ABSTRACT
Traditional contact tracing for COVID-19 tests the direct contacts of those who test positive even if the contacts do not show any symptom. But, by the time an infected individual is tested, the infection starting from the person may have infected a chain of individuals. Hence, why should the testing stop at direct contacts, and not test secondary, tertiary contacts or even contacts further down? One deterrent in testing long chains of individuals right away may be that it substantially increases the testing load, or does it? We investigate the costs and benefits of such multi-hop contact tracing for different number of hops. Considering a large number of contact topologies, spanning synthetic networks of divergent characteristics and those constructed from recorded interactions, we show that the cost-benefit tradeoff can be characterized in terms of a single measurable attribute, the initial epidemic growth rate. Once this growth rate crosses a threshold, multi-hop contact tracing substantially reduces the outbreak size compared to traditional contact tracing. Multi-hop even incurs a lower cost compared to the traditional contact tracing for a large range of values of the growth rate. The cost-benefit tradeoffs and the choice of the number of hops can be classified into three phases, with sharp transitions between them, depending on the value of the growth rate. The need for choosing a larger number of hops becomes greater as the growth rate increases or the environment becomes less conducive toward containing the disease. Author summaryThe COVID-19 pandemic has wrecked havoc on lives and livelihoods worldwide. Other epidemics may well emerge in future and one needs a preparedness to prevent their growth into another pandemic. During the early stages of a new epidemic, or even a mutated version of an earlier epidemic, pharmaceutical interventions may not be available but contact tracing and timely quarantine are among the few available control measures. We show that 1) traditional contact tracing may not successfully contain the outbreak depending on the rate of growth of the epidemic, but 2) the cost-benefit tradeoffs may be substantially enhanced through the deployment of a natural multi-hop generalization which tests contact chains starting from those who test positive.
ABSTRACT
Traditional contact tracing for COVID-19 tests the direct contacts of those who test positive even if the contacts do not show any symptom. But, why should the testing stop at direct contacts, and not test secondary, tertiary contacts or even contacts further down? The question arises because by the time an infected individual is tested the infection starting from him may have infected a chain of individuals. One deterrent in testing long chains of individuals right away may be that it substantially increases the testing load, or does it? We investigate the costs and benefits of testing the contact chain of an individual who tests positive. For this investigation, we utilize multiple human contact networks, spanning two real-world data sets of spatio-temporal records of human presence over certain periods of time, as also networks of a classical synthetic variety. Over the diverse set of contact patterns, we discover that testing the contact chain can both substantially reduce over time both the cumulative infection count and the testing load. We consider elements of human behavior that enhance the spread of the disease and lower the efficacy of testing strategies, and show that testing the contact chain enhances the resilience to adverse impacts of these elements. We also discover a phenomenon of diminishing return beyond a threshold value on the depth of the chain to be tested in one go, the threshold then provides the most desirable tradeoff between benefit in terms of reducing the cumulative infection count, enhancing resilience to adverse impacts of human behavior, and cost in terms of increasing the testing load.
