In this crucial step, retrieved articles are screened and sorted using the predetermined inclusion and exclusion criteria. The risk of bias for each included study is also assessed around this time (see below).
It's best if you import search results into a citation management tool to clean up the citations and remove any duplicates. You can then use a tool like Covidence or Rayyan to screen the results.
You should begin by screening titles and abstracts only, and then you'll examine the full text of any remaining articles. Each study should be reviewed by a minimum of two people on the project team.
Each included study is coded and the quantitative or qualitative data contained in these studies is then synthesized. You'll have to either find or develop a coding strategy or form that meets your needs.
An important part of conducting a systematic review or meta-analysis is the risk of bias or quality assessment. During this step of the process, authors identify, evaluate, and report any biases in the included studies. All studies are susceptible to what are known as "systematic errors" in their design and execution that can bias the results. You can skip this step if you're conducting a scoping or systematized, rather than systematic, review.
Risk of bias tools provide reviewers with a standard set of criteria – in the form of questions or statements – against which they can judge a study they’re looking at and determine if any biases crept into the results, as well as how these biases might have impacted the study's internal validity. The tools are usually methodology-specific.
Finding validated risk of bias tools outside of the health and behavioral sciences can be a challenge, particularly for non-experimental methodologies. Searching subject databases or well-known journals in your field for systematic reviews is typically a good first step to take when trying to find an appropriate tool. Published systematic reviews on topics that are similar to yours will cite quality assessment tools that you can find and use yourself.
The “Quality Assessment and Risk of Bias Tool Repository,” is another a good resource for finding tools, although it is centered on the health sciences and may be less useful to those in other disciplines.
Most teams use custom-made software to screen and manage studies found during the course of conducting a systematic review.
Covidence is recommended for those intending to undertake a formal, publishable review. It is not free, but those affiliated with the University can request an account through the Libraries. Visit the Covidence tab/page on this guide for more instructions on how to do so.
Rayyan is free but is missing many of the project management features built into Covidence. It is recommended for graduate students learning about the systematic review process or for individuals conducting traditional reviews.
