Assessing cancer incidence in a small community

Many small communities are concerned about health effects from environmental pollutants. One such community, Orwell, in Oswego County, New York, was investigated to determine if the level of cancer was significantly high in comparison to the county and the state. A pilot health study was administered among residents living in a defined area. Results of the survey demonstrated that cancer incidence was higher than both Oswego County and New York State, while cancer mortality was higher for men and lower for woman. A closer examination of the data showed that the way in which health surveys are usually analyzed may not be appropriate for small communities. In particular, some parametric statistics, such as chi-square analysis, might show significant differences between populations when standardized incidence and mortality rates are derived from small samples, yet standard deviations calculated from these data are so large as to cast doubt on the analysis. Researchers have discussed this problem theoretically, but empirical studies illustrating the difficulties have not heretofore appeared in the published literature. For health studies in small communities, greater consideration should be given to risk ratios and the implications of standard deviations.