Mobility network reveals the impact of spatial vaccination heterogeneity on COVID-19

Abstract

Massive vaccination is one of the most effective epidemic control measures. Because one’s vaccination decision is shaped by social processes (e.g., socioeconomic sorting and social contagion), the pattern of vaccine uptake tends to show strong social and geographical heterogeneity, such as urban-rural divide and clustering. Yet, little is known to what extent and how the vaccination heterogeneity affects the course of outbreaks. Here, leveraging the unprecedented availability of data and computational models produced during the COVID-19 pandemic, we investigate two network effects—the “hub effect” (hubs in the mobility network usually have higher vaccination rates) and the “homophily effect” (neighboring places tend to have similar vaccination rates). Applying Bayesian deep learning and fine-grained simulations for the U.S., we show that stronger homophily leads to more infections while a stronger hub effect results in fewer cases. Our simulation estimates that these effects have a combined net negative impact on the outcome, increasing the total cases by approximately 10% in the U.S. Inspired by these results, we propose a vaccination campaign strategy that targets a small number of regions to further improve the vaccination rate, which can reduce the number of cases by 20% by only vaccinating an additional 1% of the population according to our simulations. Our results suggest that we must examine the interplay between vaccination patterns and mobility networks beyond the overall vaccination rate, and that the government may need to shift policy focus from overall vaccination rates to geographical vaccination heterogeneity.

Generalizing Bayesian Optimization with Decision-theoretic Entropies

“…In Figure 2 (top row), we compare performance by plotting the negative loss versus iteration on two higher dimensional examples. We also compare each method on the Vaccination function (provided by [53]), which returns the vaccination rate for locations in the continential United States, given an input (latitude, longitude). The goal of this task is to efficiently find a set of five diverse locations with highest vaccination rates.…”

“…EHIG t (x; , A) = EI t (x). [53], which uses county-level vaccination data provided by the CDC, and uses small area estimation 2 to interpolate the vaccination rate of every location. We restrict the optimization domain to be a rectangle focusing on the state of Pennsylvania.…”

The Effect of Vaccine Mandates on Disease Spread: Evidence from College COVID-19 Mandates

“…There are strong geographic and political components to vaccination rates, public health measures, and compliance with these measures (Murthy et al 2021;Yuan et al 2021;Gollwitzer et al 2020;Ye 2021;Clinton et al 2021;Adolph et al 2021;Fraser, Juliano, and Nichols 2021). Hence, we include not only county and calendar week fixed effects in our models but also region-byweek fixed effects, interactions between week fixed effects and the Democratic vote share in the 2016 presidential election, and interactions between week fixed effects and the baseline county vaccination rate (measured during the last week of June 2021, before any college in our data began fall classes).…”

Revealing geographic transmission pattern of COVID-19 using neighborhood-level simulation with human mobility data and SEIR model: A Case Study of South Carolina

“…In general, there is a dearth of studies exploring the relationships between human mobility and COVID-19 transmission at the neighborhood level, especially in less populous metropolitan statistical areas (MSAs). Their human mobility patterns may be different compared with the large ones such as New York MSA, which have been intensively studied Yuan et al 2022).…”