Skip to Main Content

Literature Reviews (Health Sciences)

Collaborate with a Librarian

It is highly recommend that you collaborate with a librarian on developing your search strategy. Librarian are experts on identifying appropriate databases, developing comprehensive search strategies, writing the search methodology, and providing documentation for the line-by-line search strategies for use in the manuscript appendices.

According to the National Academies of Sciences, Engineering, Medicine: Standards for Systematic Reviews Standard 2.1.3, the review team should "include expertise in searching for relevant evidence".

Build Search Strategy

Before starting your database search, think about terms that can be used to describe the key concepts in your research question. Start your search with terms that you think make sense. When you find citations that are highly relevant to your research, take a closer look at those records. Examine those records for two types of terms that you can use in your search: subject headings and keywords.

  • Subject Heading: A single, assigned term that stands for a concept. For example, in PubMed, any paper that discusses acetylsalicylic acid would be assigned the Medical Subject Heading (MeSH) term aspirin. A search for the MeSH term Aspirin in PubMed should find papers written about aspirin whether or not the word actually appears in the title or abstract.
  • Keyword: Term used for a concept in everyday language. For example, if you need to find articles written about bedpans, the Medical Subject Heading (MeSH) term Toilet Facilities in PubMed may be too broad. Just searching for bedpan OR bedpans by typing this directly into the search box might work better.

Subject headings and keywords have different advantages and disadvantages. Keywords can retrieve new articles that do not yet have subject headings assigned to them. You can also use keywords to capture alternative spellings. Subject headings, however, will help you find highly relevant articles, and may mitigate the need to search for synonyms.

When you conduct your search, consider whether it makes sense to use keywords, subject headings, or both.

See "Documenting Your Search" to learn how to keep track of useful terms.

Boolean Operators

 

    Connecting terms with AND requires all terms to appear in the same article.

 

 

 

    Connecting terms with OR results in articles that include one term, a combination of terms, or all of the terms.

Apply Filters/Limits

Many databases allow you to filter your search. You can usually find filters are on the left-hand side of your results page. Based on your selection criteria, you may want to filter your results based on:

  • Age group
  • Sex
  • Publication date range (e.g. last 10 years)
  • Species
  • Source (e.g. journal name)
  • Language
  • Article type (e.g. review, research report, etc.)
  • Study type (e.g. randomized-control trial, cohort, etc.)

Precision vs sensitivity in Systematic Reviews

Searches should seek high sensitivity, which may result in relatively low precision.

  • Recall (Sensitivity): The number of relevant reports identified divided by the total number of relevant reports in existence
  • Precision: The number of relevant reports identified divided by the total number of reports identified. 

More info: Higgins JPT, Green S (editors). Sensitivity versus precision (section 6.4.4) Cochrane Handbook for Systematic Reviews of Interventions. Version 5.1.0 [updated March 2011]. The Cochrane Collaboration. Available from http://handbook-5-1.cochrane.org. Accessed June 1, 2018.

Translate Search Between Databases

If you're searching different databases for information, keep in mind that you may need to adjust your search terms for each database. For instance, the equivalent subject heading for "Heart, Artificial" in PubMed is "Heart, Mechanical" in CINAHL. Additionally, because CINAHL is an allied health and nursing database, you will find specialized subject headings such as "Toileting" in CINAHL that you won't find in PubMed.

Keywords are more likely to stay consistent across databases.