Skip to Main Content
It looks like you're using Internet Explorer 11 or older. This website works best with modern browsers such as the latest versions of Chrome, Firefox, Safari, and Edge. If you continue with this browser, you may see unexpected results.

Systematic Reviews in the Health Sciences

A guide to systematic reviews in the health sciences.

Your Literature Search

The most important thing to note is that your literature search is your data collection method for your review. As such, your search must be comprehensive, unbiased, documented, and reproducible. Meaning if someone wanted to, they could do the same thing you’re doing.

Before you even start your search, you’ll want to identify potential data sources. These include:

  • Research databases (like Ovid MEDLINE or PubMed)
  • Reference list searching (you’ll want to pay special attention to the references of studies you’ll be including in your review)
  • “Cited by” reference searching (again, looking at who has cited the studies you end up selecting for your review 
  • Contacting experts regarding research in progress
  • Published relevant systematic reviews (these can be particularly helpful for developing your search strategy)
  • Grey literature

Research Databases: 

There are many available through the University of Calgary Library.

To choose relevant databases, you’ll need to identify which disciplines inform your research. For example, if your research question is interdisciplinary, you might want to include a interdisciplinary database like Web of Science or Scopus in your search.

Keep in mind that you should only search one database at a time. Typically you’ll start with the most rigorous database, like Ovid MEDLINE, and then translate your search into other databases.

The tricky thing is that different database interfaces require different search techniques and what works in one might not work in another. You can meet with a librarian to discuss searching within a specific database. The University of Toronto Library has created great "how to" videos which you can also use to learn how to search within the different interfaces, such as OVID Medline, CINAHL, or Scopus

Some of the most common databases searched within systematic reviews are: 

  • OVID Medline
  • OVID Embase
  • Cochrane Central Register of Controlled Trials
  • APA PsychINFO
  • CINAHL
  • Scopus 
  • Web of Science
  • CAB Abstracts
  • ERIC 

Building your Literature Searc

We touched on this in step one, but this is when you'll build out your search strategy.

  • Identify the key concepts to be searched using the PICO framework
  • Consider whether key concepts could be represented differently based on the discipline
  • Brainstorm related terms, synonyms, and sub-concepts
  • Determine whether each database you are searching includes subject headings
  • Review known relevant articles (seed articles) for potential keywords and subject headings
  • Review search strategies from relevant systematic reviews
  • Consult a librarian 

It takes a LOT of time to develop a good search strategy and it is an iterative process. Once you’ve finalized your search strategy, you’ll want to pre-test it in one or two databases to make sure it retrieves known relevant articles, or articles that you already know you’re going to include in your review, like the seed articles you previously found. If it doesn’t, you should check what subject headings those articles are indexed with and what keywords they use in their abstract and in the article to express your PICO concepts.

Along the way, you should be saving your search strategies in each database, noting the database name, date, and version. Remember that some interfaces like Ovid have multiple databases, so noting the database name will help you identify which search to run in each database later on.

Once you’ve finalized your searches, you’ll run them once more, documenting the date of your search, the databases searched, your exact search strategy, the number of results in each database, and the number of duplicates. You’ll need this information to complete the PRISMA flow chart. 

Grey Literature: 

Grey literature is basically any literature that has not been published through traditional means.

It’s literature that you won’t typically find in large databases or other mainstream sources. Grey literature includes government reports, technical reports, unpublished clinical trials, program evaluation reports, committee reports, theses/dissertations, conference papers/abstracts, and standards/best practices. When searching for grey literature, you’ll need to consider what organizations produce literature on your topic. These can include government agencies like the Public Health Agency of Canada, professional societies like the Canadian Medical Association, and advocacy groups. You can also use CADTH's Grey Literature finding tool called "Grey Matters" as a starting point for your grey literature search. 

Other sources include: 

  • Conference websites
  • Clinical trial registries 
  • Canada Commons database
  • OpenGrey
  • ProQuest Dissertations and Theses Global 

Including grey literature in a systematic review helps to avoid reporting and publication bias. Reporting bias occurs when studies are not reported or published because of their negative results. Publication bias occurs when only formally published literature is included in a systematic review (for example, only peer-reviewed journal articles). 

Data Collection

Once you’ve done all of your searching, you can either use a citation management tool, like EndNote, to export the references, or you can import the references straight into Covidence which is a tool the university licenses for systematic reviews. 

This tool is free for UCalgary affiliates when you access it through the UCalgary databases. And you can still use Covidence if you’re on a review team with people from other institutions, as long as one member of the team is affiliated with the University of Calgary. 

Covidence can be used to: 

  • Remove duplicates from imported records. This is important because databases like MEDLINE and EMBASE index some of the same journals which will always result in some duplicates. Often the more databases you search, the more duplicates you’ll end up seeing.
  • Screen titles and abstracts
  • Screen full text 
  • Assess risk of bias 
  • Extract data

Follow this link to request access to Covidence through the University of Calgary.