The UK Health Security Agency (UKHSA) investigates the potential of using artificial intelligence (AI) to help detect and investigate the outbreaks of diseases carried out by food.
UKHSA experts rated different types of AI on their ability to detect and classify text in online restaurant reviews, which can be used to identify and possibly target research into outbreaks.
Food -transmitted gastrointestinal disease is a major burden in the UK, making millions of people unwell every year. It is estimated that most cases are not formally diagnosed, which sets challenges for traditional security methods.
Estimates of the Food Standards Agency (FSA) show that 2.4 million diseases are associated with illnesses of diseases through food due to pathogens. Of these, 16,400 cases get hospital treatment and 180 result in death, which costs British society £ 10.4 billion.
Potential data issues
UKHSA technical experts and scientists looked at a series of large language models and rated their ability to train thousands of online reviews for information about symptoms that may relate to food -handed diseases – such as diarrhea, vomiting and abdominal pain – as well as different food people of food, such as meat and fish.
Scientists believe that collecting such information can become routine and give more information about the rates of diseases that are not recorded by current systems, as well as instructions on possible sources and causes during outbreaks.
How but one Study increased challenges That should be overcome, especially around access to real -time data. Experts said there are restrictions in the use of data from restaurant rating that emphasize the need for cautious interpretation of the results.
Although it is possible to use the approach to collect general information about the type of food that people have eaten and that can be linked to illness, it is difficult to determine which specific ingredients or other linked factors. Variations in spelling and the use of jargon were also identified as potential challenges, as well as people who misunderstand their illness to a certain meal.
Support for existing supervision
Researchers have applied their efforts to the Yelp Open data set of assessments and contain information about symptoms and foods. They randomly tasted 3,000 reviews to be manually annotated using a protocol designed by Ukhsa -epidemiologists. As a result, 1,148 reviews such as gastrointestinal.
Only a fraction of all food consumed is provided food that might generate a review on a website like Yelp. Restaurant Review Surveillance can disproportionately record diseases in people with a higher income. It is unclear whether income is a motorcycle of people who post reviews online.
Incubation periods of pathogens can vary from hours to days, which means that people can incorrectly inflicting their symptoms to a specific food company if it is not actually the source.
“The use of AI in this way can soon help us to identify the likely source of more outbreaks of diseases transferred by food, in combination with traditional epidemiological methods, to prevent more people from getting sick. Further work is needed before we use these methods in our routine approach to tackle food -based diseases,” Professor said Professor said Professor.
UKHSA also expands the pathogens who have to inform the registered medical professionals and laboratories in England.
The aim is to strengthen local and national surveillance and to support a quick response to outbreaks of infectious diseases.
From 6 April, medical professionals must inform UKHSA if they suspect that a patient has the Middle East Respiratory Syndrome (MERS), Influenza of Zoönotic origin, the disease of Creutzfeldt-Jakob and various other disorders.
Laboratories that test human samples in England must report another 10 agents, including norovirus, by drawing encephalitis virus, toxoplasm, trichinella and yersinia.
(To register for a free subscription to Food Safety News, Click here))