The trade network in the dairy industry and its implication for the spread of contamination.

In case of an outbreak of a foodborne disease, administrative decisions in the context of crisis management are only efficient if they follow standard practices and are specifically adapted to the outbreak situation in a timely manner. These goals are hard to achieve. The complexity of national and global trade structures obscures a clear view of trade flows and, consequently, it is often impossible to unravel complex trade links quickly. Furthermore, increasing public concerns about possible health hazards caused by global trade put additional pressure on decision makers. The aim of this paper was to unveil the specific trade structures of the German milk supply chain, to highlight how these structures could affect the spatial spread of a hypothetical contaminant, and to quantify the risk of the contaminant reaching the consumer. To achieve this goal, the vertical and horizontal trade links between milk producers, dairies, and consumers were taken into account. The horizontal flow of milk between dairies (inter-dairy trade), which is intended to compensate a temporary over- or undersupply of milk, is of special importance in this respect. We hypothesized that the extent of inter-dairy trade would significantly influence the spatial spread of contaminated milk and the contamination risk. This hypothesis was tested using a computer simulation model that predicts the hypothetical spread of a contaminant via trade of milk. The model parameters were estimated using trade data collected in 2004 and 2010. The results of our study indicate that inter-dairy trade significantly influenced the contamination risk. Compared with a scenario with no inter-dairy trade, the risk that contaminated milk will reach the consumer was up to 4 times higher, even with moderate inter-dairy trade. The contamination risk depended on the extent of inter-dairy trade in a nonlinear way and reached its maximum asymptotically when inter-dairy trade increased. The contamination risk exhibited considerable spatial variation, which could be utilized to implement more accurate food control interventions in times of crisis caused by a foodborne disease.