When is an apparent excess of neurologic cases epidemiologically significant?

Insights into disease may be gained in a variety of different ways. Sometimes epidemiologists, those who study the occurrence of a disease, are able to show that a disease correlates with a specific food, poison, infectious organism, or parasite. A community physician may occasionally notice that he has seen more cases of a disease than might be expected. He might then report the existence of a cluster of cases in his community. Experience indicates that a cluster of a new syndrome is likely to prove significant, as have clusters of AIDS in the last decade or clusters of Legionnaires disease in the previous one. However, clusters of more common diseases observed in communities are likely to provide little information or to have little statistical validity. The authors discuss the statistical problems that make it difficult to determine if an observed cluster is actually related to some aspect of the disease or simply the result of random chance. For example, a physician in a community of 40,000 might suspect that more cases of amyotrophic lateral sclerosis have been seen over the last 10 years than would be expected in a community of this size; analysis of case records may confirm this hunch. To establish that this is a significant cluster of cases, other factors must be considered, such as the geographic stability of the community. If 20 percent of the people in this community move out each year and the same number move in, within approximately 10 years about 120,000 different people will pass through the physicians offices of the community. Three times as many cases of any disease are likely to be seen than would be expected in a population of 40,000. Other considerations may influence the appearance of clusters as well. It would be inappropriate to ignore all reports of disease clusters. Similarly, it would be impossible to follow-up all such reports with a detailed investigation. Careful data analysis is necessary to distinguish the community-based reports of disease clusters that are more likely to yield useful epidemiological insights. (Consumer Summary produced by Reliance Medical Information, Inc.)