Human mobility models to forecast disease dynamics and the effectiveness of public health interventions

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Human mobility underlies infectious disease transmission and determines the spatial-temporal dynamics of outbreaks and endemic disease dynamics. Yet, we do not understand how best to incorporate individual or population mobility patterns into models of infectious diseases. Human travel has been successfully incorporated into models used for planning, surveillance, and reactive responses to influenza pandemics, the COVID-19 pandemic, malaria, and others. However, little validation or comparison of approaches used in these models has been performed. Further, there has been no systematic investigation of the extent to which the many different existing sources of human travel data quantify travel patterns, or which descriptions of human mobility are most relevant to disease processes. The small amount of human mobility data available globally requires generalization or extrapolation of features of one dataset to another setting, time or circumstance. This generalization may work for some features of pathogens for a subset of pathogens or transmission routes but may fail miserably in others. It is unlikely that all travel patterns are relevant for all types of diseases. The life history of each pathogen, transmission routes, age structure of incidence and outbreak context will all dictate the importance of specific types of movement. For mobility data to be useful in planning for outbreaks and monitoring interventions, transmission models utilizing mobility data and models must be confronted with epidemiological data (including contact tracing, traditional surveillance, and genetic data) from a variety of sources. Here, we propose to perform the first systematic analysis of existing mobility data and models to identify which models perform best under multiple assumptions using a range of simulations and data from historic outbreaks. We will also identify circumstances when generalized models or non-local data are misleading. To do this, we will collate and standardize a large number of mobility datasets collected by various methods. We will statistically characterize these datasets to identify sources of variation in human mobility at individual, household, community, and larger scales. We will develop multiple candidate models describing mobility and incorporate these candidate models into a range of commonly used models of infectious disease transmission. Proceeding with the principle that human mobility is only useful to models of infectious diseases if it improves our ability to recapitulate the dynamics of observed outbreaks, we will test the ability of each of these candidate mobility models to explain observed patterns of contacts and sequenced pathogens observed in outbreaks of dengue, Zika, Ebola, and COVID-19. In doing this, we will identify conditions under which human mobility can improve our understanding of the transmission and pathogens, inform response strategies and create a resource that can inform responses to multiple current and future outbreaks.