The use of modelling and probabilistic methods in cumulative risk assessment

Abstract

This thesis was realized as part of the EU integrated project SAFE FOODS, the overall objective of which was to change the scope of decision-making on food safety from single risks to considering foods as sources of risks, benefits and costs associated with their production and consumption, and taking into account the social context in which decisions are made. One of the main goals within this project was the development and application of new quantitative risk assessment methods that better meet these demands of decision-makers. These new methods should account for variability in consumer populations and uncertainties in the assessment, and they should be fit to address the exposure to combinations of various chemicals present in the diet. This thesis combines several contributions to the development of this new approach. Two different computational tools are used for this purpose. Biologically based mathematical models (or mechanistic models) are models that quantitatively describe the behavior of chemicals in a biological system by use of differential equations. In this thesis, these models have been used to study the combined effects of chemicals, and how these could be predicted from information of the chemicals individually. Probabilistic methods are computational methods that combine statistical distributions of variables instead of point estimates. These methods are incorporated in the risk assessment framework to account for the variability in a population and uncertainty in the assessment. The final result is an estimate of potential health risks that may occur in the population, and a statement on how precise that estimation is.