Vaccine hesitancy and erosion of herd immunity: harnessing big data to forecast disease re-emergence

项目摘要 世界卫生组织（世卫组织）估计，5岁以下儿童中有150万人死亡， 由于疫苗可预防的疾病（VPD），主要是在发展中国家， 消除所需的覆盖面已证明是一项挑战。令人惊讶的是，即使在富裕的国家， 在美国，尽管历史上疫苗覆盖率很高，但群体免疫已被证明是难以捉摸的。这在很大程度上是 由于疫苗犹豫的现象，或希望延迟或拒绝接种，这构成了障碍 为麻疹和脊髓灰质炎等VPD的全球疾病根除工作提供最后一英里。实现 “最后阶段”需要及早发现疫苗犹豫，并清楚地了解其影响 这种行为对VPD的（重新）出现和传播的影响。现在迫切需要的是一种整合的手段 在机械传播模型中的不同数据源，以量化人群免疫力， 区别对待不同的行动方针。拟议的工作旨在利用一种独特的 高分辨率医疗索赔数据集，以地理定位疫苗接种犹豫，以确定其社会 经济驱动因素，并剖析这种行为与麻疹和百日咳的流行病学后果， 案例研究。最终，我们的目标是开发一个预测模型的空间和时间动态的， 预防接种行为和受影响传染病的发病率。这种模式将是我们能够 预测疾病再次出现，并制定公共卫生政策，积极改变疫苗接种格局。

Project Summary The World Health Organization (WHO) estimates that 1.5 million deaths among children under 5 years are due to vaccine-preventable diseases (VPDs), predominantly in the developing world where achieving vaccine coverage necessary for elimination has proved a challenge. Surprisingly, even in wealthy countries such as the United States, herd immunity has proved elusive despite historically high vaccine coverage. This is in large part due to the phenomenon of vaccine hesitancy, or the desire to delay or refuse vaccination, which poses a barrier to covering the last mile in global disease eradication efforts for VPDs such as measles and polio. Achieving the “endgame” will require early detection of vaccine hesitancy, as well as a clear understanding of the impact of this behavior on the (re)emergence and spread of VPDs. What is urgently needed is a means of integrating disparate data sources within a mechanistic transmission model in order to quantify population immunity and discriminate among alternative courses of action. The proposed work aims to harness the potential of a unique high-resolution data set of medical claims to geographically localize vaccination hesitancy, to identify its socio- economic drivers, and dissect the epidemiological consequences of this behavior with measles and pertussis as case studies. Ultimately, our goal is to develop a predictive model for the spatial and temporal dynamics of both vaccination behavior and the incidence of affected infectious diseases. Such a model will be key in our ability to forecast disease re-emergence and design public health policy to positively change the vaccination landscape.