Systematic review with network meta-analysis
How do several competing interventions compare across an evidence network?
When to use this project
Use this starter when a review seeks comparisons among multiple interventions, including comparisons with little or no direct evidence. Specify the treatment nodes and eligible comparisons before fitting a model.
How it differs
Network meta-analysis adds indirect evidence to direct comparisons under additional assumptions. A connected diagram is necessary for connected-network estimation but does not establish transitivity.
Decisions before you start
- Review effect modifiers across comparisons before combining evidence. Participants should be plausibly eligible for the interventions being compared.
- Account for multi-arm correlation; distinguish within-comparison heterogeneity from disagreement between direct and indirect evidence.
- Report comparison-specific effects and uncertainty. Rankings alone do not establish the clinically best treatment.
How the flows work together
Each flow below links to its diagram, steps and tables. Flow order organizes the project; it does not imply that all work is strictly sequential. Data assessment, analysis planning and protocol refinement may inform one another.
1. Network review overview
Overview · Question through network comparison
Use this overview to state the purpose and link to detailed decisions. Keep a versioned protocol and decision register, and add evidence references for the rationale.
2. Search and study selection
Selection · Identify eligible treatment comparisons
Keep records, reports and studies as separate counting units. The starter partition covers duplicate removal only: add explicit branches for automation or other removals if used. Supplementary searches may need separate identification routes. A final report-to-study mapping is not an exclusion count.
3. Extraction and network definition
Data · Define treatments and comparison geometry
Distinguish treatment nodes in the statistical network from the boxes in this authoring canvas. Specify intervention groupings and effect modifiers; separate disconnected components instead of forcing comparisons between them.
4. Network synthesis plan
Statistics · Specify direct and indirect comparisons
The comparison identity explains direction on an additive scale. The analysis still needs model, heterogeneity, multi-arm covariance, uncertainty and diagnostic specifications. Rankings are optional outputs.
5. Coherence, bias and certainty
Statistics · Assess assumptions and evidence confidence
Use transitivity for the comparability assumption and coherence/incoherence for direct–indirect agreement. Specify the confidence framework and assessment level. A nonsignificant test is not evidence that assumptions are satisfied.
Equations and statistical interpretation
These explanations describe notation and assumptions. Flow does not calculate these values.
Indirect comparison on an additive scale
dXY means the effect of Y relative to X on the same additive analysis scale (for example log odds ratio). A is a common comparator.
This identity expresses the consistency model; it is not proof that transitivity holds. Uncertainty and correlated comparisons must be handled by the fitted model.
Published example
This Lancet systematic review includes placebo-controlled and head-to-head trials in a network. Its treatment definitions and two outcome domains provide useful examples of network planning.
This publication illustrates a study design, not an endorsement of every method or a reproduction of its results. Follow its protocol, supplement and update history when studying the example.
Hypothetical scenario
A guideline panel wants to compare three exercise programmes and usual care for a common depression outcome.
- Network review overview: Define the target population, outcome and assessment time; register the review where applicable.
- Search and study selection: Find studies across all eligible comparisons, not only those involving a preferred intervention.
- Extraction and network definition: Define treatment nodes and inspect effect modifiers, multi-arm designs and disconnected components.
- Network synthesis plan: Specify a network model, dependence handling and planned relative-effect displays.
- Coherence, bias and certainty: Assess transitivity, coherence, sensitivity to node definitions, and confidence for each decision-relevant comparison.
Deliver a populated protocol, decision log, data specification, and planned outputs. Enter numerical findings only after analysis and verification. Keep decisions made before seeing results distinguishable from later amendments.
Sources
- Cochrane Handbook, Chapter 11: Network meta-analysis
- PRISMA-NMA: Network meta-analysis
- PRISMA 2020 flow diagrams: identification and selection
Source check: 2026-09-17. Record the version of each reporting or appraisal tool used in your actual protocol.
Compare all Evidence synthesis project types · Open editor in a new tab