Disease Attribution to Multiple Exposures Using Aggregate Data

Abstract

Background: Identifying which exposures cause disease and quantifying their impacts is essential in promoting and monitoring public health. When multiple exposures are involved, measuring individual contributions becomes challenging. Methods: The authors propose a disease attribution method based on aggregate data or summary statistics of individual-level data, possibly from multiple data sources. Results: Using the proposed method, the burden of disease is apportioned to the independent and interaction effects of each of its major risk factors and all the other factors as a whole. This scheme guarantees that 100% is the total share of the burden. Conclusion: The calculation is simple and straightforward; therefore, it is recommended for use in studies on disease burden.

Keywords

• disease attribution
• attributable fraction
• burden of disease
• interaction
• causal pie model