Risk models for managing allergens in food

Allergens are an unusual food hazard in that they are harmless to more than 95% of consumers, but it can be possible for allergens to find their way into food products unintentionally, especially when shared facilities and processing lines are used.

Accidental reactions

Working with patient organisations, incidents of reactions from Spain, France, Poland and Germany were analysed, which showed pre-packed food was responsible for reactions in adults and children. The majority of reactions took place in the home for children, but only around half for adults.

Consumer attitudes to precautionary allergen labelling (PAL)

PAL is used by the food industry to tell consumers when food products may contain unintended allergens that could cause them to have a reaction. Working with different stakeholders, iFAAM identified that:

  • the lack of agreed reference doses means PAL is not applied in a consistent manner and does not reflect the presence of unintended allergens;
  • phrases such as ‘This product is not suitable for’ and ‘May contain’ were the most acceptable to food-allergic consumers;
  • trust in PAL has been lost, reducing the ability of consumers with food allergies to make informed choices, reducing quality of life and increasing risk-taking by consumers who often ignore PAL;
  • trust in PAL could be improved if it was used together with an indication that a risk assessment process had been used.

Food recalls

Analysis of recalls in Europe, North America, Hong Kong, Australia and New Zealand showed the vast majority of the recalls were as a result of labelling errors, with up to 17% being caused by the unintended presence of an allergen.

Number of recalls per known allergen

  • Milk: 258
  • Sulphur dioxide and sulphites: 182
  • Cereals containing gluten: 102
  • Soybeans: 60
  • Peanuts: 57
  • Eggs: 55
  • Nuts: 45
  • Celery: 23
  • Mustard: 13
  • Molluscs: 7.6
  • Fish: 6
  • Sesame seeds: 5.4

Methods have been developed to assess the risk that such unintended allergens pose of causing an allergic reaction. iFAAM has contributed the development of these models by:

  • enriching a pre-existing database of oral food challenge data on different foods, with data from the EuroPrevall and iFAAM studies;
  • a common set of food consumption data (consumption in a meal) was developed using data from Denmark, the Netherlands and France, and grouped to allow their use in allergen risk assessment;
  • the data on threshold doses and consumption were then used to develop the iFAAM Tier 1 and Tier 2 risk assessment approaches, together with an allergen tracking tool and a risk mitigation matrix. To find out more about the tools developed in iFAAM


Further reading