A new computer-assisted method developed jointly by the German Federal Institute for Risk Assessment (BfR), the IBM Almaden Research Centre and the Johns Hopkins Bloomberg School of Public Health analyses sales data to help identify contaminated foods involved in disease outbreaks.
The probability-based method takes data on the types and quantities of foods sold in particular locations and compares the distribution patterns of specific foods with the distribution pattern of cases of illness linked to an outbreak. This approach can quickly narrow down the possible culprits to a small, localised group of products and speed up the identification of the contaminated food.
The method works on the assumption that there will usually be a close geographic link between sales locations and sales volumes of contaminated food and the occurrence of cases of illness. According to the BfR it is especially suitable for investigating outbreaks attributable to a single food produced by a single manufacturer.
The method has been successfully tested in Germany using real sales data combined with fictitious computer-generated outbreak scenarios. The next stage in the research will be the development of the method to identify the source of an outbreak when several contaminated foods are involved, such as in the case of contaminated common ingredient.
A detailed description of the method is published in the journal PLOS Computational Biology and can be found here.