Component Network Meta-Analysis (CNMA)
Component Network Meta-Analysis (CNMA) is a method that compares complex treatments made up of multiple components by estimating the individual effects of each component and how they combine. It is useful when treatments are combinations of simpler parts, such as drugs or behavioral therapies.
At a glance
Use when
Analyzing multicomponent interventions where component contributions are of interest, comparing treatments that share subcomponents, working with disconnected networks containing common elements
Avoid when
Component effects are highly context-dependent or non-additive, insufficient data on component combinations, when component definitions are inconsistent across studies
Inputs
Treatment arms composed of multiple components, effect sizes from comparative studies, network structure of component combinations
Outputs
Estimated effect sizes for individual components, interaction effects (if modeled), overall treatment effects based on component composition
How it works
Component Network Meta-Analysis (CNMA) extends standard network meta-analysis by modeling treatment effects as functions of shared components. The additive CNMA model assumes treatment effects are the sum of individual component effects, while interaction models allow for component interactions. Parameters are estimated using weighted least squares regression within a frequentist framework, implemented in the R package netmeta. CNMA can be applied even in disconnected networks if subnetworks share common components, enabling indirect comparisons across networks.
- HTA domains
- Clinical Effectiveness
- Categories
- Evidence SynthesisIndirect Comparisons
- Assumptions
- In additive models: treatment effect is sum of component effects; in interaction models: specific interactions between components can be estimated; common components have consistent effects across different combinations
- Strengths
- Enables decomposition of complex interventions into component effects, allows comparisons across disconnected networks via shared components, can be implemented using frequentist methods in R (netmeta package)
- Limitations
- Relies on assumption of additivity or specified interaction forms, component effects may not be generalizable across contexts, requires sufficient data on overlapping components
- Also known as
- CNMA, Component NMA, Component-based Network Meta-Analysis
Questions this answers
- › What are the individual effects of components within complex interventions?
- › How do treatment combinations compare when they share common elements?
- › Can indirect comparisons be made across disconnected networks using shared components?
- › What is the additive contribution of each component to overall treatment effect?
- › How do interaction effects between components influence outcomes?
- › Is a frequentist approach to CNMA feasible and valid?
References & sources
Similar by meaning
Beta record. Generated from the primary source via AI extraction and independent audit, pending final human review.