Moving from two- to multi-way interactions among binary risk factors on the additive scale

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Existing research often examines interactions between two risk factors.
• There's a literature gap in analyzing additive interactions involving more than two risk factors.

Purpose of the Study:

• To present a novel approach for assessing deviations from additive interaction with three or more binary exposures.
• To extend the analysis of relative excess risk due to interaction (RERI) for multiple factors.

Main Methods:

• Proposes decomposing the total RERI for three risk factors into contributions from all three and pairwise interactions.
• Extends this decomposition method to analyze interactions among more than three binary risk factors.
• Applies the method to data from the Greek EPIC cohort, examining mortality risks associated with Mediterranean diet, BMI, and smoking.

Main Results:

• The developed formulae enhance the interpretability of additive interaction deviations involving multiple risk factors.
• Provides a straightforward way to communicate public health implications by linking excess relative risk to specific factor combinations.

Conclusions:

• The new approach offers improved interpretability for additive interaction analysis with multiple risk factors.
• Facilitates clearer communication of complex interaction effects for public health decision-making.