6 Abstract: This study compares the results of data collected from a longitudinal query analysis of 7 the MEDLINE database hosted on multiple platforms that include PubMed, EBSCOHost, Ovid, 8 ProQuest, and Web of Science in order to identify variations among the search results on the 9 platforms after controlling for search query syntax. We devised twenty-nine sets of search 10 queries comprised of five queries per set to search against the five MEDLINE database 11 platforms. We ran our queries monthly for a year and collected search result count data to 12 observe changes. We found that search results vary considerably depending on MEDLINE 13 platform, both within sets and across time. The variation is due to trends in scholarly publication 14 that include publishing online first versus publishing in journal issues, which leads to metadata 15 differences in the bibliographic record; to differences in the level of specificity among search 16 fields provided by the platforms; to database integrity issues that lead to large fluctuations in 17 monthly search results based on the same query; and to database currency issues that arise due to 18 when each platform updates its MEDLINE file. Specific bibliographic databases, like PubMed 19 and MEDLINE, are used to inform clinical decision-making, create systematic reviews, and 20 construct knowledge bases for clinical decision support systems. Since they serve as essential 21 information retrieval and discovery tools that help identify and collect research data and are used 22 in a broad range of fields and as the basis of multiple research designs, this study should help 23 clinicians, researcher, librarians, informationalists, and others understand how these platforms 24 differ and inform future work in their standardization.
Methodological Issues with Search in MEDLINE: A 26 Longitudinal Query Analysis
Introduction28 Bibliographic databases are used to identify and collect research data, and therefore function as 29 scientific instruments [1,2]. Studies that rely on these instruments include research on 30 information literacy, bibliometrics/scientometrics, information seeking, systematic reviews, 31 literature reviews, and meta-analyses [3]. These systems, in particular, PubMed and MEDLINE, 32 are also used to inform clinical decision-making in the health professions [4] and construct 33 knowledge bases for clinical decision support systems [5]. 34 Research on search queries that inform the development of bibliographic databases or on how 35 queries influence information retrieval sets were once common lines of inquiry [6] but these have 36 subsided in recent decades [7]. Search query research has largely shifted away from a Boolean 37 model of information retrieval and has focused on ranked-based keyword systems [8] or on 38 database coverage [9-12].39 Researchers, librarians, information scientists, and others rely on bibliographic databases to 40 conduct research, to instruct future information and other professionals how to conduct literature 41 searches, and to ...