RAPID: Fitting Ebola multi-type branching process models to data

该项目的主要目标是为科学界提供一套定量建模工具，以帮助实时识别对非洲埃博拉病毒持续传播贡献最大的因素，并允许直接测试合理的干预策略。 这将通过开发埃博拉病毒传播的分支过程模型来实现，这些模型与可用的数据源相连接，并允许人类行为的变化，以及纳入旨在减轻传播的现实情景。 对关键传播率和相关不确定性指标的估计将有助于提高在一系列情景下的预测精度。虽然多类型分支过程模型是捕捉传播中关键异质性的理想方法，并且可以很容易地包括与真实的时间内出现的严重疾病威胁相关的动态景观，但只有当我们开发出将其与数据相结合的技术时，它们的潜力才能最大化。这是一种双向通信，其中现有数据允许估计关键过程速率和概率，此外，精确的参数化模型允许预测。对于西非正在爆发的埃博拉疫情，这一工作流程两翼的效用具有重要意义。更广泛地说，多类型分支过程模型对新出现的传染病威胁具有广泛的适用性。改进这种模型的拟合技术填补了关于灵活建模框架和可用数据源之间的连接的知识空白。通过识别传输中简单的关键异质性，建模方法超越了不容易扩展到其他系统的专用模型。用于分析埃博拉疾病相关数据的模型和相关拟合程序也将与未来新出现的疾病威胁相关。

The main goal of the project is to provide the scientific community with a set of quantitative modeling tools that will assist in the real-time identification of factors contributing most to the ongoing spread of Ebola in Africa, and that will allow straightforward testing of plausible intervention strategies. This will be done by developing branching process models of the spread of Ebola that connect to available data sources, and that allow for changes in human behavior as well as incorporate realistic scenarios aimed at mitigating transmission. Estimates of the key transmission rates and associated measures of uncertainty will allow for improved precision in forecasting under a range of scenarios.While multi-type branching process models are an ideal way to capture key heterogeneities in transmission, and can readily include the dynamic landscape associated with serious disease threats unfolding in real time, their potential is maximized only when we develop techniques to interface them with data. This is a two-way communication in which existing data allows estimation of critical process rates and probabilities, and in addition, the accurately parameterized models allow forecasting. For the ongoing Ebola outbreak in West Africa, the utility of both wings of this workflow have important implications. More broadly, multi-type branching process models have wide applicability to emerging infectious disease threats. Improving fitting techniques for such models fills a knowledge gap regarding the connection between a flexible modeling framework and available data sources. By identifying simple, key heterogeneities in transmission, the modeling approach transcends purpose-built models that are not easily extended to other systems. Models and associated fitting procedures for the analysis of data related to the Ebola disease will be relevant also for future emerging disease threats.