Page 342 - Biosystems Engineering

Food Safety Management 319

and preparation on the hazard in terms of levels and consumption
patterns (e.g., portion size). Model validation is a welcome addition
to the process where possible (Whiting and Buchanan 1994).
When we characterize risk, we recognize the fact that pathogenic
bacteria may vary widely in terms of their ability to cause illness,
depending on their numbers, virulence, and host susceptibility. Thus,
a dose response is used to translate the exposure assessment into a
response in terms of infected or ill individuals, depending on the
nature of the dose–response relationship. The dose–response assess-
ment evaluates the impact of a certain level of exposure on human
health in terms of illness or mortality levels. Changes in exposure can
be assessed in terms of change in disease incidence or mortality rate.
Of course, susceptibility of the host and exposure levels will influ-
ence response levels and are critical in formulating a dose–response
relationship. Such information can be difficult to ascertain, usually
being based on outbreak studies or limited challenge tests; however,
some data are available (Teunis et al. 1996).
All the data generated in previous stages are pulled together in
the risk-characterization stage of the process. Input uncertainty and
variability, including measurement and sampling errors for microbi-
ological methods, need to be accounted for regularly (Nauta 2002).
This can be achieved by representing uncertain inputs using proba-
bility distributions. This allows us to identify those inputs having
the greatest impact on model inputs and allows us to identify risk-
reduction strategies. This can also direct future research efforts and
provide a scientific basis for policy decisions.

10.4.2 Methodologies
There is no standard methodology for conducting a risk assessment.
The type of assessment that is performed will vary depending on the
complexity of the system being assessed, the available resources, and
the original risk question posed by risk managers (Coleman and
Marks 1999). Qualitative assessments and fully quantitative assess-
ments, which incorporate elements of uncertainty and variability,
have been used. Qualitative risk assessments are usually conducted
initially to assess the risk of a hazard and use descriptors to detail the
level of risk. Descriptions of risk can vary from “negligible” to “low,”
“medium,” or “high,” resulting in a ranking or categorizing system. In
contrast, quantitative assessments generate numerical estimates of risk.
Qualitative risk assessments are used to give an initial estimate of risk.
They usually require less time and resources than a fully quantitative
risk assessment. HACCP is primarily a qualitative approach to food
safety. Thus, it does not lend itself to assessing the effects of parameter
uncertainty or the effect of control measures (Hoornstra et al. 2001).
This generates the need for a quantitative approach that can be used
to provide more insight into the effect of control measures.