Causal Estimation of Exposure Shifts with Neural Networks: Evaluating the Health Benefits of Stricter Air Quality Standards in the US
Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Keywords: causal inference; neural networks; air pollution; stochastic interventions; doubly robust inference
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
TL;DR: We develop a neural network method for doubly-robust estimation of shift response functions motivated by a recent proposal from the US EPA which impels us to measure the effect of changing the air quality regulations.
Abstract: In policy research, one of the most critical analytic tasks is to estimate the causal effect of a policy-relevant shift to the distribution of a continuous exposure/treatment on an outcome of interest. We call this problem *shift-response function* (SRF) estimation. Existing neural network methods involving robust causal-effect estimators lack theoretical guarantees and practical implementations for SRF estimation. Motivated by a key policy-relevant question in public health, we develop a neural network method and its theoretical underpinnings to estimate SRFs with robustness and efficiency guarantees. We then apply our method to data consisting of 68 million individuals and 27 million deaths across the U.S. to estimate the causal effect from revising the US National Ambient Air Quality Standards (NAAQS) for $\text{PM}_{2.5}$ from 12 to 9 $\mu g/m^3$ . This change has been recently proposed by the US Environmental Protection Agency (EPA). Our goal is to estimate, for the first time, the reduction in deaths that would result from this anticipated revision using causal methods for SRFs. Our proposed method, called Targeted Regularization for Exposure Shifts with Neural Networks (TRESNET), contributes to the neural network literature for causal inference in two ways: first, it proposes a targeted regularization loss with theoretical properties that ensure double robustness and achieves asymptotic efficiency specific for SRF estimation; second, it enables loss functions from the exponential family of distributions to accommodate non-continuous outcome distributions (such as hospitalization or mortality counts). We complement our application with benchmark experiments that demonstrate TRESNET's broad applicability and competitiveness.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 8476
Loading