The socio-spatial determinants of COVID-19 diffusion: the impact of globalisation, settlement characteristics and population

Abstract

Background: COVID-19 is an emergent infectious disease that has spread geographically to become a global pandemic. While much research focuses on the epidemiological and virological aspects of COVID-19 transmission, there remains an important gap in knowledge regarding the drivers of geographical diffusion between places, in particular at the global scale. Here, we use quantile regression to model the roles of globalisation, human settlement and population characteristics as socio-spatial determinants of reported COVID-19 diffusion over a six-week period in March and April 2020. Our exploratory analysis is based on reported COVID-19 data published by Johns Hopkins University which, despite its limitations, serves as the best repository of reported COVID-19 cases across nations.

Results: The quantile regression model suggests that globalisation, settlement, and population characteristics related to high human mobility and interaction predict reported disease diffusion. Human development level (HDI) and total population predict COVID-19 diffusion in countries with a high number of total reported cases (per million) whereas larger household size, older populations, and globalisation tied to human interaction predict COVID-19 diffusion in countries with a low number of total reported cases (per million). Population density, and population characteristics such as total population, older populations, and household size are strong predictors in early weeks but have a muted impact over time on reported COVID-19 diffusion. In contrast, the impacts of interpersonal and trade globalisation are enhanced over time, indicating that human mobility may best explain sustained disease diffusion.

Conclusions: Model results confirm that globalisation, settlement and population characteristics, and variables tied to high human mobility lead to greater reported disease diffusion. These outcomes serve to inform suppression strategies, particularly as they are related to anticipated relocation diffusion from more- to less-developed countries and regions, and hierarchical diffusion from countries with higher population and density. It is likely that many of these processes are replicated at smaller geographical scales both within countries and within regions. Epidemiological strategies must therefore be tailored according to human mobility patterns, as well as countries' settlement and population characteristics. We suggest that limiting human mobility to the greatest extent practical will best restrain COVID-19 diffusion, which in the absence of widespread vaccination may be one of the best lines of epidemiological defense.

Keywords: COVID-19; Coronavirus; Globalisation; Quantile regression; Spatial diffusion; Urbanisation.

Fig. 1

Distributions of cumulative reported COVID-19 cases per million population (log transformed). Graphs show the 10th week (ending March 4th) until the 15th week (ending April 8th) of 2020. The red line indicates the mean and the black lines quantiles

Fig. 2

Choropleth map of reported cases of COVID 19 per million population for the 84 countries included in the analysis over weeks 10 to 15 (ending March 4th and April 8th 2020, respectively)

Fig. 3

Diffusion of reported COVID-19 cases per million population (log transformed) over weeks 10–15 (ending 4th March and April 8th 2020, respectively) across 84 countries

Fig. 4

Standardised coefficient value of reported COVID-19 cases at the 25th, 50th, 75th and 90th quantiles the 10th week (ending March 4th) until the 15th week of 2020 (ending April 8th)