Utilizing radar graphs in the visualization of simulation and estimation results in network meta-analysis

Area of Science:

• Statistics
• Medical Informatics
• Biostatistics

Background:

• Traditional forest plots in meta-analysis visualize pairwise treatment effects.
• Network meta-analysis (NMA) involves complex comparisons across multiple treatments.
• Visualizing NMA results against a fixed reference value presents unique challenges.

Purpose of the Study:

• To introduce radar graphs as an effective visualization tool for NMA.
• To illustrate comparisons of estimated results or performance measures against a predefined reference value.
• To enhance the understanding of complex, high-dimensional data in NMA and simulation studies.

Main Methods:

• Proposed utilizing radar graphs for visualizing NMA results.
• Applied radar graphs to compare estimated treatment contrasts and simulation study performance measures to a fixed reference value.
• Demonstrated the ability to capture the full network picture efficiently.

Main Results:

• Radar graphs provide a comprehensive overview of NMA results in a single visualization.
• This method effectively handles complex, high-dimensional data structures.
• Facilitates discussion of results, especially in large simulation studies with multiple scenarios.
• Can incorporate additional information like Monte-Carlo error and network connectivity.

Conclusions:

• Radar graphs are a valuable tool for visualizing network meta-analysis results, particularly when comparing against a reference value.
• This visualization method improves the interpretation of complex NMA data and simulation outcomes.
• Offers a space-efficient way to display multifaceted results, aiding in research discussions and comparisons.