The search protocol outlines the parameters of the search in relation to the question to be answered and the exclusion/inclusion criteria. The search framework (e.g. PICO) enables you to break the search down into the relevant concepts which define the search. The next step is to devise the search strategy you will use to conduct the systematic searches across the different databases. This means working out the keywords and controlled vocabulary terms which will need to be included in the search strategy to ensure that you find all of the relevant literature on the topic.
For systematic reviews, it is recommended that both keyword (free-text) searching and controlled vocabulary searching is undertaken. This enables the search to be as comprehensive as possible and helps avoid any problems inherent with the ambiguity of free-text searching as this is reliant on the keywords you use being the same as those used by the author/publication.
Keyword, free-text, or ‘natural language’ searching is identifying the different keywords which can be used in the literature to define a concept. This includes considering synonyms and related terms, differences in English and American terminology, acronyms and abbreviations, and spelling variations. Keyword searching will find results if the terms searched are present in the article title, abstract or keyword fields. Finding relevant articles is dependent on the keywords selected being the same as those used by the author(s)/publication. Problems include the ambiguity of ‘natural language’ and the lack of consistency in terminology.
Controlled vocabulary searching uses the pre-set subject terms defined in the individual database within their thesaurus of subject terms. These are hierarchical structures containing broader and narrower terms which facilities identification and selection of relevant subject terms. Not all databases use controlled vocabulary: for example, Medline and CINHAL do, whereas Scopus does not.
Each time an item is added to a database that uses controlled vocabulary, an indexer will tag the item with the relevant subject terms that define the content. Using controlled vocabulary searching helps overcome the problems inherent in keyword searching - i.e. the ambiguity of natural language and ensuring that results are not missed because different keywords are used to describe that concept.
Examples of databases that use controlled vocabulary
| Database | Controlled vocabulary |
|---|---|
| Medline | MesH (Medical Subject Headings) |
| PubMed | MesH (Medical Subject Headings) |
| Cochrane Library | MesH (Medical Subject Headings) |
| CINAHL | CINAHL Subject Headings |
| APA PsycInfo | APA PsycInfo thesaurus |
| SPORTDiscus | SPORTDiscus thesaurus |
This short video tutorial by EBSCO shows how to use the CINAHL/MeSH Headings feature within the EBSCO databases.
The following example of formulating a search strategy focuses on the question:
Does cognitive behavioural therapy improve self-esteem in people with eating disorders?
The PICO framework can be used to break this search into its relevant concepts as follows:
Then, the relevant keyword and controlled vocabulary terms are identified for each concept, for example:
Note: 'MH' denotes MeSH Heading so is the controlled vocabulary term used in Medline.
Use the Search history function within databases to search using a combination of free-text and controlled vocabulary terms. Search for each term separately and then create sets for each concept using the OR operator within the search history function. The large concept sets can then be combined using the AND operator to create one final set.
This video provides an overview of using the Search history function in PsycINFO to provide an illustration of how to combine free-text keywords and controlled vocabulary searching.
