Prepare your topic. Do a preliminary search of the literature to make sure there is enough literature on your topic. Conduct preliminary searching of databases and protocol registries such as PROSPERO or OSF to see what evidence is already available and whether a systematic review has already been done, or if the research question needs updating.
Create a focused research question for your systematic review. Use the PICO framework to identify key concepts of the question. Determine inclusion/exclusion criteria for studies to be evaluated.
Establish a systematic review team Create a team of two or more to reduce the risk of bias. Select team members based on their areas of expertise including subject experts, a search strategy expert (such as a librarian) to create a comprehensive search strategy and identify relevant databases, reviewers to screen studies, and a statistician (if you are doing a meta-analysis).
Create a research protocol Create a written protocol outlining your study methodology. A protocol includes your research question(s) and the rationale for the systematic review. It should define search terms for comprehensive literature searching, inclusion and exclusion criteria of studies, critical assessment methods of individual studies, and what data will be extracted and analyzed as well as methods of analysis, etc.
Register your review protocol with a registry such as PROSPERO or OSF. Registering your protocol promotes transparency, avoids redundancy and helps you to "stake your claim" to the review topic. It also allows for peer review of the planned methods of the research. The PRISMA guidelines have a protocol checklist:PRISMA-P 2015 checklist of recommended items to address in a systematic review protocol.
Search for studies Conduct a literature search to thoroughly search for all the relevant studies on your topic. To avoid potential duplication of research, literature searches must be replicable, comprehensive and detailed. Also search the grey literature, which is information that is produced outside of traditional publishing and distribution channels, and which is often not found in indexed databases. Document your search process. Use a citation manager such as Mendeley or Zotero to manage your results. You can also use a tool specifically designed for managing systematic reviews such as JBI Sumari (you will need to create a free account).
Screen studies and select studies per your predefined protocol At least two reviewers should screen studies using the inclusion/exclusion criteria from your protocol. The first screening can be done on titles and abstracts and the second screening will review the full-text of selected studies. Keep a log of excluded studies and the the reasons they were excluded.
Analyze and synthesize data. Create a table of your study results. Analyze the data and look for variation across studies and sensitivity of findings. Consider doing a meta-analysis of your results.
Interpret results and report findings. Evaluate the strength of the evidence in your systematic review. Note any limitations or biases. Summarizing findings and conclusions based on the quality of the evidence. Use the PRISMA Checklist when reporting your findings. See Cochrane Handbook for guidance on interpreting results.
A quick overview of the steps involved in conducting a systematic review from the Evidence Synthesis Academy at Brown University: