Space-time statistical models for detecting emerging outbreaks of rare events.

我们开发了新的统计模型，作为检测威胁人类健康的罕见事件新爆发的重要工具。现有的方法不能及时、适当地检测出形状不规则的暴发区。首先，我们提出了一种新的受限似然比，以避免现有的扫描统计量倾向于通过吞噬具有非提升风险的邻近区域来检测比真实簇大得多的簇的特性。其次，我们开发了一种新的时空扫描统计量，它将观察到的病例数与无条件预期的病例数进行比较，考虑了泊松均值的时间变化，并实现了一个暴发模型，以更及时和正确地捕获局部新发疾病暴发。所提出的模型以2006年日本基多九州小学缺课人数的每周监测数据为例。

We have developed new statistical models as important tools for detecting emerging outbreaks of rare events threatening human health. The existing methods cannot detect irregularly shaped outbreak areas timely and appropriately. First, we developed a new restricted likelihood ratio to avoid the property of existing scan statistics that tend to detect the cluster much larger than the true cluster by swallowing neighboring regions with non-elevated risk. Second, we developed a new space-time scan statistic which compares the observed number of cases with the unconditional expected number of cases, takes a time-to-time variation of Poisson mean into account and implements an outbreak model to capture localized emerging disease outbreaks more timely and correctly. The proposed models are illustrated with data from weekly surveillance of the number of absentees in primary schools in Kita-Kyushu, Japan, 2006.