Statistical Model for evaluating health effects of dioxins around waste incinerators

公众非常关注的是，有害的生殖健康影响可能与城市固体废物焚化炉排放气体中的二恶英有关。最近，日本首次进行了大规模的全国流行病学研究，研究了居住在日本63个高二恶英排放水平（超过80纳克国际毒性当量TEQ/m^3）的生活垃圾焚化炉10公里范围内的母亲与不良生殖结果的关系。我们研究的主要目的是研究统计方法来分析上述研究中收集的数据，并提出一种适用于更一般的问题的方法，即围绕假定的点源增加疾病风险。作为检测此类聚类的统计方法，Stone的最大似然比检验和分数检验是比较流行的，分数检验是对每一种情况与点源距离的倒数进行评分。在我们的研究中，我们提出了基于小区域数据的分数测试的扩展，因为(i)它允许我们在预先指定的参数暴露函数中选择最佳值，以避免多个测试问题；（ii）它可以应用于有害物质水平倾向于在距离点源一定距离处达到峰值的可能情况。用假设数据检验和说明了所提出的试验的性质及其功效，并对上述流行病学研究的研究设计进行了功效计算。仿真研究表明，在所考虑的所有替代方案中，所提出的测试的功率高于Stone的测试。

There has been great public concern that adverse reproductive health effects may be associated with dioxins in emission gases from municipal solid waste (MSW) incinerators. Recently, Japan's first-ever large scale nationwide epidemiological study has begun to examine the association of adverse reproductive outcomes with mothers living within 10km from 63 MSW incinerators with high dioxin emission levels (above 80ng international toxic equivalents TEQ/m^3) in Japan. The primary purpose of our study was to investigate the statistical method to analyze the data collected in the above study and also to propose a method applicable to more general problems of raised disease risk around putative point sources.As statistical method to detect such clusters, Stone's maximum likelihood ratio test and score tests which score each case the reciprocal of the distance from a point source are popular. In our study, we have proposed extensions of score tests based on small area data in that (i)it allows us to select the best among prespecified parametric exposure functions to avoid multiple testing problems and (ii)it can be applied to a possible situation where the hazardous substance levels tend to peak at some distance from a point source. The properties of the proposed tests and their power have been examined and illustrated with hypothetical data and also the power calculation has been conducted on the study design of above-stated epidemiological study. Simulation study shows that the powers of the proposed tests are higher than that of Stone's test over all the alternatives considered.