Identifying Conditions Favouring Multiplicative Heterogeneity Models in Network Meta-Analysis

TL;DR

This paper compares multiplicative-effect and additive random-effects models in network meta-analysis, finding the multiplicative model often offers better fit and robustness, especially with heterogeneity and publication bias.

Contribution

It introduces the multiplicative-effect model as a viable alternative to the conventional additive model in network meta-analysis and evaluates its performance empirically.

Findings

ME model often fits data better than RE model.

ME model is more robust to outliers and publication bias.

RE models are sensitive to extreme observations.

Abstract

Explicit modelling of between-study heterogeneity is essential in network meta-analysis (NMA) to ensure valid inference and avoid overstating precision. While the additive random-effects (RE) model is the conventional approach, the multiplicative-effect (ME) model remains underexplored. The ME model inflates within-study variances by a common factor estimated via weighted least squares, yielding identical point estimates to a fixed-effect model while inflating confidence intervals. We empirically compared RE and ME models across NMAs of two-arm studies with significant heterogeneity from the nmadb database, assessing model fit using the Akaike Information Criterion. The ME model often provided comparable or better fit to the RE model. Case studies further revealed that RE models are sensitive to extreme and imprecise observations, whereas ME models assign less weight to such observations and hence exhibit greater robustness to publication bias. Our results suggest that the ME model warrant consideration alongside conventional RE model in NMA practice.