Define your research topic. The topic question should have a clear focus and may require preliminary searching of the evidence available. This is known as 'Scoping your topic' and allows you to put your research in context and justify its importance.
Develop a protocol. A protocol is a pre-defined plan for your systematic review and is developed before starting your searches. Protocols can be refined during this process.
Check registration sites. Check all the appropriate systematic review registration sites to see if your topic is already being studied. Register your protocol while you’re at it! Get an ORCID (Open Researcher and Contributor ID) if you don’t already have one.
Conduct a comprehensive search. Use a range of databases to search for published studies on your topic.
Grey literature. Use Google Advanced and grey literature sites to search for unpublished studies.
Hand searching. Involves examining manually key journals, conference proceedings and other relevant publications to overcome deficiencies in indexing or database coverage. The citation databases, Web of Science and SCOPUS are useful for identifying key journals and authors, as well as tracking research and citation searching.
Inclusion/Exclusion criteria. Will you restrict by dates, study designs, populations (age, gender, ethnicity, diseases, location, psychosocial or emotional factors), languages of references, interventions, outcomes etc.? Clear eligibility criteria will make it easier to identify relevant articles at the screening stage.
Citation management. Put the results of your searches into citation management software such as EndNote and organize each database searched into separate groups. There are other free citation management tools available e.g. CiteULike, Mendeley or Zotero.
Screening. The process in which you determine whether each individual article meets your inclusion criteria or not. Having multiple reviewers to screen results (yourself and someone else from your team) reduces bias. Various item screening tools will make this process easier e.g. Covidence.
Extract data. Extract the data that is relevant to your particular systematic review topic. Data extraction software can be used for this process.
Data synthesis/meta-analysis. Homogenous or similar data from your review can then be synthesised using a process called “meta-analysis". It uses a statistical approach to pool together results from multiple studies.