Towards individual-based approaches for risk assessment in food safety

Foodborne diseases are caused by pathogenic microorganisms and chemical noxes present in the foods and drinks we ingest. The World Health Organization estimated that foodborne diseases were responsible for 600 million cases of illness and 420,000 deaths worldwide in 2010. With the intention of reducing the risk and ultimately the burden of disease in the population, food safety policies, legislation and interventions are designed based on the available scientific evidence. The identification and prioritization of the most effective food safety interventions is supported by evidence on the public health impact of the different foodborne diseases, on the sources and routes of transmission of the disease, on the of options for intervention in the food chain. Currently, this evidence is generated and presented through a range of data-driven tools.

Consumer exposure to foodborne hazards depends on a variety of factors, several of which are dependent on consumer practices, such as food consumption, storing and preparation choices, whereas other factors are outside of their control, such as microbial and chemical contamination. Analyses of differences between consumer profiles and their impact on exposure assessments may contribute to the development of more targeted risk assessments for different population groups. Microbial risk assessments have traditionally used exposure assessment approaches that disregard individual-level behaviour and other characteristics. The overall aims of the PhD thesis were to develop individual-based dietary exposure assessment in chemical and microbial risk assessment, and burden of disease studies of chemicals, and to make the approaches available for risk assessors, risk managers, and the scientific community.

This thesis consists of two main studies, one assessing the risk of a foodborne pathogen (Listeria monocytogenes) through consumption of ready to eat foods (RTE) in France, and another one estimating the burden of disease of dietary exposure to four chemical contaminants in the same population. Both studies focused on the development of individual-based exposure assessment using harmonized national data, so that it is possible to reproduce the models in another country. In manuscript I (“Quantitative microbiological risk assessment based on individual consumer data: A model for listeriosis from RTE foods in France (Part I)”) the risk of listeriosis was assessed based on the consumer´s food choices and the food storage on RTE foods. The model estimated 393 annual cases of listeriosis, from which 66% were linked with the storage and consumption of smoked fish and pâte. We concluded that the majority of the cases were attributed to specific food storage behaviour regarding the storage before consumption and the temperature of storage. This study demonstrated that considering individual-based data in QMRA opens the way for the establishment of risk-based measures with a greater resolution specific for distinct individuals within the population.

In manuscript II (“Quantitative microbiological risk assessment based on individual consumer data: a comparison with traditional QMRA approaches (Part II)”), we tested the effect of assumptions of manuscript I and measured the effect of data treatments on food storage and food consumption applied, using the same individual-based model developed for manuscript I. Out of the data treatments tested, we concluded that storage time has the largest effect on the number of cases, linked with the specific assumption that these individuals do not consider “use-by” date or purchasing time as reference.

In manuscript III (“Harmonized approach to estimate the burden of disease of dietary exposure to four chemical contaminants - a French study”), we estimated the burden of disease of dietary exposure to four chemical contaminants to the French population. The models were based on previously developed burden of disease models, modified to ensure harmonization according to European food categorization standards, and to be able to use data from harmonized total diet studies. Results showed that dietary exposure to lead and inorganic arsenic caused the highest burden, linked with intellectual disability caused by early childhood exposure, and lung, skin and bladder cancer, respectively.

This thesis fostered the use of individual based approach across food safety risk assessments, when data are available and representative of the population. The models and data used in this thesis were released in Food Safety Knowledge Exchange format with the intention of sharing findable, interoperable, available, and reusable knowledge. Further research should extend the generation of food safety risk models in FSKX format.