Characterizing US spatial connectivity: implications for geographical disease dynamics and metapopulation modeling

BackgroundHuman mobility is expected to be a critical factor in the geographic diffusion of infectious diseases, and this assumption led to the implementation of social distancing policies during the early fight against the COVID-19 emergency in the United States. Yet, because of substantial data gaps in the past, what still eludes our understanding are the following questions: 1) How does mobility contribute to the spread of infection within the United States at local, regional, and national scales? 2) How do seasonality and shifts in behavior affect mobility over time? 3) At what geographic level is mobility homogeneous across the United States? Addressing these questions is critical to developing accurate transmission models, predicting the spatial propagation of disease across scales, and understanding the optimal geographical and temporal scale for the implementation of control policies. MethodsWe address this problem using high-resolution human mobility data measured via mobile app usage. We compute the daily connectivity network between US counties to understand the spatial clustering and temporal stability of mobility patterns. We then integrate our mobility data into a spatially explicit transmission model to reproduce the national invasion of the first wave of SARS-CoV-2 in the US, and characterize the impact of the spatio-temporal scale of mobility data on disease predictions. FindingsTemporally, we observe that intercounty connectivity is annually stable, and was unperturbed by mobility restrictions during the early phase of the COVID-19 pandemic, despite significant changes in overall activity. Spatially, we identify 104 geographic clusters of US counties that are highly connected by mobility within the cluster and more sparsely connected to counties outside the cluster. Together, these results suggest that intercounty connectivity in the US is relatively static across time and is highly connected at the sub-state level. We find that the stability in temporal patterns allows static mobility data to effectively capture infection dynamics. On the other hand, spatial uniformity at the sub-state (cluster)-scale does not capture spatial dynamics; instead, mobility data at the county-scale is necessary to better predict spatial disease diffusion. InterpretationOur work demonstrates that intercounty mobility was negligibly affected out-side the lockdown period of Spring 2020, explaining the broad spatial distribution of COVID-19 outbreaks in the US during the early phase of the pandemic. Such geographically dispersed outbreaks place a significant strain on national public health resources and necessitate complex metapopulation modeling approaches for predicting disease dynamics and control design. We thus inform the design of such metapopulation models to balance high disease predictability with low data requirements.