Social distancing can work in Australia, modelling suggests
It's a matter of everyone getting involved.
By Nick Carne
Social distancing can help curb COVID-19 in Australia in three months, modelling carried out at the University of Sydney suggests.
However, this will only happen if 90% of the population complies with requirements and restrictions. If less than 70% does so, it will be an unproductive measure, according to complex systems expert Mikhail Prokopenko and colleagues.
And this assumes continuation of two other interventions – restrictions on international arrivals and case isolation (with increasing testing and monitoring resources).
“There is a clear trade off,” Prokopenko says. “Stricter measures imposed earlier would reduce how long our lives are impacted by this disease. On the contrary, laxer protocols could mean a longer, more drawn out and ineffective struggle against COVID-19.”
For every day the stricter social distancing measures are delayed, Prokopenko adds, society would need to endure several more days under a longer suppression policy.
“The longer we delay the peak, the more time our healthcare system has to prepare for it by accessing more resources such as ICU beds, ventilators, antivirals and trained health workers”.
The study also found that while school closures have the potential to compensate for 10% of a lack of social distancing compliance, they would only delay the peak of the pandemic by two weeks.
School closures did not significantly reduce new cases for older adults, but slightly increased the fraction of new cases in children around the peak of the pandemic in Australia.
Prokopenko’s team used a simulator comprising more than 24 million software agents, each with attributes of an anonymous individual, such as age, gender, occupation, susceptibility and immunity to diseases.
Contact rates within different social contexts – such as households, household clusters, local neighbourhoods, schools, classrooms and workplaces – are built into the program. The set of generated agents captures average characteristics of the real population and is calibrated to 2016 Australian Census data with respect to key demographic statistics.
“This model is calibrated to reproduce several characteristics of COVID-19 transmission, accounting for its reproductive number, the length of incubation and generation periods, age-dependent attack rates, and the growth rate of cumulative incidence during a sustained and unmitigated local transmission,” the researchers write in a paper available on the pre-print server arXiv.
“An important calibration outcome is the age-dependent fraction of symptomatic cases, with this fraction for children found to be one-fifth of such fraction for adults. We then apply the model to compare several intervention strategies, including restrictions on international air travel, case isolation, social distancing with varying levels of compliance, and school closures.
“The interactions result in transmission of the disease from infectious to susceptible individuals. Given the contact and transmission rates, the simulation computes and updates agents’ states over time, starting from initial infections, seeded in international airports around Australia.”
The researchers say the next phase of work will address limitations of the study, including a more fine-grained implementation of natural history of the disease, incorporation of more recent ABS data, inclusion of home quarantine strategy for the affected households, and a more refined school closures strategy.