Methods to Estimate the Effect of Interventions on the Incidence and Transmission of Tuberculosis

结核病(TB)是全球传染病死亡的主要原因，最近超过了艾滋病毒/艾滋病。 世界卫生组织已经制定了最终消灭结核病的目标，但几乎没有可靠的监测工具 朝着这一目标取得进展。此外，我们对结核病的空间和人口变异性的理解 传播和发病率不足，使有限资源的分配具有挑战性。这个项目 建议开发使用以下数据源实时估计结核病传播和发病率的方法 随处可见。在目标1中，我们将改进对疾病的序列间隔和持续时间的估计 在诊断时使用现有数据(我们的和已发布的)。这些估计将利用统计数据 对这些方法进行新的修改，使其与结核病相关。这些估计提供了 监测结核病的基础，进一步建模工作以指导政策，并提供关于结核病的有价值的见解 流行病学。目标2将侧重于开发实时估计结核病再生量的方法， 提供对传播模式的洞察。我们建议修改我们和其他人为其开发的方法 结核病潜伏期短的疾病，依赖于常规收集的监测数据。我们将延长这些措施 在空间上估计生殖数量的异质性的方法，跨越人口统计特征， 并根据有无耐药情况进行分析。在目标3中，我们解决了评估的挑战 疾病发病率。目前的估计是在全国范围内给出的，结果是根据新的 个别国家向世界卫生组织提供的确诊个体。此数据不提供粒度 估计发病率，并受到重大报告问题的阻碍，而这些问题还没有得到很好的理解。我们 提出了一种新的贝叶斯方法来从结核病流行调查中估计结核病发病率，使用我们的 目标1.结核病流行情况调查是详细的整群抽样调查， 所有参与者的空间、临床和人口统计信息。他们在整个地区表演，有一个 结核病的高负担。这一方法将提供更详细的结核病发病率和其 异质性，允许更有针对性地分配资源。我们将创建R包来传播 我们开发的所有方法都旨在广泛使用。这项工作将使决策者能够更好地 了解结核病发病率以及疾病负担最重的群体和地区。 这些知识将指导干预工作，最终目标是消除艾滋病。

Tuberculosis (TB) is the leading cause of infectious disease deaths globally, recently surpassing HIV/AIDS. The WHO has set goals for the eventual elimination of TB, but there are few, if any, reliable tools to monitor progress toward this goal. Additionally our understanding of the spatial and demographic variability of TB transmission and incidence is inadequate, making the allocation of limited resources challenging. This project proposes to develop methods to estimate transmission and incidence of TB in real time using data sources that are widely available. In Aim 1, we will develop improved estimates of the serial interval and duration of disease at time of diagnosis by using existing data (ours and published). These estimates will make use of statistical methods and novel modifications of these approaches to make them relevant to TB. These estimates provide the foundation of monitoring TB, further modeling work to guide policy, and provide valuable insights into TB epidemiology. Aim 2 will focus on developing methods to estimate the reproductive number of TB in real time, providing insight on transmission patterns. We propose to modify methods we and others have developed for diseases with short latency to TB, relying on routinely collected surveillance data. We will extend these methods to estimate heterogeneities in the reproductive number spatially, across demographic characteristics, and according the presence/absence of drug resistance. In Aim 3 we address the challenge of estimating disease incidence. Current estimates are given on a countrywide basis and result from annual reports of newly diagnosed individuals supplied to the WHO by individual countries. This data does not provide granularity in incidence estimates and is hampered by substantial reporting issues, which are not well-understood. We propose a novel Bayesian approach to estimate TB incidence from TB prevalence surveys, using our estimates of duration derived in Aim 1. TB prevalence surveys are detailed cluster sampled surveys with spatial, clinical, and demographic information on all participants. They are performed throughout areas with a high burden of TB. This approach will provide more detailed information on TB incidence and its heterogeneities, allowing for more targeted allocation of resources. We will create R packages to disseminate the methods that we develop in all the aims for widespread use. This work will enable policy makers to better understand the incidence of TB as well as groups and areas that are carrying the highest burden of disease. This knowledge will guide intervention efforts, with the eventual goal of elimination.