Gå till huvudinnehållet

Search strategies for systematic reviews

Get started

Create a search strategy - Step by step

Before you start your search

  • Have you already found relevant studies?
Search for the studies in one of the selected databases and document which keywords are used in the title and abstract as well as which subject headings/descriptors they have been indexed with, which are relevant to your search.

Do a test search

  • Have you done a test search?
Do a test search in one of the databases (if you know which one will be most important, start in that one). The search does not have to be comprehensive. Document words in title, abstract, key words and subject headings.
  • Have you found a literature review?

Have a look to see how they have reported their search strategy,  you can document the words in the title, abstract and subject words that are relevant to your search.

  • Have you found relevant studies in your test search?
Keep searching to see if you can find more relevant subject headings, key words and concepts. Make a note when you find new ones. You can change your search strategy when you find new keywords. Also look if the database provides information about similar articles, which can be called "Similar articles", "Related articles" etc.

Use search filters

  • Have you found any published search filters?
Search filters are search strategies designed to capture a certain aspect, such as study design. There are search filters that are validated, tested against a gold standard of relevant articles to make sure that these are captured by the search filter. The aim is to find a search strategy that captures as many relevant studies as possible while limiting the number of irrelevant studies. The search filters are designed to fit different databases. There are search filters for study design, age, geography etc.

ISSG Search Filter Resource

CADTH Search Filter Database

SIGN Search Filters

Search techniques

How to search with boolean operators

At least one of the words should be included in the search - OR

OR indicates that one or more words should be included in the search. Use OR to search for both the words as subject heading and free text words or on synonyms and related terms. You get more results and do a more exhaustive search.

Nurses[Subject Heading] OR nurses

All the words must be included in the search - AND

AND indicates that all keywords or search blocks that you combined should be included in the search. You get fewer but more relevant hits.

(Nurses[Subject Heading] OR nurses) AND (covid-19 OR sars-cov-2)

Search for parts of a word - truncate

Truncation means shortening a word to its word stem and then putting an asterisk (*) which indicates that different endings, inflections and spellings of a word should be searched for.

nurse* (searches for nurse, nurses etc.)

Phrase search - search for phrases or concepts - citations

When you want to search for terms that contains two or more words. Specifies that the words are to be searched together in the exact order and with the spelling that is written. You get fewer and more relevant hits.

"nursing care"

Example search strategy

Example Search Strategy:

Nurses experience of psychological stress when caring for patients with covid-19

PEO

Population Nurses Nurse[Subject Heading] OR nurse[Free text word] OR nurses[Free text word]
Exposure Caring for patients with Covid-19 Covid-19[Subject Heading] OR SARS-CoV-2[Subject Heading] OR covid-19[Free text word] OR SARS-CoV-2[Free text word]
Outcome Psychological stress Stress, Psychological[Subject Heading] OR psychological stress[Free text word] OR burnout[Free text word]

Search string

(Nurse[Subject Heading] OR nurse[Free text word] OR nurses[Free text word])  AND (Covid-19[Subject Heading] OR SARS-CoV-2[Subject Heading] OR covid-19[Free text word] OR SARS-CoV-2[Free text word]) AND (Stress, Psychological[Subject Heading] OR psychological stress[Free text word] OR burnout[Free text word])

Supplementary Searches

Supplementary searches

In order to make searches as exhaustive as possible, methods other than searching databases can also be used. You can, for example:

  • Look in the reference lists of the included studies.
  • See how the included studies have been cited by others.
  • Identify important journals that you search by hand
  • Contact important researchers in the field.

Citation chaining

Citation chaining, sometimes called snowballing, is a search method where you look at the bibliographies of key articles to find other related articles. The articles retrieved this way might not have been retrieved in your database searches, making it a complementary search method. There are two main types of citation chaining, backward chaining and forward chaining.

Backward & forward chaining

  • Backward chaining is a good way to find earlier research on your topic. Check the reference lists of key sources, to see which works they have cited.
  • Forward chaining is a good way to find more updated research on your topic. Check citations for key sources to see which works that have cited them. This works best for articles that are a few years old, as they are more likely to have been cited than newer ones.

Forward chaining in subject specific databases

Several of the subject specific databases in the list found under "Searching in databases" lets you do forward chaining.

Example from a database record in Criminology collection

Forward chaining in citation databases

The are also citation databases specifically intended for forward chaining, where you can find out how an article has been cited by other researchers.

Example from a database record in Web of Science

Forward chaining in Google Scholar

You can also use Google Scholar to do forward chaining, search for the title and click "Cited by..." to get a list of other articles where the authors have cited the article you're looking at.

Related articles

Another chaining method, is using the database feature "Related articles". The articles featured shares similar features with your key article, so this might be a good way of finding more articles. Several of the the databases from the library shows related articles. The databases uses different names for this feature, for example "Suggested sources", "Related records" and "Similar articles".