In general, a knowledge synthesis review will involve the following steps:
Develop a focused research question
Knowledge synthesis reviews address clear and answerable research questions. They also have clear criteria about the studies that are being used to address the research questions. A well formulated question will guide the direction of the review, help identify the terms to be used in the literature search, demonstrate the relationship between the terms, and identify the outcomes being sought.
Develop inclusion/exclusion criteria
This is the criteria that will be used to determine which research studies will be included. The inclusion and exclusion criteria must be decided before you start the review. Inclusion criteria is everything a study must have to be included. Exclusion criteria are the factors that would make a study ineligible to be included. Criteria that should be considered include (but not be limited to):
- Type of studies
- Type of participants
- Types of intervention
- Types of outcome measures
Search the literature systematically
An systematic search should be exhaustive and identify all publications and as much grey literature as possible that meet study requirements. At least three databases should be searched. The searcher should document and report the exhaustive search in such a way that it can be replicated for updates and reproduced by others after publication.
A systematic search is very complex and requires expert knowledge to build and translate across databases. Research has shown that librarian involvement in knowledge synthesis reviews leads to increased quality of search strategies (source) and reporting (source; source).
Screen articles for eligibility
There are two stages to the screening process, as identified in the PRISMA flow diagram.
Stage 1: Title and abstract screening
After removing duplicates, as least 2 reviewers independently screen all of the titles/abstracts identified in the search against the inclusion/exclusion criteria. Reviewers should be over-inclusive at this stage; a maybe is always a yes at this point.
Stage 2: Full text screening
Acquire all of the full texts identified in stage 1. Two reviewers will independently screen each article based on the studies selection criteria. Only Yes's are included in the review but all No's must have a reason listed for exclusion
Data Extraction
Data extraction is conducted using data extraction forms and independently/blinded by two reviewers, with a 3rd reviewer serving as a tiebreaker. Commonly extracted fields for most SRs include article citation information, study characteristics, participant characteristics, interventions and setting, and outcome data & results.
Synthesize and report
This step will vary depending on what type of review you are conducting, but may include risk of bias assessment, critical appraisal, and/or meta-analysis.