The Archway Campus - Site Title
Archway Healthcare Library
       
Layout Graphic Layout Graphic Layout Graphic Layout Graphic Layout Graphic Layout Graphic Layout Graphic
Layout GraphicHomeSearch Layout Graphic Layout Graphic Layout Graphic Layout Graphic Layout Graphic
The Archway Healthcare LibraryLibrary InformationLibrary CatalogueHealth Information TrainingDatabasesRecommended WebsitesElectronic Journals Layout Graphic Archway Healthcare Library logoSearching databases: a brief guide

 
You should break a topic or question down into separate concepts. For example, if you want to find articles on the use of aspirin to prevent stroke, stroke and aspirin are separate concepts.
Databases work best if you search for articles on each concept separately and combine those separate searches at the end of the process.
Searches on any particular concept should take account of synonymous terms (eg stroke, cerebral vascular accident, CVA etc…).
Detailed searching for the purpose of good quality research should cover both defined subject headings (known as MeSH in medline) and textword equivalents.
Every article that is added to medline (and most other databases) is given a number of defined subject-headings to describe the contents of the article. On Medline this system is called MeSH (Medical Subject Headings). An article may have more than 20 subject headings, which come from a Thesaurus of controlled terms. For example all articles concerning stroke that are added to the Medline database should receive the subject heading Cerebral Vascular Accident. There is no subject heading of stroke. A subject heading search will usually provide a fairly refined search where the term or terms you have searched for are always in context.
A thorough search should also look for textwords. A textword search looks for a specific pattern of letters in the title, abstract or subject-headings of references. A textword search on diabetes, for example, will find all the articles that have this pattern of letters in their title, abstract or subject headings. However, it will not find articles containing the word diabetic. Wild-cards can get round this problem by allowing variant endings on words. Unfortunately, wild-cards differ from database to database. On OVID versions of Medline, EMBASE and CINAHL use $ i.e. diabe$. On the Cochrane Library or Pubmed use * i.e. diabet*
Once you have completed the searches on each concept the searches should then be combined. Firstly, searches on a single concept should be combined with OR (eg stroke$ (textword) OR Cerebral Vascular Accident (subject heading)). Finally searches on different concepts should be combined with AND (eg stroke searches AND aspirin searches) to get to your final result.
If you have too many hits, a further concept could be introduced to narrow your search down. Alternatively, most databases allow limits to be applied to a search (for example by publication type, or age-group of patients).
There a number of databases to choose from and a thorough search will include more than one database. Most sources, such as Medline and CINAHL list single journal articles. The Cochrane Library is different, offering high quality systematic reviews which bring together high quality controlled trials to provide a best picture of the evidence on a particular healthcare intervention.
After a successful search be prepared to look at the articles you find with a critical eye. To assist this process see the checklists provided by the Critical Skills Appraisal Programme CASP

further reading:  Greenhalgh, Trisha    How to read a paper   3rd ed. BMJ: 2006

return to databases page

return to home page                                                                                                                                                                                                                   

Layout Graphic
Layout Graphic Layout Graphic
Layout Graphic Layout Graphic Layout Graphic
Layout Graphic Layout Graphic
Maintained by: Richard Peacock Most recent modification: 25 September, 2008 10:32 AM
Hosted by CHIME
Copyright © 1999-2004 CHIME, Royal Free & University College Medical School, University College London. All rights reserved.