En savoir plus

Rodney Feliciano, doctorant dans le cadre du projet européen ITN PROTECT, a présenté ses travaux dans le cadre de la conférence internationale virtuelle d'IAFP le 28 avril 2021.

Modélisation probabiliste de l'exposition à E. coli dans le lait cru

Rodney Feliciano, Secalim, INRAE / ONIRIS, Nantes, France
 Géraldine Boué¹, Fahad Mohssin², Muhammad Mustafa², Jeanne-Marie Membré¹

¹Secalim, INRAE, Oniris, Nantes, France

²AlSafi Danone, Al-Kharj, Saudi Arabia

Introduction: Probabilistic modelling tools are increasingly being developed to take into account the different sources of variability and uncertainty along the food safety continuum in exposure or risk assessment models. These models are also popular to reflect real-life data and generate what-if scenarios to inform decision makers.
Purpose: Probabilistic modelling tools were utilized in developing a coliform bacteria exposure assessment model in raw milk in Kingdom of Saudi Arabia (KSA). This country was taken as proxy of what will happen in Europe in the near future due to climate change.
Methods: The initial coliform concentration in raw milk was derived from industrial dairy farm data (around 1695 data) while microbial growth was determined across various scenarios of time and temperature storage, using existing databases. The exposure to coliform was interpreted considering KSA and EU standards. The probabilistic model, with uncertainty and variability separated, was implemented in R using the mc2d while to fit the data, the nls function and the packages nlstools and fitdistrplus were used.
Results: The level of exposure to coliform concentration was compared to KSA and EU standards as function of various time and temperature conditions. The impact of raw milk, storage time, chilled temperature conditions on the compliance regarding these standards was analysed in details. The upcoming climate change may affect the storage temperature but also the milk quality due to potential cow heat stress.
Significance: The application of probabilistic modelling tools to assess current exposure can be expanded to other food systems. The ways on how the variability and uncertainty from data inputs and storage scenarios were addressed, using second order Monte Carlo procedure, provide an added-value to develop realistic exposure assessment models and suggest mitigation options.