Environmental Program, University of Vermont, 05401, Burlington, VT, USA.
Environ Monit Assess. 1987 Mar;8(2):151-61. doi: 10.1007/BF00403110.
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.
许多小型社区都担心环境污染物对健康的影响。纽约州奥索格县的奥威尔镇就是这样一个社区，为了确定其癌症发病率与该县和该州相比是否显著偏高，对其进行了调查。对居住在一个特定区域的居民进行了一项试点健康研究。调查结果表明，癌症发病率高于奥索格县和纽约州，而男性的癌症死亡率较高，女性的癌症死亡率较低。对数据的进一步分析表明，通常用于分析健康调查的方法可能不适用于小型社区。特别是，当从小样本中得出标准化发病率和死亡率时，一些参数统计，如卡方分析，可能会显示出人群之间的显著差异，但从这些数据计算得出的标准偏差却大得令人对分析结果产生怀疑。研究人员已经从理论上讨论了这个问题，但迄今为止，在已发表的文献中还没有出现说明这些困难的实证研究。对于小型社区的健康研究，应更多地考虑风险比和标准偏差的含义。
