Suburban Road Networks to Explore COVID-19 Vulnerability and Severity

Shahadat Uddin, Arif Khan, Haohui Lu, Fangyu Zhou, Shakir KarimInternational Journal of Environmental Research and Public Health 2022 · DOI

TL;DR

This study reveals that the structure of suburban road networks significantly predicts COVID-19 vulnerability and severity. By analyzing road connections in Greater Sydney, Australia, researchers found that suburbs with higher road network centrality (more connections and easier accessibility) were more susceptible to early outbreaks and experienced higher case numbers, offering crucial insights for pandemic preparedness.

What Problem Does This Paper Address?

The COVID-19 pandemic, particularly with the emergence of highly transmissible variants like Delta, has highlighted the challenge of understanding and predicting disease spread. While global and national factors are often studied, the intricate dynamics of local transmission, especially within suburban areas, remain less understood. Human mobility, influenced by local infrastructure, plays a critical role, but its impact on the vulnerability (timing of the first case) and severity (number of cases) of outbreaks at a granular, suburban level has not been thoroughly explored. This paper aims to bridge this gap by investigating how the connectivity of suburban road networks influences these two key COVID-19 outcomes.

Methods

The researchers constructed a suburban road network for 19 local government areas (LGAs) in Greater Sydney, Australia, using postcodes as nodes and roads connecting them as edges. The weight of each edge was determined by the number of roads connecting two suburbs. To analyze this network, they employed several graph theory concepts:

  • Centrality Measures: Four standard measures were used to quantify the importance and influence of each suburb (node) within the network:
  • Degree Centrality: Measures the number of direct connections a suburb has to other suburbs.
  • Closeness Centrality: Assesses how easily a suburb can reach all other suburbs in the network.
  • Betweenness Centrality: Identifies suburbs that lie on the shortest paths between other pairs of suburbs, indicating their role as bridges.
  • Eigenvector Centrality: Measures a node's influence based on the influence of its neighbors.
  • Core-Periphery Structure: This analysis helps identify central "core" suburbs and more peripheral ones within the network.

These network measures were then used as independent variables in statistical models (Multiple Linear Regression and Random Forest Regression) to predict two dependent variables:

  • COVID-19 Vulnerability: Defined as the date difference between the first locally acquired Delta variant case in a suburb and a reference "date zero" (June 16, 2021). A lower score indicates higher vulnerability (earlier outbreak).
  • COVID-19 Severity: Defined as the number of locally acquired Delta variant cases per thousand population up to October 10, 2021. A higher score indicates higher severity.

Datasets

The study utilized the following datasets:

  • COVID-19 Data: Publicly available data from the NSW Government's COVID-19 data and statistics for the Delta variant outbreak in Greater Sydney.
  • Geographic Data: Google Maps was used to calculate road connectivities between suburbs.
  • Sociodemographic Data: Information from the Australian Bureau of Statistics (2016 Census QuickStats) for the 137 postal areas within the 19 LGAs was summarized but not directly used in the primary network analysis.

The study did not introduce a new benchmark dataset but rather applied network analysis to existing COVID-19 and geographic data.

Key Insights and Ideas

  • Road Network as a Proxy for Mobility: The study effectively demonstrates that the structure of suburban road networks can serve as a reliable proxy for human mobility patterns, which are crucial for disease transmission.
  • Connectivity Drives Vulnerability and Severity: The core idea is that suburbs with higher connectivity within the road network are more exposed to the virus. This is because they have more opportunities for people to travel in and out, increasing the chances of an early introduction of the virus and subsequent spread.
  • Degree Centrality is King: The most striking finding is the consistent and strong predictive power of degree centrality. Suburbs with more direct road connections are demonstrably more vulnerable and experience higher severity.
  • Lockdowns Might Not Be Enough: The findings suggest that simple radius-based lockdowns (e.g., 10km or 5km) might be less effective than anticipated if they don't account for the underlying road network connectivity that allows virus movement across these boundaries.

Key Results

  • Degree Centrality: Was a statistically significant and strong predictor for both COVID-19 vulnerability (negative correlation, meaning higher centrality leads to earlier cases) and severity (positive correlation, meaning higher centrality leads to more cases).
  • Closeness and Eigenvector Centrality: Also showed statistically significant predictive power for COVID-19 vulnerability and severity, respectively.
  • Correlation Matrix: Highlighted that degree centrality had the highest correlation coefficient with both vulnerability (r = -0.452, p < 0.01) and severity (r = 0.519, p < 0.01).
  • Regression Models:
    • Multiple Linear Regression (MLR) for vulnerability showed degree centrality as the most significant predictor (p=0.000).
  • MLR for severity also identified degree centrality as the most significant predictor (p=0.000), with eigenvector centrality also being significant.
  • Random Forest Regression (RFR) models achieved higher R² values (82.44% for vulnerability, 91.51% for severity) compared to MLR (23.30% and 35.80%, respectively), indicating a better fit and predictive power.
  • Feature Importance (RFR): Confirmed degree centrality as a top predictor for both vulnerability and severity.

Limitations and Future Work

  • Study Area Specificity: The study focused on Greater Sydney, Australia. While the methodology is generalizable, findings might vary in different urban or suburban contexts with different road network structures and population densities.
  • Data Granularity: While using postcodes as nodes is a good proxy, finer-grained data (e.g., individual street segments) could offer even more detailed insights.
  • Sociodemographic Factors: Although sociodemographic data was collected, the primary focus was on network analysis. Future work could integrate these factors more deeply with network measures to provide a more holistic understanding of vulnerability.
  • Future Directions: The authors suggest that this network-based approach can be extended to study other pandemics or different variants of COVID-19. Comparing vulnerability and severity across different pandemics (e.g., COVID-19 vs. H1N1) using this methodology is also proposed. The study also implies that suburb-wide lockdowns might need to be re-evaluated in light of network connectivity.

Summary updated

Abstract

The Delta variant of COVID-19 has been found to be extremely difficult to contain worldwide. The complex dynamics of human mobility and the variable intensity of local outbreaks make measuring the factors of COVID-19 transmission a challenge. The inter-suburb road connection details provide a reliable proxy of the moving options for people between suburbs for a given region. By using such data from Greater Sydney, Australia, this study explored the impact of suburban road networks on two COVID-19-related outcomes measures. The first measure is COVID-19 vulnerability, which gives a low score to a more vulnerable suburb. A suburb is more vulnerable if it has the first COVID-19 case earlier and vice versa. The second measure is COVID-19 severity, which is proportionate to the number of COVID-19-positive cases for a suburb. To analyze the suburban road network, we considered four centrality measures (degree, closeness, betweenness and eigenvector) and core-periphery structure. We found that the degree centrality measure of the suburban road network was a strong and statistically significant predictor for both COVID-19 vulnerability and severity. Closeness centrality and eigenvector centrality were also statistically significant predictors for COVID-19 vulnerability and severity, respectively. The findings of this study could provide practical insights to stakeholders and policymakers to develop timely strategies and policies to prevent and contain any highly infectious pandemics, including the Delta variant of COVID-19.
BibTeX
@article{uddin2022suburban,
  author = {Shahadat Uddin and Arif Khan and Haohui Lu and Fangyu Zhou and Shakir Karim},
  title = {{Suburban Road Networks to Explore COVID-19 Vulnerability and Severity}},
  journal = {International Journal of Environmental Research and Public Health},
  year = {2022},
  doi = {10.3390/ijerph19042039},
}
Figure