Transporting stochastic direct and indirect effects to new populations

Area of Science:

• Causal inference
• Biostatistics
• Health services research

Background:

• Understanding intervention effectiveness across diverse populations is crucial.
• Transported mediation effects explain differential intervention impacts.
• Existing statistical methods lack estimators for transported effects.

Purpose of the Study:

• To propose novel, doubly robust, and efficient estimators for transported stochastic direct and indirect effects.
• To evaluate the performance of these new estimators using simulation studies.
• To apply the preferred estimator to real-world data for enhanced intervention analysis.

Main Methods:

• Development of two doubly robust, efficient estimators for transported stochastic direct and indirect effects.
• Simulation study to assess finite sample properties of the proposed estimators.
• Application of the preferred substitution estimator to longitudinal data from the Moving to Opportunity Study.

Main Results:

• The proposed estimators demonstrate reliable performance in simulation studies.
• The study successfully transported stochastic indirect effect estimates across different sites.
• Analysis revealed insights into drivers of site-specific differences in intervention outcomes.

Conclusions:

• The developed estimators provide a robust framework for analyzing transported mediation effects.
• This methodology enhances understanding of intervention generalizability and adaptation.
• Findings from the Moving to Opportunity Study highlight the utility of these estimators in public health research.