While the initial trials are being undertaken by NSF with one of its leading foodservice clients, which operates more than 1,500 sites nationwide, the new approach could also find use in food and drink manufacturing plants, said NSF.
The new approach to auditing would ensure hygiene compliance by scheduling audits to maximise risk reduction, rather than relying on the standard model of interval-based audits followed by remedial intervention, according to the company.
Although initially targeted at large, multi-site foodservice and retail outlets, it could also have applications for food manufacturing, where enough production lines were at a site so that a sufficient quantity of data could be generated to predict when particular lines were about to go out of control, said NSF.
Go out of control
The predictive scheduling approach to food safety auditing was described by Chris Pratsis, commercial director of NSF at a food safety conference organised by the company in London last month.
It uses statistical analysis and mathematical algorithms to predict when risk conditions will be at their highest and identify why.
NSF reported that it expected to start rolling out commercial versions of the system over the next six months.
“There seems to be an inescapable benefit to the value of not only being able to target the highest-risk businesses but to be able to predict where the business is most likely to be at risk of a significant food safety breakdown,” said Pratsis.
‘Food safety intervention scheduling tool’
“Over the past two years, we at NSF have been working to develop our own algorithm-based approach using correlated risk-determining factors and historical data and current data feeds to create a predictive food safety intervention scheduling tool – Smart predictive scheduling.
“This we see very much as being the next generation approach of the traditional model for interval-based food safety and health and safety risk compliance auditing.”
It focuses on where and when there might be a breakdown in the perceived level of control, creating the opportunity to intervene as necessary. Because of the large amounts of relevant data required, it was better suited to operators of large numbers of sites and more mature operators, he added.