Choosing an evidence synthesis approach
Choose the review process for your question, then decide whether and how the findings can be combined.
Fit
When it matters, and when it does not
It matters when
- You must decide whether to map a field or answer a focused evidence question.
- You need to explain why pooling is appropriate, limited or omitted.
It matters less when
- For a question requiring new participant data, choose a primary study instead. Read: Choosing a study design from the research question
Before you start
What to check in your study
- Review purposeChoose mapping, effect estimation, experience or another explicit aim.
- Unit of evidenceDecide whether to review primary studies, existing reviews or individual participant data.
- Feasible scopeSpecify search coverage, eligible designs and available appraisal time.
- The search and selection match the review purpose.How to check: Use explicit eligibility rules and record exclusions.
- Any pooled studies address a compatible target.How to check: Compare populations, interventions, outcomes, follow-up and risk of bias before pooling.
Plain language
How it works
A systematic review uses explicit methods to find, select and appraise evidence for a stated question. A scoping review maps concepts, evidence types and gaps. Grant and Booth describe review types; Tricco and colleagues set out scoping review reporting. A rapid review uses declared shortcuts, which may reduce coverage.
An overview reviews existing reviews and must handle overlapping studies. Qualitative synthesis brings together findings about experience or meaning. Meta-analysis is a statistical step that may sit inside a systematic review. Pairwise pooling compares the same contrast; network synthesis compares connected alternatives under stronger assumptions. Individual participant data synthesis changes the data available, not the need for a sound review.
Worked example
A worked example
- Setting
- A team finds 10 eligible trials after a systematic search.
- Question
- Can all trials support one pooled treatment effect?
- Approach
- Four trials address the same outcome and horizon. Six use different questions or incompatible outcome definitions.
- Result
- The comparable subset is 4 of 10 trials, or 40 in 100. The other 6 remain in the review with a structured explanation.
- What it means
- Eligibility does not guarantee compatibility. Decide whether to pool the four after checking bias and clinical differences; report why the six were not pooled.
In practice
Recent studies PubMed search · not vetted
Watch out
Common mistakes
The one to remember
A systematic review need not contain a meta-analysis. Pooling cannot make incompatible questions comparable. A scoping review maps evidence and does not by itself establish treatment effectiveness.
- Treating systematic review and meta-analysis as synonyms.Instead: Name the review process and the synthesis separately.
- Pooling because the software accepts the inputs.Instead: Check the clinical contrast before statistical heterogeneity.
- Calling a rapid search exhaustive.Instead: Declare each shortcut and its possible effect on coverage.
Inside Mallard
How Mallard uses this
Mallard includes this comparison in Learn search and the Atlas. Plans do not link to this comparison automatically.
concept.choosing_evidence_synthesis_types, the id a plan cue records.Related Learn pages
Related designs and methods
Evidence
Sources & further reading
- A typology of reviews: an analysis of 14 review types and associated methodologies (opens in a new tab) · Health Info Libr J · doi 10.1111/j.1471-1842.2009.00848.x
- PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation (opens in a new tab) · Ann Intern Med · doi 10.7326/M18-0850