Skip to Main Content
St George's University of London Logo

Systematic reviews: finding and managing the evidence

This guide gives an overview of how to plan, execute and organize literature searches to support systematic reviews and other projects and research requiring in-depth searches.

Finding your evidence

 Identifying research evidence

Systematic reviews require a comprehensive and reproducible search of a wide range of sources to identify as many relevant studies as possible. An exhaustive search process minimizes the risk of biased reporting of results and thereby strengthens the validity and trustworthiness of the review. A good search is the most important element of a well-produced systematic review.

Literature searching is an iterative process so be prepared to work on and refine your search strategy as you go; it may take a few iterations of the search to get it right. Please make sure you allocate enough time to first create and review the effectiveness of your search strategies.

Typically, identifying research evidence to include in systematic reviews encompasses the following:

  • Planning the the search strategy and generating your terms
  • Searching  for published and unpublished data
  • Managing references
  • Locating articles and documents
  • Updating literature searches
  • Keeping up to date with new papers
  • Documenting the search

There are a number of methods used to identify published and unpublished data including:

  • Searching electronic databases
  • Scanning references from relevant studies
  • Hand searching key journals and conference proceedings
  • Contacting study authors, experts or manufacturers
  • Searching internet resources
  • Cited reference searching

See the Database searching tab  detailed guidance on approaches to searching electronic databases. Further information on finding unpublished resources can be found under the Other search methods tab.

See below for more information on the range of healthcare databases you can access, library database help-sheets and guides, and scoping searches.

 

Scoping searches

What is a scoping search?

Scoping searches are fairly brief searches of existing literature designed to help you gain an overview of the range and depth of research that exists for a particular research idea.

Scoping searches can also help you identify alternative words and phrases to add to your search strategy. 

Databases to consider running basic, initial searches include:

  • TRIP (open access)
  • Cochrane Library (open access)
  • PROSPERO (open access)
  • Medline/PubMed (open access)
  • Embase
  • CINAHL (nursing or allied health topics)
  • PsycINFO (psychological/mental health topics)

It's recommended that you carry out a scoping search before you produce your actual search as  a test bed.

The healthcare databases

The healthcare databases

There are a number of healthcare and social care databases available to track down published literature. As no single database is completely comprehensive, it’s important to search as many databases as possible to capture research from the widest pool of sources.  This is a major factor in distinguishing systematic reviews from traditional literature reviews and helps to minimise bias. A search of Medline alone is not enough.

Databases are provided via different platforms or hosts, such as:

EBSCOhost 

Ovid 

Proquest

The following key databases should be searched for most questions around healthcare interventions:

  • MEDLINE 
    • The National Library of Medicine's (NLM) key bibliographic database of 24 million references from 5,600 biomedical & life sciences journals.  A distinctive feature of MEDLINE is that the records are indexed with NLM's controlled vocabulary, Medical Subject Headings (MESH). A password is required to access via  HDAS, Ovid or EbscoHost, but MEDLINE is freely available via PubMed.
       
  • Embase 
    • Produced by Elsevier, Embase covers 32 million references to 8k+ biomedical journals. It is particularly strong on drugs information.  Login to access via NHS HDAS and Ovid platforms.
       
  • Cochrane Central Register of Clinical Trials (CENTRAL) 
    • Freely available database from the Cochrane Library,  of reports of randomised and quasi-randomised controlled trials
       
  • Web of Science- Clarivate Analytics
    • A multidisciplinary database of scholarly articles that can help you find highly cited articles

A paper  (Bramer et al, 2017) has also suggested that Google Scholar, Google's academic search engine, is also a useful adjunct to these key databases, but it should not be used to replace traditional databases.

Other databases focus on different population groups or specialties or type of information, and depending on your topic it will also be important to search some of these resources in addition to the ones listed above.  These other databases include:

  • AMED (Allied Health and Complementary Medicine)
  • Campbell Collaboration (Social science systematic reviews)
  • CINAHL (Nursing and Allied Health)
  • EMCARE (Nursing and Allied Health)
  • ERIC- Educational Resource Information Center (Education)
  • HMIC (Health management)
  • OT Seeker (Occupational Therapy)
  • PEDRO (Physiotherapy)
  • PsycINFO (Behavioural Sciences)
  • SpeechBite (Speech and Language Therapy)
  • NICE Evidence (Guidelines and other evidence-based material)
  • TRIP (Evidence-based material and primary studies)
  • Clinical Trials & Results Registers (A collection of national and international trials registers)

Use the library’s database webpage for an overview of  all these databases and more, on-site access links and information about off-site login procedures such as OpenAthens (for NHS users) or Shibboleth/Institutional Login (for St George's staff and students).

Database helpsheets

Database help-sheets, guides and video tutorials

The library web-site's help page houses a number of guides and links to videos to help you navigate different  database platforms such as Ovid, EbscoHost and Proquest.

You can also see our Libguide on Literature searching to get you going.