Systematic Review

Starting your search

Searching systematically is typically quite extensive and a critical phase of the systematic review process. It is important to try to strike a balance between sensitivity (comprehensiveness) and precision (relevancy or accuracy) in your search strategy.

A strategy that has a higher sensitivity often lowers the precision of the results and higher precision often lowers sensitivity of results.

To find balance in a search strategy it is important to consider synonyms and the variations in which others might describe or articulate the concepts researchers identified when formulating the research question.

By combining both keywords and subject headings in a search strategy, a researcher is more likely to strike the balance required to find the evidence needed to assist in answering the research question.

PICO Framework

Using the PICO framework (or one of the other frameworks mentioned in the Research Question section) helps in developing a systematic search strategy.

In sections 5.2 – 5.6 of The Cochrane handbook for Systematic Reviews of Interventions (2011) researchers will find factors for consideration when developing the search criteria. These factors align with the elements of the PICO framework.

Medical Subject Headings - MeSH

The National Library of Medicine publishes Medical Subject Headings (MeSH).

MeSH is a controlled vocabulary used for indexing and cataloguing, enabling health professionals, medical librarians, researchers and students to search for health related literature.

Medline and PubMed databases use NLM MeSH and subject descriptors for indexing articles. When conducting a search in either database the search term entered will be mapped automatically to the equivalent MeSH term.

Other primary databases have their own controlled vocabulary for indexing and cataloguing articles, for example, in CINAHL, the thesaurus is called CINAHL Headings and in Embase their thesaurus is known as EMTREE.

When searching for subject headings, remember they can differ from one database to the next, for example, the MeSH term in Medline is “Physical Therapy Modalities” and the subject heading for physiotherapy in the CINAHL thesaurus is “Physical Therapy”.

Medline

Mesh headings and sub headings in Medline

CINAHL

Mesh headings and sub headings in CINAHL

 

*Note – Where appropriate subject headings should be ‘exploded’ to prevent missing any relevant articles. Subheadings can assist in retrieving papers with similar subheadings.

Scope notes can also help in deciding if the subject heading or subheading are appropriate for the search.

screen shot of subheadings for physical therapy

 

For a quick tutorial, please watch the YouTube clip by VCU Libraries

Keyword searching

Keywords are the terms health professionals use in everyday practice. When developing your search strategy it is important to compile a list of synonyms, abbreviations and alternative spellings of the keywords identified in the development of the research question.

Keep in mind when looking at alternative spelling that spelling mistakes are considered. Here is an example alternative and spelling mistakes of the word haemorrhage.

 

Set#

Searched for

Databases

Results

S1

hemorrhage

MEDLINE®

268640 

S2

haemorrhage

MEDLINE®

35787 

S3

hemorhage

MEDLINE®

19 

S4

haemorhage

MEDLINE®

12 

S5

hemmorrhage

MEDLINE®

47 

S6

haemmorrhage

MEDLINE®

S7

hemmorhage

MEDLINE®

29 

S8

haemmorhage

MEDLINE®

11 

S9

hemorrage

MEDLINE®

154 

S10

haemorrage

MEDLINE®

83 

S11

hemorage

MEDLINE®

S12

haemorage

MEDLINE®

S13

hemmorage

MEDLINE®

S14

haemmorrage

MEDLINE®

 

In a scenario such as the example above, the recommendation is that each term be searched individually. Researchers are aware of the terms within their own professional area that are often misspelt or have alternate spelling.

The use of truncation and wildcard options are not recommended in this scenario.

Text mining

Specialised software has been developed to assist in identifying subject headings and keywords. Below are some of the text mining tools available.

Yale MeSH Analyser – Use the PubMed artcle ID number (PMID).

PubMed PubReMiner – Enter your search terms to find details of articles including author keywords and MeSH

NLM MeSH on Demand – identifies MeSH terms from your research question.

Voyant – Analyse the full text of all or part of an article.

Boolean operators

Boolean operators are used to combine or exclude the concepts identified in the PICO framework. Using boolean operators can narrow or broaden a search strategy.

 

Diagram on how to search using Boolean operators

Boolean Operators – Mississippi College

Truncation and wildcards

Truncation is used to retrieve all the possible variations of a search term, for example, therap* = therapy, therapies, therapists, therapeutic, etc.

Wildcards allow you to search for variations of spelling, for example, wom?n = woman or women.


Examples of truncation and wildcards

Truncation and Wildcard – Monash University

Proximity operators

Proximity operators are useful when looking for a combination of keywords within a phrase or paragraph of text.

 

Examples of proximity operators

Proximity operators – Monash University

Search limits – your exclusion criteria

Researchers do not have to enter all of the exclusion criteria into the search strategy. Sections of the exclusion criteria can be applied using the database limiters.

 

Examples of exclusion criteria and questions to ask