Authors: Ying Zhou, Eric Tchetgen Tchetgen

Published on: May 15, 2024

Impact Score: 7.2

Arxiv code: Arxiv:2405.09080

Summary

• What is new: Introduces a novel semiparametric EM algorithm for improved causal inference when dealing with unobservable treatment variables using surrogate measurements.
• Why this is important: Difficulty in causal inference when direct observation of a treatment variable is impossible, relying only on an imperfect surrogate.
• What the research proposes: A new method that incorporates both the surrogate measurement and a proxy of the hidden treatment to effectively identify the causal effects.
• Results: Demonstrated successful identification and estimation of causal effects in simulations and an Alzheimer’s study with the proposed method.

Technical Details

Technological frameworks used: Semiparametric EM algorithm, multiple robustness property semiparametric estimators

Models used: Hidden treatment causal models

Data used: Alzheimer’s Disease Neuroimaging Initiative dataset

Potential Impact

Healthcare analytics, particularly sectors involved in neuroimaging and chronic disease management.

Want to implement this idea in a business?

We have generated a startup concept here: Causalytics.