Combining immunity and climate date streams to forecast infectious disease transmission

传染病的传播是由人口免疫力、人口行为以及在某些情况下气候变化之间复杂的相互作用驱动的。旨在预测未来传染病传播的预测模型在公共卫生规划和决策中非常有用，并为预警和反应系统提供信息。预测模型框架越来越多地纳入多个数据流，以更好地捕捉传染病传播的驱动因素，并提高预测的准确性和可靠性。该博士学位的目的是将血清学和气候数据整合到统计和数学建模框架中，以研究人口免疫力，气候和控制干预措施在驱动传染病动态中的作用。作为一个案例研究，我将重点关注多米尼加共和国的两个紧迫的健康威胁：SARS-CoV-2和登革热病毒。在前两章中，我将使用美国的纵向血清学数据来量化SARS-CoV-2再感染的风险。然后，这将为房室模型的参数化提供信息，以调查多米尼加共和国SARS-CoV-2传播的驱动因素，并随着疫苗接种覆盖率的增加模拟未来可能的流行情况。在我的第三章，我将量化的影响，气候变化对登革热的风险在多米尼加共和国使用时空贝叶斯建模框架和评估气候驱动的登革热预警系统的可行性。最后，我将开发一个传播动态模型，整合监测数据，血清学数据和气候数据，以调查多米尼加共和国登革热传播的驱动因素。我将开发和评估登革热疫情早期预警的预测，并将这些预测的成功与前一章中基于气候的统计方法进行比较。总的来说，我的目标是将新的数据流纳入建模框架，以改善传染病预测，这种方法可能适用于其他加勒比岛屿和小岛屿发展中国家。这项研究将与多米尼加共和国的流行病学总局（DIGEPI）密切合作，以确保项目产出符合多米尼加共和国的公共卫生需求，以及美国疾病控制和预防中心和泛美卫生组织/世卫组织等主要合作伙伴的需求。在这个项目中，我进一步提高了我的定量分析技能，特别是我的机械和统计建模知识，以及贝叶斯推理技术。此外，我期待着通过将气候驱动的建模方法的见解与机械传播动态建模方法相结合来提高我的跨学科技能，以改善传染病的预测。

Infectious disease transmission is driven by a complex interplay between population immunity, population behaviour and, in some instances, climate variation. Forecasting models aiming to predict future infectious disease transmission can be highly useful in public health planning and decision-making, and inform early warning and response systems. Forecasting model frameworks are increasingly incorporating multiple data streams to better capture the drivers of infectious disease transmission, and to improve the accuracy and reliability of forecasts. The aim of this PhD is to integrate serological and climate data within statistical and mathematical modelling frameworks to investigate the role of population immunity, climate and control interventions in driving infectious disease dynamics. As a case study, I will focus on two pressing health threats in the Dominican Republic: SARS-CoV-2 and dengue virus. In my first two chapters, I will quantify the risk of SARS-CoV-2 reinfection using longitudinal serological data from the United States. This will then inform the parameterisation of a compartmental model to investigate the drivers of SARS-CoV-2 transmission in the Dominican Republic and simulate possible future epidemic scenarios as vaccination coverage increases. In my third chapter, I will quantify the effect of climate variation on dengue risk in the Dominican Republic using a spatiotemporal Bayesian modelling framework and evaluate the viability of a climate-driven dengue early warning system. Finally, I will develop a transmission dynamic model integrating surveillance data, serological data, and climate data to investigate the drivers of dengue transmission in the Dominican Republic. I will develop and evaluate forecasts for early warning of dengue outbreaks and compare the success of these forecasts with the statistical climate-based approach in the previous chapter. Overall, I aim to integrate novel data streams within modelling frameworks to improve infectious disease forecasting, in an approach that may be generalisable to other Caribbean islands and Small Island Developing States. This research will be developed in close collaboration with the Dirección General de Epidemiologia (DIGEPI) in the Dominican Republic in order to ensure that project outputs align with the public health needs in the Dominican Republic, as well as the needs of key partners such as the US CDC and PAHO/WHO. During this project I am furthering my quantitative analysis skills, in particular my knowledge of mechanistic and statistical modelling, as well as Bayesian inference techniques. Additionally, I am looking forward to enhancing my interdisciplinary skills by integrating insights from climate-driven modelling approaches with mechanistic transmission dynamic modelling methods to improve forecasting of infectious diseases.