Bayesian unanchored additive models for component network meta-analysis

2025-07-21

AI Summary

Conventional additive assumptions in component network meta-analysis (CNMA) for multi-component interventions are often overly restrictive, leading to model misspecification. Method: This paper proposes a Bayesian anchor-free additive CNMA framework. It establishes, for the first time, a unified theoretical framework that explicitly distinguishes anchor-based from anchor-free modeling paradigms, clarifying fundamental differences in additivity assumptions across existing methods. Two novel Bayesian anchor-free CNMA models are introduced, relaxing stringent additivity constraints to enhance model flexibility and applicability. Contribution/Results: MCMC-based simulations demonstrate that the proposed models yield lower estimation bias, higher confidence interval coverage probabilities, and improved treatment ranking accuracy compared to existing approaches. Empirical analyses on real-world data further confirm their robustness and practical utility.

Technology Category

• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Reasoning under Uncertainty: Relational Probabilistic Models
• Multiagent Systems: Adversarial Agents

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Social Networks and Social Media: Fairness and bias in social network and social media analysis

Abstract

Component network meta-analysis (CNMA) models are an extension of standard network meta-analysis (NMA) models which account for the use of multicomponent treatments in the network. This article contributes innovatively to several statistical aspects of CNMA. First, by introducing a unified notation, we establish that currently available methods differ in the way they assume additivity, an important distinction that has been overlooked so far in the literature. In particular, one model uses a more restrictive form of additivity than the other which we term an anchored and unanchored model, respectively. We show that an anchored model can provide a poor fit to the data if it is misspecified. Second, given that Bayesian models are often preferred by practitioners, we develop two novel unanchored Bayesian CNMA models presented under the unified notation. An extensive simulation study examining bias, coverage probabilities, and treatment rankings confirms the favorable performance of the novel models. This is the first simulation study to compare the statistical properties of CNMA models in the literature. Finally, the use of our novel models is demonstrated on a real dataset, and the results of CNMA models on the dataset are compared.

Problem

• Distinguishes anchored and unanchored additive models in CNMA
• Develops novel Bayesian unanchored CNMA models for practitioners
• Compares CNMA model performance via simulation and real data

Innovation

• Introduces unified notation for CNMA models
• Develops novel unanchored Bayesian CNMA models
• Conducts first simulation study comparing CNMA models