Page 1 of 5
Assignment 5 Quantitative Microbial Risk Assessment
Risk Modelling of Reducing Campylobacter in Broiler Chicken
Campylobacter is a major bacterial cause of infectious gastroenteritis all over the world.
Handling and consumption of chicken has been identified as important risk factors (Kapperud et
al., 2003; Friedman et al., 2004; Wingstrand et al., 2006). A thorough assessment of
thermophilic Campylobacter spp. in broiler chickens through the systematic approach of
quantitative microbial risk assessment (QMRA) is important from both public health and
international trade perspectives. The purpose of this work is to develop a QMRA that attempts
to understand how the probability of campylobacter infections among consumers having is
influenced by various factors during chicken processing, consumer handling, meal preparation
and finally consumption. Results from this assessment will enable the consideration of the
broadest range of intervention strategies throughout the processing-to-fork chain. The risk
characterization estimates the probability of campylobacteriosis per serving of chicken meals,
which are cooked in the consumers’ kitchen for immediate consumption. A schematic
representation of the risk assessment is shown in the Figure 1.
Figure 1. Schematic representation of the risk assessment of Campylobacter spp. in broiler
chicken.
In this project, you are expected to develop a QMRA model to translate the level (concentration)
or frequency of contamination (prevalence) of a product into human health risk outcome.
Specifically, the impact of the use chlorine-based and other alternative disinfectants used in
washing and chilling steps will be evaluated on public health protection.
For you to complete the project, the following hypothetical inputs and assumptions are provided.
1. The current project is primarily interested in the impact of several chlorine-based and
other disinfectants on the contamination on chicken products, which are most commonly
used in the processing plant for washing and/or chilling steps. Both of the steps are near
the end of the processing chain, hence the earlier processing steps will be collapsed.
For this reason, the pre-washing prevalence and concentration will be used as the initial
inputs for the subsequent simulations.
Page 2 of 5
2. The prevalence of Campylobacter positive on carcasses is 80% (deterministic input),
immediately before washing.
3. The concentration on positive carcasses immediately before washing follows a normal
distribution with a mean of 3.8 log10 CFU/carcass and standard deviation of 1.3 log10
CFU/carcass.
4. The primary effect from washing is the physical removal of contamination rather than a
chemical decontamination effect. The physical action of washing with water alone
through the inside-outside bird washer (IOBW) can result in a reduction of 2.1-2.8 log10
CFU of Campylobacter.
5. The use of chlorine in IOBW produces a slightly greater reduction compared with just
water alone. It is assumed that the effect of adding chlorine to the wash water could
produce additional reduction anywhere from no effect to 0.1 log10 CFU.
6. Acidified Sodium Chlorite (ASC) is another disinfectant that is approved for broiler
chicken processing. ASC spray is tended not to be used in IOBW or high volume
system, and therefore they would be less likely to exert a physical reduction effect.
However, studies show that the decontamination washing using ASC spray at
concentrations ranging from 600 to 1200 ml/L can result in log reductions from 0.9 to 3.8
log10 CFU (covering both physical and chemical decontamination effects).
7. Trisodium phosphate (TSP) is the other alternative to chlorine-based disinfectant that is
evaluated in this project. Similar to ASC, TSP spray is tended not to be used in IOBW or
high volume system, and therefore they would be less likely to exert a physical reduction
effect. However, studies show the use of TSP at 12% solution was found to reduce
Campylobacter by on average of 1.03 log10 CFU and standard deviation of 0.6 log10
CFU, when sprayed on carcass for 15s (covering both physical and chemical
decontamination effects).
8. Although there is a consistent tend of decrease in the concentration of the bacteria after
washing, the effect of washing on the prevalence change remains uncertain. Therefore,
it is assumed that the prevalence of Campylobacter on carcasses will not change after
washing.
9. Chilling is the processing step right after washing. During chilling, when there is sufficient
free chlorine in the chill tank, the frequency of cross-contamination occurrence is
reduced; when it does occur, the amount of cross-contamination is less. The fraction of
positive chicken carcasses remaining positive after chilling ranges from a minimum value
of 9.3% to the maximum of 47.8% and most likely at 14.3% when chilled in immersion
chilling tank with sufficient free chlorine, while follows a distribution with a minimum,
most likely and maximum values as 14.3%, 22.9%, and 74.5%, respectively when
chlorine is not added.
