Skip to main content

Systematic Reviews

Process for developing a comprehensive search strategy

1. Identify the major concepts from your research question or topic.

Highlighting keywords for searching

2. List synonyms or alternative terms for each concept and organise them in a table like the one below - using a column for each major concept. Tools and tips to assist with this process: 

  1. Text mining tools including PubMed Reminer especially if you are using a database with MeSH such as Medline or Cochrane. There are many others however.
  2. Run scoping searches for your topic in databases such as Google Scholar or use the "Basic Search" in OVID databases to identify how the literature can express your concepts. Scan titles, subject headings and abstracts for words describing the same things as your major concepts.
  3. As you find something new, add it to the appropriate column on your list to incorporate later in your search.

 

     
     
     
     
     
     
     

 

3. Validated Search Filters or "hedges" can help to save time. These can help limit search strategies to specific study designs or eliminate Animal studies from results. There is always some risk however that you may eliminate good results. 

4. Be prepared to revise, reassess and refine your search strategies after you have run your initial searches to ensure you get the best possible results. If you retrieve too many false results or "noise", try to analyse why. For example, you may have used a word which has alternative meanings. Your Clinical or Research Librarian will be able to assist with this process.

Developing Inclusion and Exclusion Criteria

  • Besides the search strategy itself the inclusion and exclusion criteria comprise the parameters or scope of the Systematic Review.
  • These criteria assist with being consistent with screening results and usually some of the finer points which might be difficult to incorporate in the search, or if they were included they might eliminate some useful results.
  • It is useful to consider them at this stage, but they could be developed later in Step 6, as they will be obvious once you start screening through the results – essentially, what should be there and what shouldn’t.