Federated Causal Inference: Multi-Study ATE Estimation beyond Meta-Analysis

TL;DR

This paper introduces federated causal inference methods for estimating treatment effects across decentralized data sources, comparing different federated estimators and providing guidance for their application in RCTs.

Contribution

It proposes and analyzes federated estimators for ATE, including simple meta-analysis and multi-shot federated learning, with theoretical and empirical evaluation.

Findings

Multi-shot federated estimator leverages full data for better accuracy.

Asymptotic variance formulas guide estimator selection in heterogeneous settings.

Simulation validates theoretical insights and practical recommendations.

Abstract

We study Federated Causal Inference, an approach to estimate treatment effects from decentralized data across centers. We compare three classes of Average Treatment Effect (ATE) estimators derived from the Plug-in G-Formula, ranging from simple meta-analysis to one-shot and multi-shot federated learning, the latter leveraging the full data to learn the outcome model (albeit requiring more communication). Focusing on Randomized Controlled Trials (RCTs), we derive the asymptotic variance of these estimators for linear models. Our results provide practical guidance on selecting the appropriate estimator for various scenarios, including heterogeneity in sample sizes, covariate distributions, treatment assignment schemes, and center effects. We validate these findings with a simulation study.