Beading plot: a novel graphics for ranking interventions in network evidence.

Area of Science:

• Biostatistics
• Health Informatics
• Clinical Epidemiology

Background:

• Network meta-analysis (NMA) synthesizes evidence from multiple treatment comparisons, crucial for clinical decision-making.
• The complexity of NMA can be challenging for healthcare providers and patients to interpret.
• Effective visualization is needed to present complex treatment rankings from NMAs.

Purpose of the Study:

• To introduce a novel graphic design, the "beading plot", for comprehensive presentation of complex treatment rankings in network meta-analysis.
• To enhance the interpretability of NMA results for clinical decision-making.

Main Methods:

• Development of the "beading plot" visualization tool.
• Implementation of the beading plot using the "rankinma" R package, available on CRAN.
• The `PlotBead()` function generates beading plots summarizing treatment rankings across diverse outcomes.

Main Results:

• The beading plot effectively displays collective metrics (e.g., probability of being best treatment (P-best), surface under the cumulative ranking curve (SUCRA), P-score) for each treatment.
• Visual representation uses a number line plot, with color-coded beads for treatments and continuous lines for outcomes.
• Treatment rankings are ordered on the y-axis based on summary metrics.

Conclusions:

• The beading plot offers an intuitive visualization for treatment rankings in complex network meta-analyses.
• Enhances reader-friendliness and aids clinicians and patients in identifying optimal treatments.
• Recommends cautious use alongside an assessment of overall evidence certainty.