made possible in part by the Institute of Museum and Library Services (RE-95-17- 0025-17).

This study is IRB Exempt under HUM00128315.