Using Quantitative Risk Assessment and Accounting for Variability and Uncertainty Incorporating risk metrics into food safety regulations: L. monocytogenes in ready-to-eat deli meats 1 Daniel Gallagher Virginia Tech 12th Annual Joint Fera/JIFSAN Symposium Greenbelt, MD June 15-17, 2011
24
Embed
Using Quantitative Risk Assessment and Accounting for Variability and Uncertainty
Incorporating risk metrics into food safety regulations: L. monocytogenes in ready-to-eat deli meats. Using Quantitative Risk Assessment and Accounting for Variability and Uncertainty . Daniel Gallagher Virginia Tech. 12th Annual Joint Fera /JIFSAN Symposium Greenbelt, MD - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Using Quantitative Risk Assessment and Accounting for Variability and Uncertainty
Incorporating risk metrics into food safety regulations: L. monocytogenes in ready-to-eat deli meats
Based on a target ALOP and industry response, an uncertainty distribution for the PO was calculated. Different quantiles of this PO distribution were then set as the regulatory PO and the resulting uncertainty distribution of risk per serving generated.
18Plant Lm Distribution
For a fixed ALOP, different industry response assumptions lead to different POs
19
Plant PO results by industry response Target ALOP = -6.38 (Q50)
Plant PO, log10 cfu/g
-10 -8 -6 -4 -2 0 2Probability that risk per serving <= target ALOP (%
)
0
20
40
60
80
100
truncatedshiftedfixed
20
Industry Risk per ServingDifferent Uncertainty Assumptions
Mean Risk of Illness per Serving, log10
-7.5 -7.0 -6.5 -6.0 -5.5 -5.0
Cum
ulative Percentage (%)
0
20
40
60
80
100
industry, baselineindustry, dose response uncertaintyindustry, increased GI uncertainty
Mean Risk of Illness per Serving, log10
-7.5 -7.0 -6.5 -6.0 -5.5 -5.0
Cum
ulative Percentage (%)
0
20
40
60
80
100
industry, baselineindustry, dose response uncertaintyindustry, increased GI uncertainty
Mean Risk of Illness per Serving, log10
-7.5 -7.0 -6.5 -6.0 -5.5 -5.0
Cum
ulative Percentage (%)
0
20
40
60
80
100
industry, baselineindustry, dose response uncertaintyindustry, increased GI uncertainty
21
Incorporating Dose-Response UncertaintyTruncated industry response
Plant PO, log10 cfu/g
-35 -30 -25 -20 -15 -10 -5 0
Probability Risk per Serving <= target ALO
P (%)
0
20
40
60
80
100
Baseline, target ALOP = -6.41 (Q25 of baseline)DR uncertainty, target ALOP = -6.41DR uncertainty, target ALOP = -6.66 (Q25 industry with DR uncertain)
Plant PO, log10 cfu/g
-35 -30 -25 -20 -15 -10 -5 0
Probability Risk per Serving <= target ALO
P (%)
0
20
40
60
80
100
Baseline, target ALOP = -6.41 (Q25 of baseline)DR uncertainty, target ALOP = -6.41DR uncertainty, target ALOP = -6.66 (Q25 industry with DR uncertain)
Plant PO, log10 cfu/g
-35 -30 -25 -20 -15 -10 -5 0
Probability Risk per Serving <= target ALO
P (%)
0
20
40
60
80
100
Baseline, target ALOP = -6.41 (Q25 of baseline)DR uncertainty, target ALOP = -6.41DR uncertainty, target ALOP = -6.66 (Q25 industry with DR uncertain)
22
Conclusions Incorporating uncertainty into risk metrics is
technically feasible computationally intensive much greater technical demands on risk managers with uncertainty, adapting PO to actual regulations
difficult▪ industry-wide compliance, not individual food plant▪ need to monitor for entire distribution▪ extremely broad PO uncertainty distributions
In practice, current levels of uncertainties limit applicability for L. monocytogenes
23
Acknowledgements
Funding: FSIS Project AG-3A94-P-08-0148
Co authors at FSIS and Virginia Tech Eric Ebel, Owen Gallagher, David
LaBarre, Michael Williams, Neal Golden, Janell Kause, Kerry Dearfield
Régis Pouillot for assistance with dose-response modeling.