A novel method for controlling unobserved confounding using double confounders.

Area of Science:

• Epidemiology
• Biostatistics
• Causal Inference

Background:

• Unobserved confounding poses a significant challenge in observational studies, often violating assumptions of existing methods.
• Developing novel approaches to control for unobserved confounders is crucial for reliable causal effect estimation.

Purpose of the Study:

• To propose a new method for estimating causal effects in the presence of unobserved confounding.
• To identify conditions under which causal effects can be nonparametrically identified using accessible confounders.

Main Methods:

• The study proposes a novel method utilizing two independent or correlated confounders that satisfy a non-linear condition on the exposure.
• Asymptotic theory and variance estimators are developed for continuous and categorical outcomes.
• An extension for more than two binary confounders is also discussed.

Main Results:

• Simulations demonstrated superior performance compared to traditional regression methods.
• A real-world application assessed the effects of Body Mass Index (BMI) on various health indicators in a Chinese population.
• Increased BMI was causally associated with higher Systolic Blood Pressure, Diastolic Blood Pressure, Total Cholesterol, Triglycerides, and Low Density Lipoprotein.

Conclusions:

• A novel method is presented to control for unobserved confounding using readily accessible double binary confounders.
• The method relies on a non-linear condition related to the exposure, offering a practical approach to causal inference.