Project Description

Author: Dr. Jayanta Sarkar, Queensland University of Technology.

Epidemiological predictions of COVID-19 have so far served as critical planning tools for policymakers and health practitioners. One key shortcoming of the epidemiological models is that they presume human behaviour as given, ignoring the crucial behaviour-disease interactions. Health-seeking behaviour is known to depend on the perceived prevalence of a disease (e.g., Blythe et al., 1991; Hethcote et al., 1991; Auld, 2003). The recent economic-epidemiology models capture health behaviour only through the modelling of people’s choice of physical interactions. However, the implications of other behavioural attributes that evolve over the course of an epidemic, such as risk-taking behaviour, are not well-understood. Incorporating such behavioural nuances into epidemiological modelling can make these models more useful in informing the necessary control measures.

In a non-mandatory COVID-19 testing regime, a critical individual risk-taking behaviour is the decision to get tested. From a rational perspective, it is worthwhile to get tested when one’s perceived benefits from testing outweighs the associated costs. The benefit depends on, among other things, the perception of own health-risk and that of the immediate family members and co-workers, if infected. The costs include the private cost of self-isolation, if tested positive, which depends on his/her valuation of the foregone income and outdoor activities due to self-isolation. Indeed, for someone experiencing COVID-19 symptoms, the evident health-risk from infection is too high, implying the perceived benefit from testing is likely to be higher than the cost, given everything else.

But what about those who experience mild or no symptoms? Emerging evidence shows that a surprisingly high proportion of infected people fall in this category. This group faces a higher private cost of testing, which could well exceed the expected benefit at the margin. Moreover, if risk-attitude is prevalence-dependent, perceived health-risk from infection would be low when the visible infection curve plateaus temporarily, triggering risk-taking behaviour. Thus, demand for testing will be much lower among the ‘asymptomatics’, allowing them to stay under the radar causing new unexplained community infections. The ‘invisible’ infections may well be the main reason behind the unusually high infection rate found in the Austrian ski-resort village Ischgl.

What does this imply for health policy? First, the government messaging that urges people to test ‘if they experience symptoms’ needs to change to reflect asymptomatic infections. Second, random testing may be the best option in detecting the asymptomatic infections even though such testing may be expensive. A recent research suggests that a stratified random testing focused on at-risk groups can dramatically reduce the required testing resources, which will probably be much lower than the cost to the economy and human lives resulting from future waves of infection.

This pandemic has clearly demonstrated, perhaps not for the last time, that individual behaviours driven by self-interest are often in conflict with the socially desired behaviour. This may also be reflected in the uptake of future vaccines. Thus, there is no better time like the present to bridge the gap between behavioural economics and epidemiology.

For more information please contact Dr. Jayanta Sarkar at Queensland University of Technology (email: Jayanta.sarkar@qut.edu.au)