论文标题
EpiLocal：用于局部流行监测的实时工具
Epilocal: a real-time tool for local epidemic monitoring
论文作者
论文摘要
我们描述了Emocal，这是一个简单的R程序，旨在自动下载有关所有意大利省和地区的感染SARS-COV-2案例的最新数据，并提供简单的描述性分析。对于每个省，每天都有报告的受感染病例的累积数量。此外，目前住院的患者人数（是否分别用于重症监护），并且在该地区一级可以使用死者的累积数量。通过具有对数链路函数和多项式回归的泊松通用线性模型对数据进行分析。对于累积数据，我们还考虑了危险函数的逻辑参数化。根据相应的估计参数的统计意义和拟合优点评估，进行自动模型选择以在不同的模型规范之间进行选择。所选模型用于产生计数增长率的最新估计。将结果绘制在该国的地图上，以允许对具有不同患病率和增长率的区域的地理分布进行视觉评估。
We describe Epilocal, a simple R program designed to automatically download the most recent data on reported infected SARS-CoV-2 cases for all Italian provinces and regions, and to provide a simple descriptive analysis. For each province the cumulative number of reported infected cases is available each day. In addition, the current numbers of hospitalized patients (separately for intensive care or not) and the cumulative number of deceased individuals are available at the region level. The data are analyzed through Poisson generalized linear models with logarithmic link function and polynomial regression on time. For cumulative data, we also consider a logistic parameterisation of the hazard function. Automatic model selection is performed to choose among the different model specifications, based on the statistical significance of the corresponding estimated parameters and on goodness-of-fit assessment. The chosen model is used to produce up-to-today estimates of the growth rate of the counts. Results are plotted on a map of the country to allow for a visual assessment of the geographic distribution of the areas with differential prevalence and rates of growth.
