![](img009.GIF) |
There are several epidemiologic issues associated with the investigation
of a traditional environmental or community cluster. The entire analysis
is based on whether the disease cluster exists or not. In other words,
is it truly an increase of disease in time and/or space.
A major issue (and one extremely difficult to explain to
non-epidemiologists) is that any one disease cluster could occur solely
due to chance. This means that a group of children with the same type
of leukemia in the same neighborhood over a short period of time could
be ill due to chance rather than some aggregate exposure or etiology.
Furthermore, just because the disease cluster is statistically
significant by the previously discussed tests, there is still the
statistical possibility that this cluster exists only due to chance.
The latter issue is not helped by the fact that disease clusters,
especially of rare diseases, usually involve very few individuals so
that the application of statistical tests may be difficult. Furthermore,
the “index cases” or those persons who first were noted to be the
cluster should probably be excluded from any statistical analysis after
subsequent gathering of cases because they bias the result; the index
cases are what focused attention in the first place. Of note, most
investigators tend to apply the tests with the index cases and then
without as 2 separate analyses.
|