Search filters & other search development tools
Text mining
A number of text-mining tools exist to help you identify MeSH terms (subject headings in MEDLINE) that are relevant to a concept that you are including in your search strategies. You may need to use multiple MeSH terms to capture a single concept.
-
Yale MeSH Analyzer - Enter the PubMed ID numbers (PMID) of known relevant articles and create a file with a list of MeSH headings used to index those articles.
-
PubMed PubReminer - Useful for high precision searches on specific topics or the PMIDs of articles that are relevant to your topic. PubReminer ranks the MeSH terms or words that are most commonly associated with a set of records you provide.
-
MESH on Demand - Enter a block of text and this tool will return a list of MeSH terms relevant to your text.
Other strategies
Search the literature for other evidence syntheses that are relevant to your topic to see how other researchers developed their search strategies. Search strategies should be available in the text of the manuscript or as an appendix. Keep in mind that search strategies are specific to individual databases and platforms. See the page on Translating a Search Strategy for more.
Check the indexing of relevant articles for official subject terms and keywords that you can use for your search strategy.
Search filters or hedges
Search filters, sometimes called hedges, are search strategies that have been pre-developed for a given concept, idea, or study design. These search strategies include index terms and keywords and are formulated for specific databases and platforms (for example, the MEDLINE database via the Ovid platform). Filters may or may not be validated for sensitivity, specificity, or precision, but they can be a useful tool to take advantage of the work that others have already done to identify terms to find literature on a given concept. If you reuse a search filter that others have developed, you should cite or acknowledge them in your manuscript.
Search filters can be developed for individual concepts, or to find particular types of study designs. While useful tools, study design filters should be used with care, as it cannot be guaranteed that authors or databases have correctly identified the study design of a given manuscript in a given database, and there is always a risk of missing relevant articles that are not labeled properly in a database. Using filters to exclude certain concepts or study types should rarely be done, as this comes with a great risk of overlooking relevant articles. Filters may also be out of date and not incorporate changes to a database's subject thesaurus or interface changes.