New methods for optimizing monitoring of aflatoxins and dioxins in the food supply chain

Dioxins (including DL-PCBs) and aflatoxins, especially aflatoxin B1 (AFB1) and aflatoxin M1 (AFM1) are two of the most relevant chemical hazard groups in the dairy (and other) supply chains, due to their occurrence in dairy products and their negative impacts on human health. Food safety monitoring is essential for hazard identification in food chain to guarantee public health, but its application may be limited due to costly analytical methods and inefficient sampling procedures. Recent studies on optimal monitoring schemes for dioxins and AFB1/AFM1 along food supply chains were based on cost-effectiveness analyses. However, several aspects which could affect the cost-effectiveness of monitoring chemical hazards have not yet been explored. For instance, specific sampling strategies, such as random or systematic sampling, at each control point of the food supply chain, adapting the number of samples per product type or per period of the year, given a probability of contamination, could further improve the effectiveness of the monitoring of dioxins and AFB1/AFM1 along the chain. Furthermore, next to the monitoring effectiveness and economic criteria, other criteria like public health, customer trust, and/or the practicality of implementation, could be considered in designing optimal monitoring schemes. To fill the knowledge gaps detailed above, this thesis aimed to develop new methods for optimizing monitoring programs for dioxins and AFB1/AFM1 in food supply chains by taking into account criteria of monitoring costs, production losses, human health, complexity of implementation and customer trust according to stakeholder preferences. Firstly, we compared spatial sampling strategies used in monitoring dioxins along the dairy supply chain to study the cost-effectiveness of monitoring programs. Then a risk based approach was developed to allocate resources for monitoring aflatoxins and dioxins at different control points of dairy supply chain in order to reduce potential impacts on public health. Thirdly, we used Bayesian Network to predict probability of different agricultural products contaminated by dioxins given different conditions, and based on predicted probability, we optimized dioxin monitoring scheme in agricultural products. Lastly, we investigated the optimal chemical monitoring schemes in the dairy chain, evaluated from industry perspective (feed industry and dairy industry as a whole), with considering effectiveness, acceptability and implementation aspects of monitoring scheme. The multicriteria decision making methodology was used to rank alternative strategies of monitoring aflatoxins and dioxins contamination according to the preferences of industries’ experts.