Service providers and search interfaces

Last updated: 25 March 2024

What's new in this update

No new research relevant to the chapter was identified during the March 2024 update.

Introduction

Searching different databases will inevitably give different results (for example searching CINAHL versus searching MEDLINE). However, the interface or platform selected for searching each database also has important implications. Searching the same database via different interfaces can result in different results. For instance, searching Embase via Embase.com or via Ovid, or searching MEDLINE via Web of Science, EBSCOhost, Ovid, or PubMed. The variability of results when searching the same database via different interfaces can be a result of differences in search functionality and syntax, the currency of the database and even variations in its content across different interfaces. This all has implications for selecting which search platform/s to use, and for the reproducibility of searches across different interfaces. It is important to document and report which search platform is used for each database (1,2). Bethel and Rogers (3) produced a checklist to evaluate the suitability of database-hosting platforms for complex, systematic searching, and used it to assess EbscoHOST, Ovid and ProQuest. This checklist could be a useful tool to aid decision-making about which platform to use or purchase where more than one is available for the same database.

The inner workings of different search platforms are usually not made publicly available by the vendors, so independent investigations are needed in order to make informed comparisons (1,4). However, there is little published research that formally assesses the differences between search interfaces and their impact on search results; most papers are descriptive in nature, offering advice on how to search each interface and/or which one to use (5,6). Some evaluations have gone beyond a descriptive comparison, measuring the impact of searching via different platforms. The majority of these have compared PubMed with other versions of MEDLINE (4,7-10), although other comparisons have been undertaken for The Cochrane Library (5), AMED (6), CINAHL (11), Embase (12-14) and PsycInfo (1,15). The studies of Cochrane Library, AMED and CINAHL found that the same search can vary in different interfaces, both when doing free text searching, and when using keywords/subject headings (5,6,11). One study that compared searching Embase via Embase.com and Ovid found no difference in retrieved relevant articles (12), but both this study and a more recent comparison rated Embase.com easier to use (12,14). An additional paper about translating a search filter from Ovid Embase to Embase.com compared results retrieved by each line of the original filter to those retrieved by the translated version, providing insights into the idiosyncrasies of the two interfaces, including issues with database content, search syntax, search fields and publication date limits (13).

In relation to PubMed–MEDLINE comparisons, studies have found that PubMed contained more records than other versions of MEDLINE and may have had a higher sensitivity in some cases (7-10). The time lag between PubMed and the versions of MEDLINE offered on interfaces such as Ovid or within Embase.com is also important to consider (4,9,16), although one study was able to rule this out as the main cause of differences in numbers of results across interfaces (10). The conclusions of some of these studies may no longer be accurate, as Epub Ahead of Print, In-Process, In-Data-Review and Other Non-Indexed Citations segments are now available within the Ovid MEDLINE database (17).

All research around differences in database providers suffers from the problem of currency. Database interfaces are constantly being developed and improved, more so than the databases themselves. This means that comments about interfaces are time specific and may quickly become out of date. Also, there are few studies evaluating the performance of real-life search strategies in different interfaces, comparing differences in their recall and precision, and relevant unique results, and providing in-depth explanations of the differences.

Reference list

(1) Bedi RP, Kumar RP. Reproducibility of PsycInfo searches across platforms. Canadian Psychology-Psychologie Canadienne 2023. Advance online publication. http://dx.doi.org/10.1037/cap0000365. [Publication appraisal]

(2) Rethlefsen ML, Kirtley S, Waffenschmidt S. et al. PRISMA-S: an extension to the PRISMA Statement for Reporting Literature Searches in Systematic Reviews. Syst Rev. 2021:10:39. https://doi.org/10.1186/s13643-020-01542-z 

(3) Bethel A, Rogers M. A checklist to assess database-hosting platforms for designing and running searches for systematic reviews. Health Info Libr J. 2014 Mar;31(1):43-53. 

(4) Burns CS, Nix T, Shapiro RM, 2nd, Huber JT. MEDLINE search retrieval issues: A longitudinal query analysis of five vendor platforms. PLoS ONE. 2021;16(5):e0234221. [Publication appraisal] 

(5) Craven J, Jefferies J, Kendrick J, Nicholls D, Boynton J, Frankish R. A comparison of searching the Cochrane library databases via CRD, Ovid and Wiley: implications for systematic searching and information services. Health Info Libr J. 2014;31:54-63. [Publication appraisal]

(6) Younger P, Boddy  K. When is a search not a search? A comparison of searching the AMED complementary health database via EBSCOhost, OVID and DIALOG. Health Info Libr J 2009:26:126-135. [Publication appraisal] 

(7) Irvin, E. Does it matter which version of MEDLINE you search? 12th Cochrane Colloquium: Bridging the Gaps. Oct 2-6 2004. Ottawa, Ontario, Canada. [Publication appraisal] 

(8) Katchamart, W., Faulkner A, Feldman B,Tomlinson G, Bombardier C. PubMed had a higher sensitivity than Ovid-MEDLINE in the search for systematic reviews. J Clin Epidemiol 2011;64:805-807. [Publication appraisal] 

(9) Duffy S, de Kock S, Misso K, Noake C, Ross J, Stirk L. Supplementary searches of PubMed to improve currency of MEDLINE and MEDLINE In-Process searches via Ovid. J Med Libr Assoc. 2016 Oct;104(4):309-12. [Publication appraisal] 

(10) Burns CS, Shapiro RM, 2nd, Nix T, Huber JT. Search results outliers among MEDLINE platforms. J Med Libr Assoc. 2019 Jul;107(3):364-73. [Publication appraisal]

(11) Allison MM. Comparison of CINAHL via EBSCOhost, Ovid, and ProQuest. J Electron Res Med Lib. 2006;3:31-50. [Publication appraisal] 

(12) Fortier KJ, Kiss N, Tongbram V. What is the optimal search engine for results from embase and medline: ovid or embase.com? Value Health 2013:16(3):A25. [Publication appraisal] 

(13) Glanville J, Foxlee R, Wisniewski S, Noel-Storr A, Edwards M, Dooley G. Translating the Cochrane EMBASE RCT filter from the Ovid interface to Embase.com: a case study. Health Info Libr J. 2019 Sep;36(3):264-77. [Publication appraisal]

(14) Mangat G, Sharma S. Comparison of search design and output of Embase using two different search interfaces: Embase.com and OVID (MSR30). ISPOR 2023. May 7-10 2023 Boston, MA, USA. [Puiblication appraisal]

(15) West D, Powers R, Brundage K. A comparison of PsycARTICLES and PsycInfo as offered by APA and third-party vendors. Charleston Advisor 2011:13(2);5–9. https://doi.org/10.5260/chara.13.2.5

(16) Boeker M, Vach W, Motschall E. Time-dependent migration of citations through PubMed and OvidSP subsets: a study on a series of simultaneous PubMed and OvidSP searches.  Stud Health Technol Inform. 2013;192:1196. [Publication appraisal] 

(17) Ovid. MEDLINE® 2021 Database Guide. Wolters Kluwer Health; 2021 [cited 08 July 2021].