- Form a research team. One person cannot do a systematic review. You will need a minimum of 2 people. At least one team of two individuals is required to do independent screening of abstracts and full text as well as data extraction; be prepared to calculate inter-rater reliability statistics. Prepare your research question as a team.
- If you would like assistance, please email our librarian and fill out this contact form.
- Do a bit of preliminary, unofficial searching to get a sense of how much is out there on your topic and what kinds of studies are available. You may need to narrow or broaden your topic. Having a few pre-identified key studies can help with finding search terms and developing search strategies. Read example reviews online to see what a systematic review looks like in practice. They will provide you with a more in-depth understanding of the method generally.
- Register protocol plan on PROSPERO, run by the National Institute for Health Research and the University of York. This will help you lay out your process in detail and ensure that no other research teams are conducting a similar study. Every protocol registration process will have a set of reporting guidelines.
- Inclusion/exclusion criteria
- Create a reproducible search strategy: Key terms? Date filters? What other filters will be included? Limits on Human subjects? Article types? Age? Language? Publication dates?
- Data extraction process
- Assessment of study quality process
- Data synthesis
- Result dissemination plans
- Conduct literature search using multiple databases
- Limiting the review to only a few databases tends to bias the review. "Casting a wide net" in order to capture more citations across different databases reduces bias. It is important to search three or more databases in order to limit bias. The sources may overlap, but will be deduplicated towards the end of the review (step 8).
- Select studies based on inclusion/exclusion criteria
- Title & abstract screening
- Exclude papers that do not fit criteria
- Appraise studies using quality checklist with team members
- Conclusions should be based on the highest-quality evidence available
- De-duplicate results
- Perform Data Extraction
- Who is performing data extraction?
- What tools will they use?
- Identify relevant data from each paper and summarize the data using forms/tables.
- Analyze results
- Re-run search if many months have passed since the most recent search
- Synthesize results - Interpret Findings
-Bring together data from a set of included studies with the aim of drawing conclusions about a body of evidence. Synthesis of study characteristics as well as conclusions about the body of evidence.
-What were the main findings of my review?
-How do my findings fit with previously published research?
-What are the strengths and limitations of the included studies?
-What are the strengths and limitations of the review process?
-What conclusions can be drawn from this review?
Additional questions may be found on page 7.
12. Present Findings
- Were there best practices that you found that you want to highlight?
- Is there a need for further research?
- Celebrate!
