On Heterogeneous Treatment Effects in Heterogeneous Causal Graphs

Richard A Watson, Hengrui Cai, Xinming An, Samuel Mclean, Rui Song
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:36714-36747, 2023.

Abstract

Heterogeneity and comorbidity are two interwoven challenges associated with various healthcare problems that greatly hampered research on developing effective treatment and understanding of the underlying neurobiological mechanism. Very few studies have been conducted to investigate heterogeneous causal effects (HCEs) in graphical contexts due to the lack of statistical methods. To characterize this heterogeneity, we first conceptualize heterogeneous causal graphs (HCGs) by generalizing the causal graphical model with confounder-based interactions and multiple mediators. Such confounders with an interaction with the treatment are known as moderators. This allows us to flexibly produce HCGs given different moderators and explicitly characterize HCEs from the treatment or potential mediators on the outcome. We establish the theoretical forms of HCEs and derive their properties at the individual level in both linear and nonlinear models. An interactive structural learning is developed to estimate the complex HCGs and HCEs with confidence intervals provided. Our method is empirically justified by extensive simulations and its practical usefulness is illustrated by exploring causality among psychiatric disorders for trauma survivors. Code implementing the proposed algorithm is open-source and publicly available at: https://github.com/richard-watson/ISL.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-watson23a, title = {On Heterogeneous Treatment Effects in Heterogeneous Causal Graphs}, author = {Watson, Richard A and Cai, Hengrui and An, Xinming and Mclean, Samuel and Song, Rui}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {36714--36747}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/watson23a/watson23a.pdf}, url = {https://proceedings.mlr.press/v202/watson23a.html}, abstract = {Heterogeneity and comorbidity are two interwoven challenges associated with various healthcare problems that greatly hampered research on developing effective treatment and understanding of the underlying neurobiological mechanism. Very few studies have been conducted to investigate heterogeneous causal effects (HCEs) in graphical contexts due to the lack of statistical methods. To characterize this heterogeneity, we first conceptualize heterogeneous causal graphs (HCGs) by generalizing the causal graphical model with confounder-based interactions and multiple mediators. Such confounders with an interaction with the treatment are known as moderators. This allows us to flexibly produce HCGs given different moderators and explicitly characterize HCEs from the treatment or potential mediators on the outcome. We establish the theoretical forms of HCEs and derive their properties at the individual level in both linear and nonlinear models. An interactive structural learning is developed to estimate the complex HCGs and HCEs with confidence intervals provided. Our method is empirically justified by extensive simulations and its practical usefulness is illustrated by exploring causality among psychiatric disorders for trauma survivors. Code implementing the proposed algorithm is open-source and publicly available at: https://github.com/richard-watson/ISL.} }
Endnote
%0 Conference Paper %T On Heterogeneous Treatment Effects in Heterogeneous Causal Graphs %A Richard A Watson %A Hengrui Cai %A Xinming An %A Samuel Mclean %A Rui Song %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-watson23a %I PMLR %P 36714--36747 %U https://proceedings.mlr.press/v202/watson23a.html %V 202 %X Heterogeneity and comorbidity are two interwoven challenges associated with various healthcare problems that greatly hampered research on developing effective treatment and understanding of the underlying neurobiological mechanism. Very few studies have been conducted to investigate heterogeneous causal effects (HCEs) in graphical contexts due to the lack of statistical methods. To characterize this heterogeneity, we first conceptualize heterogeneous causal graphs (HCGs) by generalizing the causal graphical model with confounder-based interactions and multiple mediators. Such confounders with an interaction with the treatment are known as moderators. This allows us to flexibly produce HCGs given different moderators and explicitly characterize HCEs from the treatment or potential mediators on the outcome. We establish the theoretical forms of HCEs and derive their properties at the individual level in both linear and nonlinear models. An interactive structural learning is developed to estimate the complex HCGs and HCEs with confidence intervals provided. Our method is empirically justified by extensive simulations and its practical usefulness is illustrated by exploring causality among psychiatric disorders for trauma survivors. Code implementing the proposed algorithm is open-source and publicly available at: https://github.com/richard-watson/ISL.
APA
Watson, R.A., Cai, H., An, X., Mclean, S. & Song, R.. (2023). On Heterogeneous Treatment Effects in Heterogeneous Causal Graphs. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:36714-36747 Available from https://proceedings.mlr.press/v202/watson23a.html.

Related Material