Research Guides: Systematic Reviews for the Health Sciences: Finding the Search Terms

Systematic ReviewsToggle Dropdown

Part 2: Preparing for the Search

A large portion of your systematic review will be spent in creating and gathering search terms (keywords and controlled vocabulary). This is a key part where librarian assistance can be extremely helpful.

As stated in the JBI Manual for Evidence Synthesis: "Whenever possible, the search should be undertaken by librarians, information specialists or other expert searchers".

FINDING THE SEARCH TERMS

Subject heading frequency analysis of gold standard articles

Term Harvesting

Create search string 1

Test search string 1 against gold standard articles, if GSAs are not found modify the search string

Translate search string to remaining databases

Recommendations

MECIR (Methodological Expectations of Cochrane Intervention Reviews) Standards for Search Development include:

Standard	Statement	Mandatory or Highly Desired	Cochrane Handbook
C32	Inform the structure of search strategies in bibliographic databases around the main concepts of the review, using appropriate elements from PICO and study design. In structuring the search, maximize sensitivity whilst striving for reasonable precision. Ensure correct use of the ‘AND’ and ‘OR’ operators.	Mandatory	4.4.2
C33	Identify appropriate controlled vocabulary (e.g. MeSH, Emtree, including 'exploded' terms) and free-text terms (considering, for example, spelling variants, synonyms, acronyms, truncation and proximity operators).	Mandatory	4.4.4
C34	Use specially designed and tested search filters where appropriate including the Cochrane Highly Sensitive Search Strategies for identifying randomized trials in MEDLINE, but do not use filters in pre-filtered databases e.g. do not use a randomized trial filter in CENTRAL.	Highly Desired	4.4.7
C35	Justify the use of any restrictions in the search strategy on publication date and publication format.	Mandatory	4.4.5

Information taken from: https://community.cochrane.org/sites/default/files/uploads/MECIR%20Version%20February%202022.pdf

Step 7 & 8: Subject Heading Analysis & Term Harvesting

Use the Yale MeSH analyzer to gather the medical subject headings of your gold standard articles. Results can be exported and analyzed so that you can choose appropriate medical subject headings.

During term harvesting gather terms from:

Gold Standard Articles (especially from the title and abstracts)
Medical Dictionary
DynaMed Plus

Think about:

Alternate spellings
Acronyms
Phrases

These tools can help determine frequency of terms:

Example:

asymptomatic intracerebral hemorrhage
MESH	PubMed ENTRY TERMS	Emtree	Embase synonyms	Other
Cerebral Hemorrhage[Mesh]	Cerebrum Hemorrhage*	'brain hemorrhage'/exp	brain bleeding	ICH	Asymptomatic
	Cerebral Parenchymal Hemorrhage*		brain haemorrhage	aICH	Non symptomatic
	Intracerebral Hemorrhage*		brain hemorrhage		nonsymptomatic
	Cerebral Hemorrhage*		brain microhaemorrhage		symptomless
MESH Tree	Cerebral Brain Hemorrhage*	Emtree Tree	brain microhemorrhage		symptom-less
Basal Ganglia Hemorrhage		brain ventricle hemorrhage	cerebral haemorrhage
Putaminal Hemorrhage		cerebellum hemorrhage	cerebral hemorrhage
Cerebral Hemorrhage, Traumatic		massive intracerebral hemorrhage	cerebral microbleed
Cerebral Intraventricular Hemorrhage		subarachnoid hemorrhage	corpus callosum bleeding

Step 9 & 10: Create & Testing Search String #1

Once you have gathered the controlled vocabulary, keywords, synonyms, and phrases - it is time to begin combining the terms to create a search string for one database. Key components for this include: Boolean operators, truncation, proximity searching, and wildcards.

Boolean Operators

AND - Narrows the search
- Diabetes AND Nutrition
OR - Broadens the search
- Infants OR Newborns
NOT - Narrows the search
- Hearing loss NOT Hearing aids

Truncation

Truncation is used to search for multiple endings of a word and is typically indicated with an asterisk.
For example: obes* will find obese and obesity

Proximity Searching

Proximity Searching is a way to search multiple terms that appear within a certain number of words from each other. There are typically two options to use -- Near or Next (sometimes also called Within). Order does not matter with Near, meaning ice NEAR/1 cream will find ice cream and cream ice. Order does matter with Next, meaning ice NEXT/1 cream will only find ice cream. Use the table below to see how popular databases require proximity operators to be listed. (Replace x with a number to represent the amount of words present between your terms)

Proximity Operators for Popular Databases
	Near	Next (or Within)
Embase	NEAR/x	NEXT/x
EBSCO Databases (including CINAHL, Medline, PsycINFO)	Nx	Wx
Scopus	W/x	PRE/x
PubMed	does not allow proximity searching	does not allow proximity searching

Wildcards

There are a variety of wildcards that are available in different databases. They are similar to truncation but are used to stand for either a single letter or multiple letters within a word. They are indicated by a question mark or a hash mark (#) - check with the database you are using to determine which symbol is required.
For example spelling variations of hemorrhage can be found with:
- Web of Science: h$emorrhage
- Scopus: h*morrhage

Testing the Search

Once you have formatted your search string you need to test it to make sure that it finds your gold standard articles. If your search string does not find your GSAs you will need to make appropriate adjustments.

Step 11: Translating Search String to Other Databases

Unfortunately, there is not a magic search string format that will work in all databases. You will need to translate search string #1 that you have verified finds your gold standard articles. This includes potentially changing or removing:

Controlled vocabulary
Truncation & wildcard characters
Proximity operators

There are quite a few tools and guidelines that can help you with this step: