Systematic reviews

The stages of a systematic review may vary slightly depending on the discipline and study types to be included but generally follow a series of steps outlined as follows:

1. Clarifying the problem, defining the research question and its boundaries, and setting out criteria for studies to be included in the review. This information, along with a plan for carrying out the review, should be presented in a protocol.
2. Identifying studies, including selecting appropriate sources and searching for studies.
3. Screening and selecting the studies, collecting and presenting the data from the studies, and / or describing the methods and findings of the studies using a structured approach.
4. Appraising the relevance and quality of each study, which may involve assessing the risk of bias.
5. Analysing and synthesising the data from the studies in order to answer the research question, which may involve a meta-analysis.
6. Assessing reporting biases.
7. Presenting results.
8. Interpreting results and presenting conclusions.

Training and support available from UCL Library Services primarily focuses on the 'Searching for studies' stage of the process, but may include other stages in some circumstances.

Publishing a protocol is not mandatory if you are writing a review, but it is increasingly considered best practice. The PRISMA statement includes a checklist to support the development of protocols and you may wish to consider publishing a protocol on one of the sites listed below.

The JBI recommends that protocols are published in PROSPERO. The JBI Manual for Evidence Synthesis includes information about how to write protocols for scoping reviews, umbrella reviews, and various types of systematic review.

The following resources provide more in depth information about the stages of a systematic review:

See also Useful books about systematic reviews.