Network Meta-Analysis.

Area of Science:

• Biostatistics
• Evidence Synthesis
• Clinical Epidemiology

Background:

• Multiple interventions often exist for a single disease, necessitating comparative effectiveness research.
• Pairwise meta-analysis compares interventions head-to-head using direct trial data.
• Network meta-analysis (NMA) extends pairwise methods to compare multiple interventions simultaneously, including those not directly studied against each other.

Purpose of the Study:

• To provide a comprehensive overview of network meta-analysis (NMA) methodology.
• To elucidate the complexities and applications of NMA in biomedical research.
• To guide researchers in conducting, interpreting, and critically appraising NMAs.

Main Methods:

• Exploration of similarities and differences between pairwise and NMA.
• Explanation of direct, indirect, and mixed treatment comparisons.
• Description of NMA model assumptions (connectivity, homogeneity, transitivity, consistency) and outputs (network plots, SUCRAs).

Main Results:

• NMAs facilitate simultaneous comparison of multiple interventions, expanding the scope beyond head-to-head trials.
• NMAs address previously unanswerable research questions, leading to increased publications.
• Conduct and interpretation of NMAs are more complex than pairwise meta-analyses, requiring understanding of specific assumptions and outputs.

Conclusions:

• NMA is a powerful tool for comparative effectiveness research, enabling simultaneous ranking of multiple interventions.
• Understanding NMA assumptions and outputs is crucial for accurate interpretation and application.
• This chapter provides a framework for conducting and appraising systematic reviews incorporating NMA.