No MeSH, No Filters, and No Librarians: Oh My! An Examination of the Search Strategies Used in 100 Published Systematic Reviews

Objectives: The increased popularity of systematic reviews has resulted in a wide variety of methods used to construct the literature search for these publications. Our project examines the components of 100 literature search strategies used in published systematic reviews to determine if specific key elements were included in the search strategy. To our knowledge, no research exists that gathers data on this specific topic.

Methods: We identified 100 citations that were indexed in PubMed and included the following criteria: the words “diabetes” and “systematic review” were in the title, and they had been tagged as a systematic review by MEDLINE indexers. Diabetes was chosen to allow a wide variety of clinical areas, including pharmaceutical, therapeutic and behavioral. We limited results to those published in 2019 and selected the first 100 results, sorted by relevance. We analyzed these reviews to determine how the literature search string was constructed, including whether MeSH terms were used, if the results were limited to the title/abstract fields, if guidelines or checklists such as PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) or MOOSE (Meta-analyses Of Observational Studies in Epidemiology) were purported to have been followed, if a reproducible search strategy was included, the average number of databases used, and whether a librarian was a co-author or acknowledged in some other way.

Results: 55% of these reviews included a reproducible search string (many provided just keywords, and some included no information about the literature search strategy). 48% of researchers used MeSH terms, and 23% used field codes to filter by title/abstract or text word. The average number of databases searched was 3.8, and 85% of reviews stated that they followed PRISMA or MOOSE guidelines or checklists. Only 11% of the reviews acknowledged librarian involvement. Of the reviews associated with a librarian, 73% used MeSH terms and 56% used field codes.

Conclusions: Our analysis of the components of 100 literature search strategies used in published, indexed systematic reviews shows that consistency is lacking. Our research does not attempt to recommend the elements to include in a search strategy, as some elements are appropriate for one review but may not be appropriate for others. Rather, using visuals such as charts and graphs, our poster illustrates how often the literature search strategies of 100 published systematic reviews used MeSH terms, field code searching, and other common key components of a literature search strategy.