You are here:

Food & feed classification for tracing purposes (FoodClass)

06/2021-06/2023

Funding programme / funding institution: Europäische Behörde für Lebensmittelsicherheit (EFSA)

Grant number: GP/EFSA/AMU/2020/02

Project homepage: https://foodrisklabs.bfr.bund.de/

Project description:

In case of foodborne disease outbreaks, rapid identification of the causative food product is essential, since the medical and economic damages grow with the duration of the outbreak. Recent foodborne disease outbreaks in Europe illustrated that there is a need for a software system capable of supporting investigations on supply chains as well as exposure assessments in crisis situations. To address these needs, the FoodChain-Lab tracing software has been developed. With the growing number of applications of FoodChain-Lab in outbreak investigations and the availability of additional tracing tools for various purposes, further standardization of the analyzed information is needed to enable comparisons, and statistics on several investigations. Results could be used to identify food and feed items (products), contaminations (e.g. pathogens), and actors in the food chain with higher risks of incidents.

Precondition is a sufficient and user-friendly classification system for the main data elements for tracing. Tracing activities follow a food/feed item through the whole food chain from primary production via processing to the final consumers. Food and Feed classification systems (integrated in FOODEX2) are made for specific steps of the supply chain (e.g. for raw material, consumed products or processed food) and are therefore not always directly applicable. With the help of facets, e.g. for processing and packaging, some connections between the steps of the supply chain can be made, but for complex food items with many ingredients the system is limited. A fit-for-purpose classification system to be used for tracing activities should be simple and easy to apply, the system might be hierarchical with different levels of details. FOODEX2 is a hierarchical standardized food classification system developed by EFSA which incorporates several domain-specific hierarchies and 28 facets, but an evaluation on how to use it for tracing activities/investigations is missing. Furthermore, the system should be integrated into the data collection workflow developed in P-AMU-28. Other data elements of interest are package and transportation types, food and feed sectors, and actors in the food and feed chain.

Tasks:

The main purpose of the project is the development of a food and feed classification system for tracing purposes within the FOODEX2 system, a corresponding coding manual, and the test of automatic data extraction from the RASFF notifications by AI. The development of the classification system is divided into three steps. In the first work package (WP 1) a theoretical analysis will identify needs and define criteria on classification systems of food and feed for tracing. Possible hierarchies and facets for tracing purposes, esp. to describe complex products, will be identified and discussed. The second work package (WP 2) will define and execute a protocol for searching of existing classification systems used for tracing purposes by competent authorities, international organizations, and companies (e.g. retailer). A comprehensive literature/web search or survey on classification systems will be performed. The retrieved systems will be evaluated by the identified quality criteria. In the third work package (WP 3) the integration into the FOODEX2 system will be proposed. A comparison among the status-quo of FOODEX2, the existing systems outside EFSA and the theoretical needs and quality criteria should identify catalogues and facets of FOODEX2 for tracing purposes (preferred option), gaps which could be filled by external standards (second option), and gaps to be filled by additional catalogues/facets to be developed. This results in a coding manual for tracing data. In parallel, a feasibility study (WP 4) should test, if AI solutions (RASNEX tool) can be used to extract information from the RASFF notifications, follow-ups (pdf-files) and attachments in different file formats. Focus will be the extraction of address information in the right data format and classification, but could be extended to product information and similar. All project results and changes in the tracing workflow (P-AMU-28) will be continuously implemented into the FoodChain-Lab software in work package 5. At the end of this project the coding manual and automatic extraction tool will be presented, discussed and reviewed (WP 6) during a workshop with possible users from the member states. An early involvement of the RASFF network, including dissemination of the coding manual is intended in collaboration with the SCER Unit. The coding manual should be accessible to food business operators, competent authorities, and investigators during outbreak investigations.

Project partners

  • Europäische Behörde für Lebensmittelsicherheit (EFSA) - Italien

Up

Keywords

 (1)

nach oben

Cookie Notice

This site only uses cookies to offer you a better browsing experience. Find out more on how we use cookies in our Data Protection Declaration.