Multicriteria Decision-Making Methods for Optimal Treatment Selection in Network Meta-Analysis.

Area of Science:

• Medical research methodology
• Decision science

Background:

• Network meta-analysis (NMA) is crucial for comparing multiple interventions using randomized data.
• Selecting optimal treatments in NMA becomes challenging with multiple, conflicting outcomes.

Purpose of the Study:

• To integrate multicriteria decision-making (MCDM) methods into NMA for selecting the best intervention.
• To address the complexity of choosing treatments when multiple outcomes are considered.

Main Methods:

• Employed TOPSIS, VIKOR, and PROMETHEE algorithms for partial and absolute rankings.
• Utilized objective (entropy method) and subjective (analytic hierarchy process) weighting for outcomes.
• Explored alternative preference functions within the PROMETHEE method.

Main Results:

• Demonstrated high agreement among TOPSIS, VIKOR, and PROMETHEE II methods in a simulated study.
• Showcased visualization techniques like score-adjusted scatterplots and PROMETHEE I league tables.
• Confirmed method agreement increases with a higher number of interventions.

Conclusions:

• MCDM offers a flexible and direct approach for selecting compromise interventions in NMA.
• TOPSIS, VIKOR, and PROMETHEE are effective for optimal treatment selection with conflicting outcomes.
• Further research is needed to validate MCDM in real-world NMA settings.