Using empirical likelihood to combine data: application to food risk assessment

Area of Science:

• Environmental Health
• Food Safety
• Statistical Modeling

Background:

• Accurate food safety risk assessment requires integrating diverse contamination and consumption data.
• Existing methods struggle with large datasets and complex risk calculations.
• Quantifying the probability of exceeding safe contaminant doses is crucial for risk management.

Purpose of the Study:

• To develop an integrated methodology for food safety risk assessment.
• To create a computable risk index combining multiple data sources.
• To provide risk managers with a robust measure of contaminant exposure risk.

Main Methods:

• Introduced an empirical likelihood methodology for data integration.
• Defined a risk index as the probability of exposure exceeding a safe dose.
• Employed linearization techniques and approximated U-statistics for computational feasibility.
• Developed an alternative Euclidean likelihood program.

Main Results:

• The empirical likelihood approach provides a tractable method for risk index computation.
• Asymptotic confidence intervals were derived for the risk index.
• The methodology was validated using simulated data.
• Applied to assess methylmercury risk in seafood.

Conclusions:

• The proposed empirical likelihood method effectively combines food safety data for risk assessment.
• The developed risk index offers a valuable tool for risk managers.
• The methodology is applicable to real-world scenarios like methylmercury in seafood.