Double Debiased Machine Learning for Mediation Analysis with Continuous Treatments
This work addresses a methodological gap for practitioners in fields like medicine who need to isolate direct and mediated effects with continuous treatments, representing an incremental improvement over existing mediation analysis techniques.
The paper tackles the problem of estimating causal mediation effects with continuous treatments by proposing a double machine learning algorithm that uses a kernel-based doubly robust moment function, achieving asymptotic normality with nonparametric convergence rates. The method was evaluated on simulated data and applied to real-world medical data to analyze glycemic control's effect on cognitive functions.
Uncovering causal mediation effects is of significant value to practitioners seeking to isolate the direct treatment effect from the potential mediated effect. We propose a double machine learning (DML) algorithm for mediation analysis that supports continuous treatments. To estimate the target mediated response curve, our method uses a kernel-based doubly robust moment function for which we prove asymptotic Neyman orthogonality. This allows us to obtain asymptotic normality with nonparametric convergence rate while allowing for nonparametric or parametric estimation of the nuisance parameters. We then derive an optimal bandwidth strategy along with a procedure for estimating asymptotic confidence intervals. Finally, to illustrate the benefits of our method, we provide a numerical evaluation of our approach on a simulation along with an application to real-world medical data to analyze the effect of glycemic control on cognitive functions.