The goal of systematic review searches is to identify all relevant studies on a topic. Therefore, systematic review searches are typically quite extensive. It is necessary, however, to strike a balance between striving for comprehensiveness and maintaining relevance when developing a search strategy. Increasing the comprehensiveness (or sensitivity) of a search will reduce its precision and will retrieve more non-relevant articles.
The goal of a systematic review search is to maximize recall and precision while keeping results manageable. Recall (sensitivity) is defined as the number of relevant reports identified divided by the total number of relevant reports in existence. Precision (specificity) is defined as the number of relevant reports identified divided by the total number of reports identified.
Issues to consider when creating a systematic review search:
|
|
|
|
|
|
|
|
|
|
|
A good rule of thumb for picking databases to search is find five seminal papers and figure out where the papers are indexed. When selecting databases, use one BIG database that's multidisciplinary abbd two smaller, subject specific databases.
Another important thing to consider is if you are working with a team of researchers form other universities, make sure all members of the team have the same databases with the same access.
Most databases offer a thesaurus or list of available subject headings that can be allocated to an article by the author or indexer.
Using controlled terminology can be a great way to access information in your field. Many databases have a preferred vocabulary or thesaurus built into them. The lists of preferred terms in Medline (MeSH), Psycinfo, and SocINDEX are particularly extensive.
Each is a hierarchical arrangement of broader terms, preferred and related terms, and narrower terms, designed to map the context and content of their respective fields.
This can be a great way to add terms to your search or even build your search. Each database uses different search terms so the search that works in one database, may or may not be successful in the next.
Evidence synthesis methods require authors to search multiple databases, and not all databases accept the same search "syntax." Each individual database requires use of specialized search syntax, and therefore evidence synthesis search strategies must be 'translated' between databases.
For example, a search for vitamin D[tiab] in PubMed will show you all citations with the phrase "vitamin D" in the title, abstract, or keywords, but a search for vitamin D[tiab] in Web of Science will not work at all.
Below is a template that you can use to document your search strategy translations and results, as well as search syntax translation tools and examples. Contact a librarian for assistance with search syntax translation.
Librarians can recommend databases and other sources to search for a systematic review. The sources you choose will depend on your research question and the disciplines in which relevant research may be conducted. Below is some guidance for choosing scholarly databases in a number of research areas. Check the library's database list for a full list of available sources across all disciplines.
Librarians can also help with designing complex searches using the specialized syntax of individual databases. Consult with a librarian if you have questions.
Note that these databases largely focus on the published, peer-reviewed literature. For guidance and resources for searching other types of information (i.e., gray literature), see the next tab in this guide.
Link to all of the subject specialists at LSU and their contact information.
All of the guides (Course, Research, Topic, etc.) that we have.
There are three Boolean operators that are used to connect terms and tell databases how and what to search for: AND, OR, NOT.
AND is to combine terms, usually unlike terms/concepts. AND narrows a search. Example: social media AND teenagers
OR is typically used with synonyms and similar terms. OR broadens a search. Example: teenagers OR adolescents
NOT is used to exclude something. Example: teenagers NOT bullying
We use parentheses to help group parts of the search query, especially when we have several parts, and to tell the database the order of the query. Think about the search query as a mathematical equation.
All put together, they look like this:
social media AND (teenager OR adolescent) NOT bullying
Truncation allows you to find different endings to a word. The symbol in many databases is: *
Example: teenage* captures teenager, teenagers, teenaged.
Be careful not to truncate too far into the word. For example: car* will capture car, cardiology, carbohydrate, caramel, carabidae, carassius, and thousands more words.
carbohydrat* would be a better way to truncate.
Wildcards are symbols used within a word to represent a letter for a variation on spelling. While not every database uses them anymore, for those that do, the symbol is often ? or $, though always best to check the database documentation.
Example: behavio$r captures both the American spelling, behavior, and the British spelling, behaviour
Don't forget to document your search. For each search record make sure to record: