Monte Carlo Simulations Assessing the Risk of Salmonellosis from Consumption of Almonds

Abstract

Recent outbreaks of salmonellosis associated with raw almonds have raised awareness of this food as a vector for foodborne illness. We performed a quantitative assessment of the risk of contracting salmonellosis from consumption of raw almonds, accounting for factors that become important after almonds reach the processor. We estimated the risk associated with the consumption of raw almonds and the risk reduction associated with almonds treated with a theoretical 5-log reduction process or treated with propylene oxide using a standard commercial process. Probability distributions were chosen to describe the chance of almond contamination and the effects of storage time, storage temperature, and processing from currently available data. A β-Poisson model for the dose-response relationship for Salmonella was obtained from published literature. The simulation estimated a 78% chance of one or more cases of salmonellosis per year from consumption of raw almonds. The application of a commercial propylene oxide treatment reduced this risk to 0.01%. Hypothetical 5-log reduction treatments with different standard deviations (±1, ±0.5, ±0.1, and ±0) reduced the predicted yearly risk of salmonellosis to 0.69, 0.35, 0.30, and 0.21%, respectively. These results suggest that the risk of one or more U.S. cases of salmonellosis per year from consumption of raw almonds can be reduced from 78% to less than 1% by using a process achieving a 5-log reduction in Salmonella with a process standard deviation as large as 1 log unit or by using a commercial propylene oxide treatment.