Professional Documents
Culture Documents
Presented by:
Edmond Sacla Aide (2159278)
Melvin Estolano (2159122)
Fan Huang (2263527)
Lecturer:
Prof. Dr. Christel Faes
Prof. Dr. Marc Aerts
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
Project Summary
In this paper, we will perform a quantitative microbial risk assessment (QMRA) to evaluate the
risk posed by Campylobacteriosis infection considering the food pathway for the whole döner kebab
processing.
Firstly, we will describe and simulate the number of Campylobacter in positive samples of chicken
meat. Secondly, we will describe and simulate the number of Campylobacter in döner kebab before
and after cooking. From this, we will estimate the contamination level. Third, we will simulate
the probability of infection and illness and predict the number of cases per 106 servings. Fourth,
we will investigate the impact of ingested dose on infection level by using Spearman’s rank corre-
lation coefficients. Lastly, three scenarios for sensitivity analyses will be performed to estimate the
number of Campylobacter in döner kebab.
All analyses were done using R software version 4.2.1. All simulations are seeded at 1 to reproduce
a particular sequence of random numbers.
1
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
Contents
1 Simulation of the number of Campylobacter in positive samples of chicken meat 3
1.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
7 Third sensitivity analysis: Impact of the negative association between the amount
an individual would consume and the concentration in the meal 13
7.1 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
7.2 Result . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
1.2 Result
Figure 1: Histogram plot of the mean Campylobacter concentration in positive samples of chicken
meat: (a) Log scale (b) Re-scaled
Table 1: Summary statistics for the mean Campylobacter concentration in positive samples of
chicken meat
3
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
Figure 2: Proportion of mean Campylobacter concentration (cfu/g) from raw chicken meat
Figure 1 shows the histogram plot of the mean Campylobacter concentration for both the log and
re-scaled version of the simulated data. Rescaling back to the original unit of concentration (cfu/g)
it can be seen that the distribution is skewed. Moreover, from the rescaled unit of concentration
in table 1, it can be seen that the median Campylobacter concentration is 45.616 cfu/g.
In addition, the pie chart of mean Campylobacter concentration from raw chicken meat was re-
constructed, revealing that the proportion of categorized concentration in the simulated data is
different as compared to the original result in the article.
Ckebabprepaf tercooking = Ckebab × 10Hs eas × 10Mm ar × 10Sf reezing × 10HS urvCooking
To estimate the number of Campylobacter in döner kebab before and after cooking, we made use
of Monte Carlo simulations. The simulation algorithm is as follows:
4
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
• Creating a vector of the weight of kebab skewers to which a vector of the weight of each
chicken was simulated from a normal distribution with mChickenunit = 1.928 and standard
deviation sdChickenunit = 0.178 until 1 kebab skewer is at least 10kg.
• Calculating the number of Campylobacter per chicken by adding the simulated log chicken’s
weight (logUchicken,i ) and the simulated raw meat Campylobacter concentration (Crawmeat ).
• Calculating the number of bacteria per contaminated kebab by summing the number of
Campylobacter to the power of 10 if its contaminated.
• Calculating the decrease of Campylobacter in the freezing step (if the time is smaller than 1,
then SF reezing = −0.56 × ST ime , otherwise SF reezing = −0.465 − 0.095 × ST ime )
• Calculating the concentration in the kebab at the moment of consumption before and after
cooking
2.2 Result
From the histogram plot, we can see that the distribution of the weight of kebab skewers is skewed.
It can also be seen that it has two peaks, around 10,000 grams and 11,500 grams. Referring from
table ??, the average Kebab skewer weight is 11,216 grams.
5
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
From the plot of the histogram, the distribution of the estimated Campylobacter in the kebab
skewer is highly skewed for both before and after cooking.
Figure 4: Histogram plot of the number of Campylobacter : (a) Before cooking (b) After cooking
From table ??, it can be seen that the average number estimated number of Campylobacter con-
centration in the kebab skewer before cooking gives a huge value as compared with after cooking.
So the concentration of Campylobacter in a kebab before cooking is around 22830 cfu/g while after
cooking this concentration is reduced to around 50 cfu/g.
Table 3: Summary statistics for the estimated number of Campylobacter concentration in the kebab
skewer
Figure 5 shows the impact of cooking on the estimated number of Campylobacter concentration
in the kebab skewer (10 kg). Similar interpretations can be gained that the estimated number of
Campylobacter concentration in the kebab gradually decreases after cooking. Thus, the cooking
step is very important in the kebab production chain in reducing the concentration of Campylobac-
ter.
6
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
7
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
It was also noted by previous studies (Black et al. [1988], Havelaar et al. [2000], Nauta et al. [2007])
that the probability of illness given the infection is given by 1/3. Thus, we can calculate the prob-
ability of illness by Pill = M.PInf × Pill|inf .
