Governments have put in place a range of interventions in a bid to slow the spread of the coronavirus pandemic, and UNSW Business School research investigates the most effective ways to contain the virus
Since the outbreak of the coronavirus pandemic, governments around the world have introduced a range of policies and non-pharmaceutical interventions (NPIs), to slow the spread of COVID-19.
Similar policies were established in a large number of countries all over the world. These policies include restrictions on international travel, closures of schools, workplaces, public transport, cancellations of public events, restrictions on internal movement, tracing infected persons contacts, extensive testing practices, and more.
Timelines of these policies vary across countries, but by mid-April 2020, due to the virus crisis around 70 per cent of countries have enacted one or more of these measures, according to Geni Dechter, a Senior Lecturer in the School of Economics at UNSW Business School.
Together with fellow researchers Gonzalo Castex and Miguel Lorca, they set out to study the effectiveness of these government response policies.
In their recently published paper COVID-19: Cross-country heterogeneity in effectiveness of non-pharmaceutical interventions, they found that the effectiveness of NPIs may vary with a range of other actions governments may implement simultaneously, including (but not limited) to enforcement actions, sick leave policies and benefits to those forced out of jobs. Effectiveness may also vary with the response of the local communities at the time of crisis, such as the degree of compliance and voluntary social distancing.
Other factors that may influence the actions of local governments and communities – and the resulting effectiveness of NPIs – include public resources, geographic characteristics, the state of the health system and population characteristics.
“Our findings show that one size NPI does not fit all communities,” said Dechter, who observed that there are a number of relevant considerations for designing different country policies.
“For example, lockdown policies could be more effective in communities where individuals have lower incentives to voluntary practice self-distancing, for example because of age distribution, better health, or perceived access to health system,” she said.
“On the other hand, lockdown policies might not appear as effective in places where individuals are voluntary reducing their outside of home activities and contacts.”
Lockdowns should be complemented with other policies, such as extensive and accessible testing, contact tracing and isolation of infected individuals, especially when the enforcement of lockdowns cannot be guaranteed.
The UNSW Business School researchers tested whether economic, geographical and public health factors as well as population characteristics may explain the differences in flattening the spread of COVID-19 when NPIs are put in place.
“Different NPIs may be more or less effective in different environments,” said Dechter.
“We show that NPIs that impose lockdowns or closures, such as school and workplace closures, are more effective in countries with lower population density, lower surface area, lower air pollution, higher health expenditure, and lower proportion of elderly population.”
On the other hand, Dechter said extensive testing is more effective in countries with lower GDP per capita, higher population density, larger surface area, higher air pollution, higher employment rate, lower health expenditure, and higher proportion of elderly.
These findings can be explained by incentive-driven behaviours and public resource constraints, and Dechter also observed that compliance with closures may demand more resources in places with higher population density and a larger geographical area – making them less effective.
Furthermore, communities with access to better health care, measured by the number of physicians and health expenditure as percentage of GDP may have lower incentive to voluntary reduce activities related to social gatherings.
“Therefore, lockdown measures in such communities could be more effective,” said Dechter.
Dechter observed that more research is needed to study the design of optimal government response policies to a pandemic crisis (where an optimal response implies flattening the spread of the virus, relieving the pressure on hospital systems, and saving lives, while minimising the negative economic and social impacts).
“Based on our results and results in other existing studies, the optimal response policy might need to be targeted,” said Dechter.
“This targeting should consider the distribution of risk, endogenous social response, the capacity of the local government to enforce compliance, and the costs of the policy.”
Dechter pointed out that the research does not examine the social and economic costs of different combinations of NPIs and how these costs vary with community characteristics.
“More research is needed to incorporate economic and epidemiological interactions and to study how changes in economic and social incentives impact the dynamics of a pandemic,” she said.
“We hope to study some of these questions in our future work.”