Interpretable meta-analysis of model or marker performance

TL;DR

This paper develops interpretable meta-analysis methods for model performance across heterogeneous populations, ensuring estimates are relevant to a specific target population, with validation through simulations and lung cancer trial data.

Contribution

It introduces new identifiability conditions, inverse-weighting, outcome modeling, and doubly robust estimators for meta-analysis in heterogeneous settings.

Findings

Methods produce interpretable performance estimates for target populations.

Simulations validate the robustness of the proposed estimators.

Application to lung cancer trials demonstrates practical utility.

Abstract

Conventional meta analysis of model performance conducted using datasources from different underlying populations often result in estimates that cannot be interpreted in the context of a well defined target population. In this manuscript we develop methods for meta-analysis of several measures of model performance that are interpretable in the context of a well defined target population when the populations underlying the datasources used in the meta analysis are heterogeneous. This includes developing identifiablity conditions, inverse-weighting, outcome model, and doubly robust estimator. We illustrate the methods using simulations and data from two large lung cancer screening trials.