Estimation of Heterogeneous Causal Mediation Effects in a Hypertension Treatment Trial
Yi Zhao, Chengyun Li, Wanzhu Tu

TL;DR
This paper introduces a new framework for estimating heterogeneous causal mediation effects in clinical trials, addressing inconsistencies in standard methods by modeling individual differences and applying advanced regularization techniques.
Contribution
A novel linear structural equation modeling approach incorporating covariate-treatment interactions and a modified covariate method with generalized lasso for high-dimensional data.
Findings
Substantial heterogeneity in mediation effects identified in SPRINT data
A subset of patients benefits from therapies targeting albuminuria
Simulation studies show good estimation and inference performance
Abstract
Hypertension is a highly prevalent condition and a major risk factor for cardiovascular disease. The landmark Systolic Blood Pressure Intervention Trial (SPRINT) showed that lowering systolic blood pressure (BP) goals from 140 mmHg to 120 mmHg leads to significantly reduced BP, cardiovascular mortality, and morbidity. However, the underlying mechanisms are not yet fully elucidated. In patients with impaired renal function, early reduction of albuminuria has been proposed as a potential mediation pathway. Evidence from the standard causal mediation analysis (CMA), however, yields inconsistent results, possibly due to heterogeneous mediation effects across individuals. To disseminate the heterogeneity, a new framework that incorporates covariate-treatment and mediator-treatment interactions within a linear structural equation modeling system is introduced. Causal assumptions are discussed…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Causal Inference Techniques · Statistical Methods and Bayesian Inference · Statistical Methods and Inference
