Network analysis reveals food safety ‘transgressors’

Researchers at Kingston University in London have developed an analytical technique which reveals emerging trends, such as particular food categories and countries, associated with high levels of food safety incidents.

The system works by interrogating international food safety alert databases, such as the EU’s Rapid Alert System for Food and Feed (RASFF), using network analysis techniques, said Declan Naughton, professor of biomolecular science in the Department of Life Sciences at Kingston University.

The research team is currently the only one applying network analysis to food safety databases, claimed Naughton.

By applying network analysis to databases, which contain huge amounts of food safety information, it is possible to apply various filters to the information. This quickly identifies incident hot spots, whether they are types of food or countries that business involved in the global food chain and regulators should be aware of.

Speaking at a one-day conference on the future of the food industry organised by food safety specialist NSF International at the offices of legal firm DWF last week (February 12), Naughton gave a demonstration of how quick and easy the system was to use.

‘Net transgressors’

Network analysis helps to differentiate “net transgressors”, those countries responsible for high numbers of incidents, from those which are picking them up, so-called “net detectors” or “notifiers”, he said.

As well as RASFF, other countries such as US, Canada, and Israel operate similar rapid alert databases, said Naughton. Ideally, these should be interrogated by the system.

“In many cases we are not doing a lot with [these databases],” said Naughton, who added that the longer term aim was to develop a global system that operated with real-time feeds from databases such as RASSF to help developing nations, in particular, improve their food safety. “Using network analysis we can rapidly get a picture of what is happening.

“Very quickly we are able to ‘bottom out’ what is happening in the database … it’s a five second job.”

Border rejections

The prototype system developed by Naughton and his co-researchers allows a multitude of filters to be applied to the databases. These might be for incidents involving metal contamination, micro-organisms and other bacteria, mycotoxins – even border rejections – which can have the biggest impact on databases. He is currently working with the UK’s Food Standards Agency in developing the system.

The conference also heard from a range of other speakers, including NSF’s Serban Teodoresco, David Edwards and Carole Payne, DWF’s Dominic Watkins and John Barnes, head of the FSA’s Local Delivery Division.

Much of the day was devoted to the fall-out from the 2013 horsemeat contamination scandal and speakers described how the food industry and regulators were responding to it in an attempt to avert future food fraud incidents.