A Review of Spatial Causal Inference Methods for Environmental and Epidemiological Applications.

Area of Science:

• Environmental Science
• Epidemiology
• Spatial Statistics
• Causal Inference

Background:

• Causal inference methods are increasingly vital across disciplines.
• Spatial causal inference presents unique challenges due to complex spatial correlations and interference.
• Recent advancements are beginning to address these spatial complexities.

Purpose of the Study:

• To review the existing literature on spatial causal inference.
• To identify key challenges and areas for future research in spatial causal inference.
• To provide practical guidance and code for applying these methods.

Main Methods:

• Exploiting spatial structure to control for unmeasured confounding variables.
• Analyzing spatial interference using simplifying assumptions.
• Extending methods to spatiotemporal data, comparing potential outcomes and Granger causality.
• Applying geostatistical analyses for spatial random fields.

Main Results:

• Demonstrated methods for accounting for spatial confounding and interference.
• Provided a framework for spatiotemporal causal analysis.
• Illustrated applications in environmental and epidemiological studies, including air pollution and COVID-19 mortality.
• Offered open-source code for implementing discussed methods.

Conclusions:

• Spatial causal inference is a rapidly developing field with significant potential.
• Addressing spatial confounding and interference is crucial for robust causal conclusions.
• The reviewed methods offer valuable tools for environmental and health research.
• Further methodological development is needed for complex spatial systems.