10. The decontamination of Campylobacter through water immersion chilling without
chlorine added could be as low as 0.71 log10 CFU, high as 1.42 log10 CFU, but most
Page 3 of 5
likely as 1.06 log10 CFU. The decontamination of Campylobacter through water
immersion chilling with chlorine added at 50 mg/L could be as low as 1.23 log10 CFU,
high as 1.8 log10 CFU, but most likely as 1.52 log10 CFU.
11. After processing, it is assumed no growth or inactivation of Campylobacter on broiler
chicken products occur during transportation, at retail, or during storage at consumers’
home before cooking.
12. It is assumed chickens are always fully cooked with all Campylobacter bacteria killed
after consumers’ cooking.
13. However, cross-contamination can occur to transfer Campylobacter from contaminated
utensils, raw materials, hands to the cooked meals, due to consumers mishandling. We
have very limited knowledge on the occurrence of mishandling in consumers’ kitchen. It
is assumed that the mishandling can happen anywhere from no time to 100% of the
time.
14. When cross-contamination occurs during mishandling, the fraction of Campylobacter
originated from the raw whole carcass transferred to the cooked whole chicken can be
as low as 0.02% and as high as 10%.
15. It is assumed the Campylobacter cells are homogeneously distributed on the chicken
carcasses and meals throughout the chain covered in this model.
16. The dose-response parameter, probability of infection from one Campylobacter cell
(r_dose) follows the distribution of Beta(0.21, 59.95).
17. Average amount of chicken consumed per meal per person is 15.6% of a whole chicken
(a deterministic input).
18. The final risk estimate, per-serving risk of campylobacteriosis can be predicted by 1-(1-
r_dose)Dose, where Dose denotes the ingested dose of Campylobacter in CFU/serving at
the time of chicken meal consumption.
To complete this assignment, please address the following questions.
Task 1. Develop a conceptual model to describe the translation of cumulative changes in
prevalence and concentration of Campylobacter from washing/chilling at processing, through
preparation, consumption, to infections in humans, integrating the factors along that may
influence the dynamic of Campylobacter along the chain. (25 points)
Task 2. Develop a quantitative model based upon the conceptual model to estimate the
probability of campylobacteriosis per serving of chicken meal. In the baseline model, please
consider the application of IOBW washing step with plain water and immersion chilling with no
additives. (30 points)
Page 4 of 5
Task 3. Please provide the probability distribution of the final output estimate, and also list the
mean value, 2.5th percentile and 97.5th percentile values. Hint: In this model, the number of
iterations required to reach convergence is high (over 800,000). Let’s use 50,000 iterations in
your setting to complete the assignment. No need to separate variability and uncertainty in the
risk distribution, so the requested statistics are based on a first order Monte Carlo simulation.
(10 points)
Which input(s) have the greatest influence on the changes in output estimate of the model
developed in Task 2? Hint: sensitivity analysis visualized in tornado charts. (10 points)
Task 4. Please quantify the risks from alternative intervention scenarios to evaluate the efficacy
of the following intervention strategies by comparing with the baseline risk. What information
would you give to the risk managers to support their decision making on which intervention(s)
can be adopted? The following Campylobacter intervention scenarios can be considered: (25
points)
1) Baseline scenario: The baseline model includes a IOBW washing step with plain water
and a chilling step in water with no free chlorine;
2) Scenario 1: Use IOBW with chlorine at 50 mg/L;
3) Scenario 2: Use ASC decontamination spray;
4) Scenario 3: Use TSP decontamination spray;
5) Scenario 4: Use chlorine in chill tank at a concentration to ensure sufficient free chlorine;
6) Scenario 5: Combination of Scenario 1 and 4 (chlorine in wash water and chill tank)
7) Scenario 6: Combination of Scenario 2 and 4 (ASC decontamination spray and chlorine
in chill tank)
8) Scenario 7: Combination of Scenario 3 and 4 (TSP decontamination spray and chlorine
in chill tank)
Hint 1: the efficacy of intervention can be measured by borrowing the concept of attributable
fraction that we discussed in the Epidemiology section, and we provided a relevant example
when we had the overview of microbial risk assessment (page 22 of 02/04 lecture slides)