Increased surveillance for both infections would
provide useful information and it is hoped that the risk
assessment could be used as a tool to prompt such
action. Throughout the analyses assumptions have
been clearly stated. Acquisition of further data, or
refinement of existing data should enable the assumptions
to be reduced. The more refined the data in the
assessment, the more confidence we can place in the
resultant estimate. A risk assessment should not be
considered complete; as new data become available,
the model could be up-dated to capture and include
the new information. Even with incomplete information,
a risk assessment is useful. It not only provides a
crude estimate of risk, but also underlines where data
clarification and expansion would be most useful.
Assumptions can be altered to demonstrate both the
impact caused by the lack of data, and also to judge
the potential gains to be made through acquisition of
more data. Tables 2 and 3 show the effect on the risk
estimates of different assumptions: the infection to
illness ratio; the number of people consuming sprouts
in the population; and the frequency with which
sprouts are consumed during the year. These scenarios
are not exhaustive in capturing the possibilities that
exist given our state of knowledge. However, they do
show that even with these basic scenarios the
estimates for Cryptosporidium range from 150 to
1970 cases per year and for Giardia the estimates
range from 190 to 2450 cases per year. These ranges
do not account for all sources of variability and
uncertainty (for instance: the range of concentrations;
the uncertainty in the prevalence; the uncertainty in
viability/genotype; the uncertainty in the doseresponse,
and its dependence upon the immune status
of the infected, etc.). A full quantification and
characterisation of the uncertainty and variability
was deemed to be non-essential for the current work,
given the obvious lack of some very basic information.
As a result, the current analysis simply demonstrates
the impact that alternative assumptions can
have on the final estimates. As the data gaps are filled,
the process can evolve, increasing in complexity and
providing a more realistic estimate. Areas where data
are particularly lacking, or where further or more
refined data would be most useful, are described in