Sensitivity analysis for transportability in multi-study, multi-outcome settings

TL;DR

This paper explores methods to generalize causal effects across multiple studies with missing primary outcomes, using proxy variables and sensitivity analysis to evaluate assumptions and improve estimation efficiency.

Contribution

It introduces a novel approach for transporting causal effects in multi-study, multi-outcome settings using proxies and develops sensitivity analysis methods for the key assumptions.

Findings

Proposes a new identification strategy leveraging proxy outcomes.

Develops sensitivity analysis techniques for the assumptions.

Enhances efficiency of causal estimators in heterogeneous data sources.

Abstract

Existing work in data fusion has covered identification of causal estimands when integrating data from heterogeneous sources. These results typically require additional assumptions to make valid estimation and inference. However, there is little literature on transporting and generalizing causal effects in multiple-outcome setting, where the primary outcome is systematically missing on the study level but for which other outcome variables may serve as proxies. We review an identification result developed in ongoing work that utilizes information from these proxies to obtain more efficient estimators and the corresponding key identification assumption. We then introduce methods for assessing the sensitivity of this approach to the identification assumption.