Using Bibliometrics-aided Retrieval to Delineate the Field of Cardiovascular Research

A hybrid search strategy, using lexical and citation based methods, is presented in this paper as a robust method to delineate the broad field of cardiovascular research. Overall, this study aims to provide scientifically reliable and accurate data driven evidence about cardiovascular research by establishing a dataset of published research in this field. A workflow is presented that outlines the methods carried out to establish a core dataset based on a core set of journals, to identify and use search terms to detect a broader dataset, and then to apply measures of similarities between the citations of these two datasets to ensure relevance of the final dataset. The final core set of journals established comprises of 120 unique journals covered in Thomson Reuters Web of Science Core Collection (WoS) database including a total of 320,647 documents from 1991 to 2013. The search terms utilised include 107 cardio-specific terms that initially identify 1.8 million unique documents when searching the title, abstract and keywords. Upon application of the citation-based similarity measures the final combined dataset consists of 845,071 publications. Overall, establishing a relevant dataset of cardiovascular research means placing a greater emphasis on having a precise dataset, reducing recall in the process. Conference Topic Methods and techniques