🤖 AI Summary
In evidence synthesis, inconsistent interpretation and poor transportability of marginal versus conditional summary measures induce aggregation bias. Method: We systematically analyze their causal interpretational differences across outcome types and data-generating mechanisms, integrating causal inference theory with covariate-adjustment strategies in indirect comparisons. Contribution/Results: We demonstrate that even for collapsible measures, effect modification can render marginal and conditional effects non-equivalent; moreover, covariates traditionally deemed non-effect-modifiers may modify treatment effects at the population level. We develop a formal transportability framework grounded in causal inference and empirically validate it via indirect comparisons. Our analysis clarifies the mechanistic origins of bias and establishes individual patient data (IPD) as critical for ensuring compatibility of summary measures across studies. This work provides both theoretical foundations and methodological guidance for cross-study evidence integration.
📝 Abstract
Marginal and conditional summary measures do not generally coincide, have different interpretations and correspond to different decision questions. While these aspects have primarily been recognized for non-collapsible summary measures, they are also problematic for some collapsible measures in the presence of effect modification. We clarify the interpretation and properties of different marginal and conditional summary measures, considering different types of outcomes and hypothetical outcome-generating mechanisms. We describe implications of the choice of summary measure for transportability, highlighting that covariates not conventionally described as effect modifiers can modify population-level treatment effects. Finally, we illustrate existing summary measure incompatibility issues in the context of evidence synthesis, using the case of covariate adjustment methods for indirect treatment comparisons. Because marginal and conditional summary measures do not generally coincide, their naïve pooling in evidence synthesis can produce bias. Almost invariably, care is needed to ensure that evidence synthesis methods are combining compatible summary measures, and this may be easier to ensure with full access to individual patient data.