As allergy diagnoses among children continue to rise, a new online calculator is said to provide fast, cheap and highly accurate predictions, with potential implications for better-targeted on-pack allergen labelling.
The primary benefit of the system is to combine simplicity with a prediction accuracy of 96%, according to researchers at University College Cork (UCC), Ireland. This compares with the accuracy of current, individual tests, put at between 61% and 81%
The calculator is currently able to predict the three most prevalent types of allergy, based on cow’s milk, eggs and peanuts. It was developed by Dr Audrey DunnGalvin and Prof Jonathan Hourihane of UCC’s Paediatrics and Child Health Department.
The ‘gold standard’ and alternative to skin-prick and blood tests is the double-blind, placebo-controlled food challenge test. But these tests are usually carried out by a specialist in a hospital, with a waiting list of up to five months in Ireland. “What’s more, they’re expensive, and there’s always an element of risk involved,” said DunnGalvin.
The UCC mathematical model correlates known food challenge test results with skin-prick and blood test results, allergic reaction history, sex and age. The theory – since confirmed – was that linking more clinical information from different sources would lead to higher accuracy in the final model.
According to DunnGalvin, the occurrence of diagnosed (as opposed to self-reported) allergies among children in North America and Europe has increased from 2% or 3% only a few years ago to between 6% and 8% today.
The prediction technique holds wider implications for the food industry, for the accuracy of on-pack allergen information, and consequently for the perceived relevance of it to younger consumers in particular. “We’re beginning to establish population threshold levels,” said DunnGalvin. “If we can join up the studies we have accessed with these levels, the results could be transferred into labelling.”
In other words, warnings on a particular product could state different levels of risk depending on the consumer’s allergic reaction threshold. “As it stands, the ‘may contain…’ type of allergen labelling tends to be ignored by adolescents, for instance,” she explained.
DunnGalvin said she is already talking to industry about the patented technology, and she has encountered high levels of interest.
The research team is now working to broaden out the basis of study to include less common allergens: fish, shellfish and other nuts. “We’ve looked at UK, Canadian and Irish data, but I’d be interested to explore French data to see how cultural and genetic differences impinge on the results,” she added.
The Technology Transfer Office at UCC said it foresaw a commercial product being launched some time this year.
UCC’s research was published yesterday online in the Journal of Allergy & Clinical Immunology.