Road networks to explore COVID-19 infection

Shahadat Uddin, Arif Khan, Haohui Lu, Fangyu Zhou, Shakir Karim, Farshid Hajati, Mohammad Ali MonimedRxiv 2023 · DOI

Abstract

The COVID-19 pandemic necessitated unprecedented restrictive measures globally. While lockdowns and social distancing slowed case growth, concerns persist regarding the disproportionate impact on socio-economically disadvantaged populations. This study utilizes quantitative analysis and network theory to explore the relationship between local road networks, mobility, and infection rates during the Delta and Omicron outbreaks in Greater Sydney, Australia. By modeling suburbs as nodes and road connections as weighted edges, the research examines how physical infrastructure predicts viral spread and whether socio-demographic variables like age, education, and income moderate these effects. The findings indicate that road connectivity is a strong predictor of infection growth, particularly under lockdown conditions, and debunk the notion that socio-economically disadvantaged suburbs inherently drive higher community transmission.

Introduction

Public health officials traditionally rely on high-resolution GPS or cellular data to predict viral spread, but such data is often inaccessible due to privacy concerns. This research proposes an alternative: using the static "skeleton" of a city—its road network—as a proxy for human mobility. The study specifically investigates whether the physical connectivity between suburbs can predict how a virus moves and if socio-economic factors influence this relationship.

Methodology: The Road Network as a Proxy for Mobility

The researchers developed a network-based approach focusing on 100 suburbs in Greater Sydney. Suburbs were modeled as nodes, with edges representing physical road connections. The "weight" of each edge was determined by the number of physical roads connecting two suburbs, sourced from geospatial data.

Figure 1: Suburb connectivity modeled as a weighted network graph based on physical road connections.

A core innovation of this study is the Neighbourhood Measure ($RNInf$), which calculates a suburb's risk based on a weighted sum of its neighbors' infection counts. To analyze the spread, the authors utilized fixed-effect panel regression across two distinct four-week periods: the Delta wave (August–September 2021) and the Omicron wave (November–December 2021).

Comparative Analysis: Delta vs. Omicron

The study integrated infection data from NSW Health with socio-demographic data from the Australian Bureau of Statistics. The model's performance varied significantly between the two variants, reflecting the impact of public health mandates on physical movement.

Figure 2: Statistical summary of the Delta variant outbreak in Greater Sydney, showing the data used to validate the predictive model during lockdown conditions.

Infrastructure as Destiny

During the Delta wave, when a 5km travel limit was in place, the physical road network became a primary determinant of viral "leakage" between neighborhoods. The model explained over 85% of the variance in new cases for Delta. In contrast, for the Omicron wave, the model's predictive power dropped ($R^2 = 0.52$). This divergence suggests that when movement restrictions are lifted and a variant is highly transmissible, individual behavior and biological factors outweigh the constraints of physical infrastructure.

Socio-Economic Factors and Transmission

A critical component of the research was moderation analysis to determine if socio-economic variables (median age, education levels, and weekly income) influenced the link between connectivity and infection.

  • Debunking Blame: Income and education levels did not significantly moderate the relationship. While disadvantaged suburbs might show higher raw case numbers due to essential work or larger households, they were not inherently "more infectious" to neighboring areas.
  • The Role of Age: Age was the only variable that significantly moderated the Delta results, with higher median ages correlating with higher case growth, likely due to increased clinical vulnerability.

Conclusion

The physical road network serves as a remarkably strong proxy for mobility during periods of restricted movement. The study demonstrates that infrastructure, rather than socio-economic status, is a primary driver of neighborhood-to-neighborhood transmission. These findings are vital for stakeholders in tailoring health decisions and avoiding the stigmatization of disadvantaged communities during future pandemics.

Summary updated

Abstract

Abstract COVID-19 pandemic triggered an unprecedented level of restrictive measures globally. Most countries resorted to lockdowns at some point to buy the much-needed time for flattening the curve and scaling up vaccination and treatment capacity. Although lockdowns, social distancing and business closures generally slowed down the case growth, there is a growing concern about the social, economic and psychological impact of these restrictions, especially on the disadvantaged and poorer part of society. While we are all in this together, these segments are often taking the heavier toll of the pandemic and facing harsher restrictions or getting blamed for community transmission. This study tries to explore this perspective using quantitative analysis and network theory. The research is set in the context of the latest delta and omicron outbreaks in the Greater Sydney area, Australia, during late 2021. We first try to model how the local road networks between the neighbouring suburbs (i.e., neighbourhood measure) and current infection count affect the case growth and how they differ between delta and omicron variants. We use a geographic information system, population and infection data to measure - road connections, mobility and transmission probability across the suburbs. We then looked at three socio-demographic variables – age, education and income and explored how they moderate independent and dependent variables (infection rates and neighbourhood measures). The result shows strong model performance to predict infection rate based on neighbourhood road connection. However, apart from age in the delta variant’s context, the other two variables – income and education level do not seem to moderate the relation between infection rate and neighbourhood measure. The results indicate that suburbs with a more socio- economically disadvantaged population do not necessarily contribute to more community transmission. The study findings could be potentially helpful for stakeholders in tailoring any health decision for future pandemics.
BibTeX
@misc{uddin2023road,
  author = {Shahadat Uddin and Arif Khan and Haohui Lu and Fangyu Zhou and Shakir Karim and Farshid Hajati and Mohammad Ali Moni},
  title = {{Road networks to explore COVID-19 infection}},
  year = {2023},
  doi = {10.1101/2023.01.31.23285228},
  howpublished = {medRxiv},
}
Figure