In the most restrictive sense, a systematic review is a type of evidence synthesis that is particularly rigorous in the finding, selecting, evaluation, and analyzing of research.
The rigorous nature of a systematic review means that it is not always an appropriate (or possible) methodology for all evidence synthesis. The ideal systematic review includes:
Some of the best known systematic reviews are produced in the healthcare field in the Cochrane Library of Systematic Reviews (SCSU login for subscription access). They have an extremely rigorous process and require training, as documented in the Cochrane Handbook.
JBI is another group specializing in systematic reviews and other types of healthcare evidence synthesis. Besides documentation on methodology, they also have several video series on defining and conducting high quality evidence synthesis projects.
A newer group is the Campbell Collaboration, which produces systematic reviews (and other types of evidence syntheses) in Business & Management, Crime & Justice, Disability, Education, International Development (including Nutrition), Knowledge Translation & Implementation, and Social Welfare.
The Institute for Education Science (out of the US Dept of Education) maintains the What Works Clearinghouse, which compiles reviews into educational intervention reports and practice guides following specific protocols.
That said, not all fields differentiate between systematic reviews and other types of evidence syntheses, calling everything systematic reviews. And the definitions have changed even within the healthcare field, where evidence syntheses have been formally used the longest.
Steps in a Traditional Systematic Review | Estimated Time Investment | |
---|---|---|
1. Assemble systematic review team and select project manager | Varies | |
2. Identify appropriate review methodology | 2 weeks | |
3. Define research question | 2 weeks | |
4. Define inclusion/exclusion criteria | 1 week | |
5. Select databases | 1 week | |
6. Select gray literature resources | 1 week | |
7. Write search strategy for “master” database | 1 week | |
8. Write and register protocol (written compilation of previous steps) | Varies | |
9. Translate search strategy to syntax for all databases (including gray literature) | 2 weeks | |
10. Search and export results into citation management software | 2 weeks | |
11. De-duplicate results | 2-4 weeks | |
12. Title and abstract screening | 2–3 months* | |
13. Retrieve full-text articles | 1 month* | |
14. Full-text screening | 2–3 months* | |
15. Risk-of-bias assessment | 2–3 months | |
16. Data extraction | 2–3 months | |
17. Meta-analysis or synthesis of results | 2–3 months | |
18. Write the manuscript | 2–3 months |
* Timeframe can vary significantly depending on number of citations identified for screening.
From: Evidence Synthesis: How Librarians Can Help from Cornell University Library (CC-BY 4.0)
Questions to think about:
If your answers are No, then you may want to do some other type of review. For semester-long projects, a Rapid review or a systematized/structured review (both follow SR techniques but with streamlined procedures.)