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.