Hein Duijf
The SARS-CoV-2 pandemic has forced almost a third of the world’s population into some form of quarantine, causing severe rights-restrictions, as well as drastic health, economic, social, and psychological harms. In many countries, when the magnitude of the COVID-19 threat was acknowledged, the virus was too widely spread for successful containment measures. The spread in these societies had gone too far, and the only viable alternatives were social distancing strategies (like some degree of quarantine and restricted movement) to prevent an overload of the health-care system. Social distancing strategies cause severe economic, social, and psychological harms, and they are prone to political abuse. Therefore, societies are seeking ways to return to normality.
Digital contact tracing technologies are heralded as an effective way of containing SARS-CoV-2 faster than it is spreading, thereby allowing the possibility of easing measures of population-wide quarantine. In a recent manuscript, we argue that digital contact tracing technologies need to address three important problems. First, there is much evidence that contract tracing policies are an effective containment strategy at the stage where there are very few infections in a society, but is it a viable strategy to return to normality after a period of population-wide quarantine? Second, despite the ethical and pragmatic concerns regarding digital contact tracing, the current public debate overlooks the crucial question of how much benefit these technologies would yield: How many people are expected to be quarantined under policies involving digital contact tracing? Third, how can these technologies ensure that the costs of the pandemic and its countermeasures are fairly allocated in society?
Behavioral factors may jeopardize the effectiveness of digital contact tracing
Given the severe risks associated with social distancing strategies, societies must find ways to safely decrease the number of people in quarantine speedily, restore infringed rights, and avoid health, economic, social, and psychological harms. Many countries are considering or implementing policies involving contact tracing and digital contact tracing technologies are a core part in these solutions. Under the assumption that the success of social distancing measures eventually allows returning to a containment approach, the first problem is as follows:
1. ‘Saving Lives’ Problem: How can the spread of the virus be contained, at least to the extent that overloading intensive care units is avoided?
There is much evidence that contract tracing policies are an effective containment strategy at the stage where there are very few infections in a society. However, the important question facing today’s societies is whether it is a viable strategy to return to normality after a period of population-wide quarantine. One of the most important epidemiological features of COVID-19 in this regard is that people infected with the virus are infectious up to seven days before the onset of symptoms. A great proportion (46%) of SARS-CoV-2 infections occur before the onset of symptoms, which means that infected individuals will likely spread the virus unknowingly. These pre-symptomatic infections put constraints on the feasibility of contact tracing approaches – both manual and digital.
One could consider the policy to quarantine only those individuals that have developed symptoms. In their Science publication, Ferretti et al. (2020) provide models that suggest this policy will not be effective. This means that any effective policy needs to quarantine individuals before the onset of symptoms. Contact tracing is one way to identify individuals that have a higher chance of being infected and to put them in quarantine (at least for several days or until tested negatively).
An important factor in the effectiveness of contact tracing is the time that it takes for a symptomatic individual to be isolated and for its contacts to be identified and quarantined. The importance of any delay is highlighted by further results of Ferretti et al. (2020): they show that every extra day it takes to isolate symptomatic individuals and quarantine their contacts would drastically reduce the efficacy of contact tracing; moreover, if the delay is two days or more, then their models reveal that it is uncertain whether contact tracing has any positive effect at all. This demonstrates that it is important to keep the delay to a minimum and essential to keep it below two days. Digital contact tracing technologies can help to decrease this delay, by reducing the time needed to identify and inform possible contacts.
Although digital contact tracing technologies may help to decrease the delay, there are good reasons to assume that several behavioral factors may jeopardize the effectiveness of digital contact tracing. First, studies show that a large uptake of the app (commonly 60% of the population) is necessary for digital tracing to be effective. Second, patient delay (a common term amongst medical practitioners which refers to the delay between the onset of symptoms and contacting a medical doctor) of only a day or two can be devastating. Third, some countries are considering policies where contacts are traced and quarantined only after the symptomatic individual has tested positively, thereby losing precious time between the onset of symptoms and the positive test. Finally, non-symptomatic contacts may delay going into quarantine and thereby risk pre-symptomatic transmissions.
Uncertain benefits of digital contact tracing
It is important to flag that these technologies are subject to several ethical concerns, including privacy concerns, the risk or mission creep (continued use of the app for other, perhaps nefarious goals), abuse by ill-minded users, reliability, and others. Given these ethical concerns and given the uncertainty regarding the effectiveness of digital contact tracing, the second problem is as follows:
2. ‘Saving Livelihoods’ Problem: How can the ratio of free, unquarantined people in a society be maximised?
To be sure, adopting contract tracing policies will entail that fewer people are quarantined, but to balance the risks with the potential benefits we need to obtain good estimates of how many people will be quarantined in a society under various policies involving digital contact tracing. After all, the abovementioned ethical concerns weigh heavier if it turns out that large portions of society will be quarantined under these policies.
Although there is little research on the ratio of unquarantined people in a society under contact-tracing policies, Hinch et al. (2020), in an unpublished manuscript, provide models based on the UK that suggests that, under plausible assumptions, it is to be expected that 30-50% of the population will be in quarantine under these policies. Of course, the more effective policies where symptomatic persons, their contacts, and all their families are quarantined yield the highest turn-in of 50%. So, even though digital tracing apps reduce the number of people in quarantine, it is an open ethical question whether the benefit of lifting the quarantine for a subset of our society outweighs the risks associated with the app (and with abandoning a lock-down). Public debates often overlook this issue and more research is needed to estimate the exact benefits of digital contact tracing.
Ensuring a fair distribution of costs
The previous problem yields a further hitherto unacknowledged problem that puts constraints and restrictions on possible solutions. Societies need to ensure that the costs of fighting the pandemic are allocated fairly, for both pragmatic and ethical reasons. Pragmatically, unequal allocations of costs, particularly to already disadvantaged groups, may further drive anti-democratic tendencies like nationalism and radicalism. Ethically, measures should benefit all parts of society equally, where inequalities ought to be justified so that they are to the greatest benefit of the least advantaged members of society. Hence, a third problem transpires:
3. ‘Ensure Fairness’ Problem: How can we ensure that the costs of the pandemic and its countermeasures are fairly allocated in society?
Ensuring fairness is of particular relevance in a discussion of digital contact tracing because several other technological solutions, such as AI-based technologies, have been shown to potentially exacerbate existing inequalities. Ample empirical evidence from the social and economic sciences suggests that the structures of social networks in societies will play a critical role in determining behavioral interactions with any technological solution. Further, recent research finds that the fraction of work that can be done from home varies substantially across industries, with those associated with low incomes and precarious work arrangements least likely to be done from home. Now is a time when existing inequalities may be rectified. Digital tracing will not be a genuine solution to our problem if it ignores or even perpetuates these problems.
(This blog post heavily draws on joint work with Michael Klenk and Christian Engels, see the full manuscript for a more elaborate discussion and references.)
Comments