[Causal graph model and its application in nutritional epidemiologic research].

Area of Science:

• Nutritional Epidemiology
• Causal Inference
• Public Health

Background:

• Suboptimal diet is a key controllable risk factor for non-communicable diseases.
• Randomized controlled trials face challenges in quantifying specific dietary factor-health outcome causal links.
• Causal inference methods offer robust tools for analyzing observational data in nutrition research.

Purpose of the Study:

• Introduce causal graph models for visualizing complex dietary relationships.
• Explain various causal analysis strategies within the causal graph framework.
• Promote the application of causal graph models in nutritional epidemiology.

Main Methods:

• Utilizing causal graph models to integrate prior knowledge and identify confounding factors.
• Applying different causal inference strategies (confounder adjustment, instrumental variables, mediation analysis) based on causal graphs.
• Reviewing the application of these methods in nutritional epidemiology.

Main Results:

• Causal graph models provide a framework for identifying confounding and determining causal effect estimation strategies.
• Various analysis strategies can be derived from causal graphs for robust evidence generation.
• The approach facilitates high-quality nutritional epidemiologic research.

Conclusions:

• Causal graph models are valuable tools for advancing nutritional epidemiology.
• These models enhance the ability to study diet-health relationships using observational data.
• The paper provides a foundation for future research and application in nutrition science.