Therefore, the cases per year, assuming that the estimated Kebab’s serving persons per year
(NServed ) is 1,000,000, can be calculated by multiplying the Probability of illness with the Ke-
bab’s serving persons per year.
3.2 Result
The mean probability of infection is given by M.Pinf which is equal to 0.043212. In addition,
the probability of illness Pill is equal to 0.014404. By looking at the histogram of probability of
infection, it can be seen that the distribution is left skewed. Moreover, the plot can also tell us that
if we relate this to a person eating kebab, he/she can have a probability of infection approximately
about 1.4%. Although this figure seems low, if we take the variability into account, then eating a
kebab that does not have good hygiene can still have risk of being contaminated.
The number of cases per year was derived by multiplying the Probability of illness with the Kebab’s
serving persons per year. This is equivalent to 14,404 cases per year.
8
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
9
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
4.2 Result
10
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
5.2 Result
Since it is easier to obtain a value that is smaller than 0.862 from the triangular distribution that
we specify, we can see from table 2 that when the contamination rate is not fixed, the number of
Campylobacter in döner kebab tends to be smaller.
Table 5: Mean of number of Campylobacter in döner kebab with fixed prevalence and sampled
prevalence
Fixed Prevalence Sampled Prevalence
Mean of number of Campylobacter 18,748,599,212 18,259,281,453
11
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
6.2 Result
When the amount of consumption is sampled from a triangular distribution, we can see from the
histograms that the probability to obtain a döner kebab with high ingested does is higher, leading to
higher cases per year. When the portion consumed is fixed, the predicted number of cases is about
14,404 per year, whereas when the portion consumed is sampled from a triangular distribution, the
predicted number of cases becomes around 15,025 per year.
Figure 9: Histogram of ingested dose with fixed consumption (left) and sampled consumption
(right)
12
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
7.2 Result
From the histograms, we can see that when there’s a negative association, the frequency to obtain
a döner kebab with high ingested does is lower (shorter tail), leading to lower cases per year. When
the portion consumed is fixed, the predicted number of cases per year is about 14,404 per year,
whereas when there’s negative association between consumption amount and the concentration in
the meal, the predicted number of cases decreases to 14,396 per year.
Figure 10: Histogram of ingested dose with fixed consumption (left) and negative association (right)
13
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
References
Robert E. Black, Myron M. Levine, Mary Lou Clements, Timothy P. Hughes, and Martin J.
Blaser. Experimental Campylobacter jejuni Infection in Humans. The Journal of Infectious
Diseases, 157(3):472–479, 03 1988. ISSN 0022-1899. doi: 10.1093/infdis/157.3.472. URL
https://doi.org/10.1093/infdis/157.3.472.
European Food Safety Authority. Analysis of the baseline survey on the prevalence of campylobacter
in broiler batches and of campylobacter and salmonella on broiler carcasses, in the eu, 2008-part
b: Analysis of factors associated with salmonella contamination of broiler carcasses. EFSA
Journal, 9(2):2017, 2011.
Sara Messad, Taha-Mossadak Hamdi, Radia Bouhamed, Nadjia Ramdani-Bouguessa, and Mo-
hamed Tazir. Frequency of contamination and antimicrobial resistance of thermotolerant campy-
lobacter isolated from some broiler farms and slaughterhouses in the region of algiers. Food
control, 40:324–328, 2014.
14
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
Appendices
R Codes
# - - - - - ----- Libraries
library ( triangle )
library ( tidyverse )
library ( Rlab )
library ( ggplot2 )
15
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
num _ Camp _ per _ chicken [ i ] <- list ( log ( w _ chickens [[ i ]] , base =10) + C _ raw _ meat [ i ])
}
16
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
for ( i in 1: sims ) {
D _ ing [ i ] <- rbinom (1 , round ( C _ kebab _ prep [ i ]) ,150 / U _ kebab [ i ])
}
# - - - - - ----- Correlation
r1 <-c ( cor ( P _ inf , t _ heating , method = " spearman " , use = " complete . obs " ) ,
cor ( P _ inf , S _ freezing , method = " spearman " , use = " complete . obs " ) ,
cor ( P _ inf , log _ Dtr , method = " spearman " , use = " complete . obs " ) )
barplot ( r1 , horiz = TRUE ,
names . arg = c ( " cooking time " ,
" freezing time " ,
" log reduction time " ) ,
xlim = c ( -0.2 ,1) ,
xlab = " Correlation " )
mean ( C _ kebab )
mean ( C _ kebab _ SA )
17
Microbial Risk Assessment (MRA) - Project1 (2022–2023)
18