The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), has caused a significant amount of illness, mortality, and hospitalization globally. This ongoing pandemic has severely impacted the healthcare sectors, along with the global economy, which saw the greatest recession since the 1930s.

Study: Road networks to explore COVID-19 infection. Image Credit: Michael Smith ITWP/Shutterstock
Study: Road networks to explore COVID-19 infection. Image Credit: Michael Smith ITWP/Shutterstock


In response to the pandemic, scientists have worked tirelessly to understand every little aspect of the novel coronavirus, SARS-CoV-2. They developed vaccines and therapeutics and gathered more knowledge about the mode of transmission to manage the pandemic. Similarly, governments implemented several strategies, such as national lockdowns, mandatory wearing of facemasks, and travel restrictions, to contain COVID-19.

Several models have been developed based on SARS-CoV-2 transmission data to effectively predict the future trend. For instance, an autoregressive integrated moving average (ARIMA) model was developed using data from 145 countries. This model predicted the COVID-19 spread pattern based on the population data. The ARIMA model demonstrated that SARS-CoV-2 transmission could be projected using variables such as humidity, culture, and climate.

The classical Susceptible-Infected-Recovery (SIR) model was modified with fuzzy parameters, such as recovery rate, infection rate, and death rate due to SARS-CoV-2 infection. Although classical statistical models could not incorporate important determining factors, machine learning models provided an effective alternative to understanding complicated datasets. The artificial neural network (ANN)-based model was developed to predict the SARS-CoV-2 transmission pattern. 

As stated above, most governments implemented restrictions on people’s mobility to protect individuals from contracting COVID-19. Not many studies have verified the impact of mobility restrictions in terms of costs and benefits. In addition, it is important to determine if the implementation of other factors, such as increasing awareness, economic support, education, and so forth, could collectively contain the pandemic.

About the study

A new study posted to the medRxiv* preprint server investigated how human mobility influenced SARS-CoV-2 transmission based on a network-based approach and panel regression methods. The authors particularly considered the suburban population’s characteristics, such as age, education, and income, for the analysis.

The current study included SARS-CoV-2 infection data for one hundred different suburbs of the Greater Sydney area of New South Wales, Australia. Two distinct periods were selected, i.e., during Delta variant circulation and during Omicron variant circulation, to determine infection statistics of the suburbs. Three moderating attributes, namely, age, education, and income, were considered to determine their impact on the relationship between the dependent and independent variables of the current study’s model.


This study was divided into two parts. In the first part, the authors investigated the impact of an individual’s mobility through the neighborhood on the COVID-19 case count. The neighborhood measure was associated with the number of shared roads to determine human movement across the suburbs.

Notably, several underlying factors changed in the two selected timeframes of the study, i.e., during the Delta and Omicron circulation. During the Delta outbreak, the lockdown was implemented, and people could move within a 5 km radius for essential items. In addition, some areas also had night-time curfews during this period. The vaccination coverage increased from approximately 26% to 43% in the area. 

In contrast, during the Omicron phase, no lockdowns were implemented. The only mandatory requirements were social distancing, wearing face masks, and business capacity capping. Double vaccination coverage increased from 77% to almost 79%. Unsurprisingly, people’s mobility within and across the suburbs during this phase was significantly more than in the Delta phase.

The Omicron variant is more infectious than the Delta strain; hence, it is crucial to determine how the neighborhood measure impacted the COVID-19 cases linked with the Delta and Omicron outbreaks. Although the fixed effect panel regression model provided a good prediction performance for the Delta variant (R-squared value of 85.66%), its performance suffered for the Omicron variant (R-squared value of 52.67%). 

The previous infection count had a significant positive influence on the present infection count for both Delta and Omicron variants. This model projected that the infection counts for a suburb during the Delta variant could be positively determined based on past infection counts and influx from surrounding suburbs (neighborhood measure). 

In the case of the Omicron variant, the regression model and the neighborhood measure needed to provide more insights because the R-square value was almost the same as the delta variant. In addition, the neighborhood measure revealed a negative impact on infection counts counter-intuitively.

In the second part of the study, researchers investigated how age, education, and income impacted the infection rate in the suburbs. They found that education did not have any moderating effect on the infection rate for both variants. In contrast, age and income significantly influenced the relationship between previous and present COVID-19 case counts.


The model captured the macro-movement during the lockdown and utilized it to predict infection rates during Delta and Omicron outbreaks. The impact of mobility on the infection rate was determined based on the road network between the neighboring suburbs that assisted in the influx and the corresponding risk of an increase in the number of cases from adjacent places.

*Important notice

medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

Source link