Table 2.

Glossary

Browse: Browsing is the practice of scanning for information by reviewing content. This may include using tables of contents, indices in books or other materials, web directories, full journal issues, specific web pages, or other types of information scanned without using a formal search strategy.

Citation index: A type of database or database function that enables searchers to analyze relationships between publications through citations, including what publications have cited, and which publications are citing the publication(s) you are interested in. Common examples include Science Citation Index, Scopus, and Google Scholar.

Cited reference: Publication referenced in a given publication.

Citing reference: Publications that have referenced a given publication.

Database: Within PRISMA-S, this refers to a literature database designed to search journal literature. Databases may be multidisciplinary or specialized. Many include specialized search features, subject headings, and structured data designed to facilitate easy and comprehensive searching. Examples include MEDLINE, EconLit, and PsycINFO.

Digital object identifier: Also called a DOI, a digital object identifier is a unique code assigned to a publication, dataset, or other online item or collection that will remain constant over time.

Field code: Unique to each database platform and database, field codes are used to specify where a term is searched for in a database record. In PubMed, for instance, the field code [tiab] is placed after a search term to tell the database to search only within the title and abstract fields.

Filter: Filters are predefined combinations of search strategies designed to locate references meeting certain criteria, usually publication type, topic, age group, or other categorization. Filters generally are combinations of keywords, subject headings or thesaurus terms, logical operators, and database-specific syntax. Many filters are validated and offer sensitivity and specificity information that allows searchers to determine their usefulness for a given search. Filters may also be called hedges or optimal search strategies and are designed for other searchers to use and reuse.

Indexing: Application of standard terminology to a reference to describe the contents of the full article. Depending on the database or other information source, indexers may add subject headings or thesaurus terms as well as list age groups, language, human studies, study design, publication type, or other descriptive terms. Examples of indexing terminology include MEDLINE’s Medical Subject Headings (MeSH) and Embase’s EMTREE.

Information source: Any database or other resource such as a web site, journal table of contents, email alert, web directory, contact with authors or industry, study registry, or preprint server, etc. searched or browsed as part of the search.

Literature search: Here, an overall term for the entire information retrieval process as part of a systematic review. This includes the full range of searching methods and information sources, including databases, study registries, regulatory datasets, web searches, government documents, unpublished data, and much more.

Limits: Features built into databases to allow searchers to quickly restrict their search by one or more categories. Common limits built into databases include publication date ranges, language, gender, age group, and publication type. Limits are different from filters (see above) and are also not the inclusion/exclusion criteria used in the screening process.

Multi-database search: Many database platforms offer more than one database on the same platform. Some platforms allow users to search these multiple databases at one time, for example using the Ovid platform to simultaneously search MEDLINE, Embase, and the Cochrane Database of Systematic Reviews.

Peer review: In PRISMA-S, this refers to the peer review of search strategies prior to executing the search. Peer review is used to identify errors, missing keywords or subject headings, and other issues within a search strategy. One commonly used tool for search strategy peer review is the Peer Review of Electronic Search Strategies (PRESS) Guideline [1].

Platform: Many databases are available on multiple different systems, each of which have their own specifications for how a search strategy can be constructed. The location or host system of the database is the platform. Platform is sometimes referred to as the interface or vendor. Common examples include Ovid, EBSCOhost, ProQuest, and Web of Science.

Records: Individual items retrieved from any type of search, though most commonly used in conjunction with database searches. Records may also be referred to as references or hits.

Repository: An online archive for varying types of electronic files, including text documents, data files, and more. Repositories may be hosted by an institution or more broadly available.

Rerun: Re-executing the same search strategy in the same database one or more times after the original search was conducted. See Updating a search strategy.

Search: Overall term for the entire information retrieval process as part of a systematic review. It can also refer to searching a specific database, web site, or other information source.

Search strategy: Structure of terms, logical operators, and syntax elements (field codes (see above), adjacency operators, phrases, etc) that is used to search a database or other information source. A search strategy may be very simple or very complex, depending on the information source and requirements of the search.

Sensitivity: A measure of how well a search strategy finds relevant articles, sensitivity (usually expressed as a percentage) is the number of relevant records found with a search strategy divided by the total number of relevant records in a given information source. Highly sensitive search strategies or filters detect most or all records that are relevant. Together with specificity, sensitivity is a measure used to assess the performance of filters. Sensitivity may also be called recall.

Specificity: A measure of how well a search strategy omits irrelevant articles, specificity (usually expressed as a percentage) is the number of irrelevant records not found with (or excluded by) a search strategy divided by the total number of irrelevant records in a given information source. Search strategies or filters with high specificity will find few irrelevant articles. Together with sensitivity, specificity is often used to assess the performance of filters.

Study registry: A database of records of research studies in progress. Originally designed for clinical trials as a location for patients to find clinical trials to join, study registries have spread beyond biomedical research to other fields. Study registries may also contain research results, posted after study completion.

Supplementary materials: Additional content for a study that does not fit in the main manuscript text. For a systematic review, supplementary materials should include full search strategies for all information sources and more complete search methods description. Supplementary materials are generally submitted with the manuscript for peer review.

Syntax: Search structure and organization, based on a set of rules governing how a search operates in a specific database and platform. Rules might include field codes, phrase and adjacency searching, Boolean operators, and truncation, amongst others.

Systematic review: For the purposes of PRISMA-S, systematic review is used for the entire family of methods-based reviews. This includes rapid reviews, scoping reviews, meta-narrative reviews, realist reviews, meta-ethnography, and more.

Updating a search strategy: To ensure currency, authors often search for additional information throughout the systematic review process or before submitting a report. The search may be updated by running the exact same search (rerunning the search) or by conducting a new or modified search to locate additional references.