Fitch S, Klaren WD, Payne L, Wikoff D. Comparison of public and private literature databases for toxicological investigations. Poster presented at Society of Toxicology Annual Meeting, San Diego, CA, March 2022.
Abstract
With the ever-increasing number of peer-reviewed publications, the ability to evaluate the current state of the science on a specific topic is a constant challenge. Multiple public and private databases are available with a variety of functionalities to aid in evaluating the published literature. That said, the inherent differences in the functioning of those databases may result in an omission of relevant publications. This work set out to compare the query outputs of different public and private databases to understand the degree of congruency and differences among them. The public database under consideration is NCBI’s PubMed (PM), while the private databases are Elsevier’s Embase (E) and EBSCO’s Academic Search Complete (ASC). Search queries were developed and utilized in all databases to evaluate how the databases handle broad search queries designed to result in many hits, narrow and specific search terms resulting in few hits, and evaluations of syntax complexity and search discretion were also investigated. The number of resulting publications were compared along with the difference in publications captured from the search. For select searches, the potential relevancy of the publications to the search terms was also assessed. A generic search for furfuryl alcohol and toxicity key words was developed for the three databases. In a query of the PM, E, and ASC databases, 228, 144, and 78 references were returned, respectively. Of these, 135 citations were reported by two or more databases, leaving 315 unique citations identified between the three. The preceding example, among others, indicates that the different databases query the search terms slightly differently, which could result in the omission or absence of a crucial publication in a literature evaluation if clear a priori search strategies are not employed. It also highlights the importance of search validation exercises and supplemental hand-searching, which are common in the evidence-based medicine community but have not been widely adopted among the toxicology community. In all, these findings further emphasize the importance of database selection or multiple database searches and especially search syntax when performing an evaluation of relevant peer-reviewed publications.