Comparison-adjusted funnel plot
A method that adjusts funnel plots in network meta-analysis to better detect small-study effects and potential publication bias by accounting for different comparisons across studies.
At a glance
Use when
Conducting a network meta-analysis and wanting to assess potential publication bias or small-study effects across multiple interventions.
Avoid when
Analyzing a single pairwise comparison with sufficient studies where standard funnel plots are adequate.
Inputs
Effect sizes and standard errors from studies in a network meta-analysis, along with the intervention comparisons they evaluate.
Outputs
Funnel plot adjusted for comparison type, with potential asymmetry indicating small-study effects or publication bias.
How it works
Comparison-adjusted funnel plots extend traditional funnel plots by incorporating the type of intervention comparison in each study, allowing for more accurate assessment of asymmetry in networks of interventions. This method addresses limitations of standard funnel plots when few studies are available per comparison and improves detection of small-study effects across a network meta-analysis.
- HTA domains
- Clinical Effectiveness
- Categories
- HeterogeneityIndirect Comparisons
- Assumptions
- Small-study effects (e.g., publication bias) may affect an entire research field uniformly and not just isolated comparisons; the direction and magnitude of asymmetry can be modeled across comparisons.
- Strengths
- Improves interpretation of funnel plots in networks with multiple interventions,Allows visualization of asymmetry while accounting for different comparisons,More informative than standard funnel plots when few studies exist per pairwise comparison
- Limitations
- Relies on sufficient number of studies across the network for reliable assessment,Cannot definitively distinguish between bias and other causes of asymmetry,Assumption of uniform small-study effects across comparisons may not always hold
- Also known as
- comparison-adjusted funnel plot, adjusted funnel plot for network meta-analysis
Questions this answers
- › Is there evidence of small-study effects in a network of interventions?
- › Does funnel plot asymmetry vary by type of intervention comparison?
- › Could publication bias affect the results of a network meta-analysis?
- › How can we visually assess bias when multiple treatments are compared?
- › Are smaller studies showing larger effects for specific comparisons?
- › Can we distinguish between true heterogeneity and bias in a network?
References & sources
Similar by meaning
Beta record. Based on the original catalogue summary; primary-source enrichment pending.