Food Science and Technology: Microbial Risk Assessment: Salmonella in Chicken

This is a reconstruction of paper from Oscar 2004. He used @Risk software to carry out the risk assessment. I am reconstructing his paper using R- software taking the same model input values. So for details, please refer to the paper of Oscar 2004. The following are the important model parameters described in that paper.

Model design

The following 5 nodes were considered during model development.

Node 1- Retail

"The incidence of Salmonella contamination of whole chickens ranged from 0% to 100% with a median value of 30%. The MPN of Salmonella on contaminated whole chickens is low, often below 10 MPN, whereas the maximum MPN per chicken ranges from >300 to >1100. A minimum value of 1 MPN (by definition, the minimum level of contamination that is possible), a median value of 10 MPN, and a maximum value of 450 MPN per chicken were used to define the input settings for the PERT distribution for the extent of Salmonella contamination of ready-to-cook whole chickens at retail."

Node 2- Consumer Transport

"Time and temperature data of fresh meat showed that the time of consumer transport ranged from 0.2 to 6.3 h with a median time of 1 h. The temperature of fresh meat when it arrived in the consumer’s home ranged from minus 3.9 to 21.1 ^oC with a median temperature of 7.8 ^oC. These values were used to define the input settings for the PERT distributions for time and temperature in the growth model, which was then simulated. A predicted incidence of potential growth events during consumer transport of 0.02% was obtained and was used to define the incidence of growth events for this node.

The growth model showed that the extent of potential growth events would range from 0.0005 to 0.15 log cycles with a median value of 0.04 log cycles. These values were used to define the input settings for the PERT distribution for the extent of potential Salmonella growth during consumer transport."

Node 3- Cooking

"A PERT distribution was used to model the variability and uncertainty of final cooked temperature, which was reported to range from 26 to 93 ^oC with a mean of 62 ^oC. However, PERT settings of 55, 62 and 70 ^oC were used because the thermal inactivation model only had a temperature range of 55 to 70 ^oC. It was assumed that chickens were cooked in a home oven and that Salmonella were exposed to the final cooked temperature for a minimum of 15 min, a median of 30 min, and a maximum of 45 min.

Results of the cooking model simulation indicated that the log cycle reduction of Salmonella ranged from minus 96 to minus 0.83 with a median value of minus 8.1; these values were used to define the PERT distribution for cooking."

Node 4- Serving

"The incidence of food handling mistakes that could lead to cross-contamination averaged 28% among three consumer surveys and was the value used to define the incidence of cross-contamination events for this node. The average values among studies of 0.021, 0.057 and 0.24 for the minimum, median and maximum transfer rates, respectively, of Salmonella were used to define the PERT distribution for cross-contamination during serving."

Node 5- Consumption

"Based on these data, the input settings for the PERT distribution for illness dose were a minimum of 1 log MPN, a median of 3 log MPN and a maximum of 7 log MPN. Dose response was modeled as a discrete event. For an illness to occur from consumption of a cooked chicken, the dose of Salmonella consumed had to exceed the illness dose that was randomly assigned to that chicken or iteration by the model. Thus, the outcome of the dose response was discrete: no illness or illness."

Model simulation

The model was simulated with R-software using mc2d package with 100001 numbers of iteration. The code has not been provided in this page, to keep it simple.

Model Output:

The output from the model in R-software is given below.

> summary(Risk)

Retail_log_contamination_Input1 :