Linking transmission models and data analysis in infectious disease epidemiology

描述（由申请人提供）：传染病流行病学使用主要从慢性病流行病学借用的统计方法。这些方法将感染视为独立事件，忽略了传染病的定义特征。我们的研究试图将随机流行病模型与传染病数据分析联系起来，包括研究设计和因果推断。在整个项目中，我们打算将我们开发的方法应用于华盛顿大学和孟加拉国国际腹泻病研究中心（ICDDR.B）以前和正在进行的研究。我们必须了解什么是重要的测量，然后学习如何测量它。通过发展随机流行病模型的分析方法，我们希望区分传染病的鲁棒性和参数敏感性行为。然后，我们将调整方法，从生存分析到来自传播模型的可能性，以开发用于疾病传播的点估计和假设检验的方法。通过模拟，我们可以研究这些检验和估计量在真实世界传染病数据中常见情况下的行为，例如未观察到的感染时间、无症状感染以及季节或地理变化。由于疾病传播可能导致复杂的时间依赖性混杂，我们必须学会识别需要控制混杂的非标准技术的情况，并将其专门用于传染病流行病学。最后，我们将调整队列、病例对照和病例交叉研究设计，以获得有效和有效的估计暴露对传染性、易感性和疾病传播的其他方面的影响，我们将探索疫苗和干预试验的新设计和随机化程序。一旦充分开发，所有这些方法都需要在流行病学家，生物统计学家和其他公共卫生从业人员可以访问的计算机程序中实现。尽管世纪中期的情况令人乐观，但传染病仍然是全世界公共卫生的巨大负担和威胁。通过开发解释疾病传播的研究设计和统计分析方法，我们希望为未来的流行病学家和生物统计学家提供工具，以便更详细和准确地分析暴露和干预措施影响感染传播的机制。对这些机制的详细、针对具体地点的了解将有助于更及时、更有效地采取地方干预措施，最终保护全球公共卫生。

DESCRIPTION (provided by applicant): Infectious disease epidemiology uses statistical methods borrowed largely from chronic disease epidemiology. These methods treat infections as independent events, ignoring the defining feature of infectious disease. Our research attempts to link stochastic epidemic models to the analysis of infectious disease data, including study design and causal inference. Throughout this project, we intend to apply the methods we develop to previous and ongoing research at the University of Washington and the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR.B). We must learn what is important to measure and then learn how to measure it. By developing methods for the analysis of stochastic epidemic models, we hope to distinguish between robust and parameter-sensitive behaviors of infectious diseases. We will then adapt methods from survival analysis to likelihoods derived from transmission models to develop methods for point estimation and hypothesis testing that account for the transmission of disease. Through simulation, we can investigate the behavior of these tests and estimators in situations common in real-world infectious disease data, such as unobserved infection times, asymptomatic infections, and seasonal or geographic variation. Since disease transmission can lead to complex time-dependent confounding, we must learn to identify situations where non-standard techniques of controlling confounding are necessary and adapt them specifically for infectious disease epidemiology. Finally, we will adapt cohort, case-control, and case-crossover study designs to obtain valid and efficient estimates of the effects of exposures on infectiousness, susceptibility, and other aspects of disease transmission, and we will explore new designs and randomization procedures for vaccine and intervention trials. Once adequately developed, all of these methods need to be implemented in computer programs that are accessible to epidemiologists, biostatisticians, and other public health practitioners. Despite the optimism of the mid-twentieth century, infectious diseases remain a tremendous burden and threat to public health worldwide. By developing methods for study design and statistical analysis that account for the transmission of disease, we hope to provide future epidemiologists and biostatisticians with tools for a much more detailed and accurate analysis of the mechanisms by which exposures and interventions affect the spread of infections. Detailed, place-specific knowledge of these mechanisms will allow more timely and effective local interventions, ultimately protecting global public health.