This is an archived version of the Handbook. For the current version, please go to or search for this chapter here.

6.4.5  Controlled vocabulary and text words

MEDLINE and EMBASE (and many other databases) can be searched using standardized subject terms assigned by indexers. Standardized subject terms (as part of a controlled vocabulary or thesaurus) are useful because they provide a way of retrieving articles that may use different words to describe the same concept and because they can provide information beyond that which is simply contained in the words of the title and abstract. When searching for studies for a systematic review, however, the extent to which subject terms are applied to references should be viewed with caution. Authors may not describe their methods or objectives well and indexers are not always experts in the subject areas or methodological aspects of the articles that they are indexing. In addition, the available indexing terms might not correspond to the terms the searcher wishes to use.


The controlled vocabulary search terms for MEDLINE (MeSH) and EMBASE (EMTREE) are not identical, and neither is the approach to indexing. For example, the pharmaceutical or pharmacological aspects of an EMBASE record are generally indexed in greater depth than the equivalent MEDLINE record, and in recent years Elsevier has increased the number of index terms assigned to each EMBASE record. Searches of EMBASE may, therefore, retrieve additional articles that were not retrieved by a MEDLINE search, even if the records were present in both databases. Search strategies need to be customized for each database.


One way to begin to identify controlled vocabulary terms for a particular database is to retrieve articles from that database that meet the inclusion criteria for the review, and to note common text words and the subject terms the indexers have applied to the articles, which can then be used for a full search. Having identified a key article, additional relevant articles can be located, for example by using the ‘Find Similar’ option in Ovid or the ‘Related Articles’ option in PubMed. Additional controlled vocabulary terms should be identified using the search tools provided with the database, such as the Permuted Index under Search Tools in Ovid and the MeSH Database option in PubMed.


Many database thesauri offer the facility to ‘explode’ subject terms to include more specific terms automatically in the search. For example, a MEDLINE search using the MeSH term BRAIN INJURIES, if exploded, will automatically search not only for the term BRAIN INJURIES but also for the more specific term SHAKEN BABY SYNDROME.  As articles in MEDLINE on the subject of shaken baby syndrome should only be indexed with the more specific term SHAKEN BABY SYNDROME and not also with the more general term BRAIN INJURIES it is important that MeSH terms are ‘exploded’ wherever appropriate, in order not to miss relevant articles. The same principle applies to EMTREE when searching EMBASE and also to a number of other databases. For further guidance on this topic, review authors should consult their Trials Search Co-ordinator or healthcare librarian.


It is particularly important in MEDLINE to distinguish between Publication Type terms and other related MeSH terms. For example, a report of a randomized trial would be indexed in MEDLINE with the Publication Type term ‘Randomized Controlled Trial’ whereas an article about randomized controlled trials would be indexed with the MeSH term RANDOMIZED CONTROLLED TRIALS AS TOPIC (note the latter is plural).  The same applies to other indexing terms for trials, reviews and meta-analyses.


Review authors should assume that earlier articles are even harder to identify than recent articles. For example, abstracts are not included in MEDLINE for most articles published before 1976 and, therefore, text word searches will only apply to titles. In addition, few MEDLINE indexing terms relating to study design were available before the 1990s, so text word searches are necessary to retrieve older records.


In order to identify as many relevant records as possible searches should comprise a combination of subject terms selected from the controlled vocabulary or thesaurus (‘exploded' where appropriate) with a wide range of free-text terms.