The UK Health Security Agency (UKHSA) is looking into the potential of using artificial intelligence (AI) to help detect and investigate foodborne illness outbreaks.
UKHSA experts assessed different types of AI for their ability to detect and classify text in online restaurant reviews, which could be used to identify and potentially target investigations into outbreaks.
Foodborne gastrointestinal illness is a major burden in the UK, causing millions of people to become unwell every year. However, it is estimated that most cases are not formally diagnosed, posing challenges for traditional surveillance methods.
Food Standards Agency (FSA) estimates show there are 2.4 million illnesses annually due to pathogens associated with foodborne disease. Of these, 16,400 cases receive hospital treatment and 180 result in death, costing UK society £10.4 billion.
Potential data issues
UKHSA tech experts and scientists looked at a range of large language models and rated their ability to trawl thousands of online reviews for information about symptoms which might relate to foodborne illness — such as diarrhea, vomiting, and abdominal pain — as well as different food types of food people report eating, such as meat and fish.
Scientists believe gathering information like this could become routine, providing more information on rates of illness which are not captured by current systems, as well as clues around possible sources and causes in outbreaks.
However, a study raised challenges that would need to be overcome, particularly around access to real-time data. Experts said there are limitations with use of restaurant review data that highlight the need for cautious interpretation of results.
While it is possible to use the approach to gather general information on the type of food people have eaten and which may be linked to illness, determining which specific ingredients or other linked factors is difficult. Variations in spelling and the use of slang were also identified as potential challenges, as well as people misattributing their illness to a given meal.
Support existing surveillance
Researchers applied their efforts to the Yelp Open Dataset of reviews and included information on symptoms and foods. They randomly sampled 3,000 reviews to be manually annotated using a protocol designed by UKHSA epidemiologists. This left 1,148 reviews as gastrointestinal related.
Only a fraction of all food consumed is catered food that would potentially generate a review on a website like Yelp. Restaurant review surveillance may disproportionately capture illnesses among higher-income individuals. It is unclear if income is a driver of people posting reviews online.
Incubation periods of pathogens can vary from hours to days, meaning people may misattribute their symptoms to a specific food establishment, when it is not actually the source.
“Using AI in this way could soon help us identify the likely source of more foodborne illness outbreaks, in combination with traditional epidemiological methods, to prevent more people becoming sick. Further work is needed before we adopt these methods into our routine approach to tackling foodborne illness outbreaks,” said Professor Steven Riley, chief data officer at UKHSA.
UKHSA is also expanding the pathogens that registered medical professionals and laboratories in England must notify the agency about.
The aim is to strengthen local and national surveillance and support a prompt response to outbreaks of infectious diseases.
Beginning April 6, medical professionals must notify UKHSA if they suspect a patient has Middle East respiratory syndrome (MERS), influenza of zoonotic origin, Creutzfeldt-Jakob disease and several other conditions.
Laboratories that test human samples in England will be required to report an additional 10 agents including norovirus, tick-borne encephalitis virus, toxoplasma, Trichinella, and Yersinia.
(To sign up for a free subscription to Food Safety News, click here)
