How Contact Tracing Apps Could Help Prevent COVID-19 Super-Spreader Events
Until an effective vaccine is available, contact tracing apps could help limit the spread of COVID-19. But first they need to be more useable, better regulated, and more private.
Few countries have done better than New Zealand, South Korea, or Taiwan in controlling the COVID-19 pandemic. Their transmission rates are among the lowest in the world. So the United States, with the novel coronavirus resurging in many regions, could profit from their example.
The keys to these countries’ early and continuing success are not new; they would have been recognized by any epidemiologist working during the past 100 years. First, the countries required individuals to reduce contact with other people when transmission of the virus started. Then they asked people to wash their hands frequently, and to wear face masks or other coverings in public or whenever they couldn’t maintain physical distance.
Finally, they used an extremely labor-intensive process known as contact tracing, where public health staff identified all the people who had recently spent significant time with an infected individual—say, on a bus, at a crowded restaurant, or during a religious service—and quickly and repeatedly tested them in order to break the chain of further transmission. This time around, though, all three countries made extensive use of mobile-phone-based applications to broaden the reach of their traditional contact tracing efforts.
The results speak for themselves. Throughout much of October, daily cases in New Zealand (population 4.9 million) and Taiwan (population 24 million) ranged between 0 and 5. South Korea (population 52 million) has recently seen an uptick in numbers—from around 60 per day to 150—largely as the result of a recent cluster centered in a Seoul megachurch that has reportedly defied the government’s contact tracing efforts. Early, intensive testing-and-surveillance efforts allowed the countries to limit the amount of time they needed to shut down their economies and to ask residents to disrupt their day-to-day lives.
Early, intensive testing-and-surveillance efforts allowed the countries to limit the amount of time they needed to shut down their economies and to ask residents to disrupt their day-to-day lives.
The United States, which still lacks a cohesive national strategy against COVID-19, has not fared nearly as well. Some 70,000 people are now testing positive for the novel coronavirus each day in the latest surge. True, with 328 million people, the United States is four times larger than the three other countries combined. But Taiwan and South Korea are 15 to nearly 20 times more densely populated than the United States, and their citizens are slightly older on average—factors that would normally work against them in this type of pandemic.
Unlike their counterparts in these and many other countries, US health officials have had a hard time finding most of the individuals who have been exposed to the virus before they have had a chance to pass the infection to others. This is a particularly important failing. Epidemiologists now believe that the largest, fastest growing clusters of COVID-19 are typically triggered by so-called super-spreading events—situations in which the virus is more easily transmitted than usual. Studies of previous coronavirus outbreaks, including the SARS-CoV epidemic of 2003-04 in Asia and the ongoing MERS-CoV infections in the Middle East, have highlighted the importance of super-spreading events in fueling the rapid spread of these infections beyond their points of origin. Preliminary results from Hong Kong, Italy, South Korea, and other regions suggest a similar pattern is occurring with SARS-CoV-2, the virus that causes COVID-19.
In addition, researchers have determined that a surprising number of super-spreading events occur in situations where people who are ill exhibit limited symptoms despite being infectious. They are, therefore, less likely to choose to isolate themselves because they don’t feel sick—at least at first.
US health officials have had a hard time finding most of the individuals who have been exposed to the virus before they have had a chance to pass the infection to others.
This is precisely where mobile applications for contact tracing—in combination with traditional approaches—could prove most helpful in slowing the spread of the virus. In theory, getting many Americans to use a smartphone app should be possible, given that 81% of America adults owned a smartphone as of February 2019, according to the Pew Research Center. If the software allowed public health officials to identify infected individuals before they could pass the virus to many other people, any new surge in cases could be limited before it could spread too far.
Still, without a national app for this task, private companies and other organizations in the United States have scrambled to offer their own versions of contact tracing apps. These smartphone-based apps typically use Global Positioning System (GPS) or Bluetooth-based technologies to notify users of their exposure status following presumed contact with someone who has tested positive for SARS-CoV-2.
However, even this potentially helpful emergence of apps is failing to make a difference—and probably won’t for the foreseeable future without significant changes in US policy and practice. Perhaps the greatest problem is that fewer than half of Americans, according to a recent survey, say they would even consider downloading such software to their phones, largely because of privacy concerns. Furthermore, given US transportation patterns and geography, no single app is likely to achieve the scale necessary to succeed. Without a coherent national policy, public health must be regulated at the state level, meaning contact tracing apps are unlikely to work across borders. Because so many different groups have jumped into the fray, the apps tend to not work with each other. Finally, mobile apps are no panacea for infectious disease surveillance; they simply will not help without rapid, broad-based testing—and traditional contact tracing—of the general population.
As members of the COVID-19 Dispersed Volunteer Research Network (a group of interdisciplinary investigators from around the world), we have thought a lot about how to make digital contact tracing apps more successful in the United States—or in other countries that are still struggling to control the pandemic. Our recommendations fall into three broad categories.
Usability
Mobile tracing apps must reflect the diversity of the people they are designed to help, without further widening already existing health inequalities. Not everyone, for example, wants or can afford a smartphone. But 96% of Americans, according to Pew, own at least a regular cell phone. Thus “low tech” methods must be developed that use voice surveys and text-message-based digital contact tracing so that any digital efforts do not necessarily have to depend on smartphone capabilities.
Mobile tracing apps must reflect the diversity of the people they are designed to help, without further widening already existing health inequalities.
More than 350 languages are spoken in US households, according to the US Census Bureau—with English, Spanish, and Chinese being the most common. Ensuring that digital contact tracing tools are offered in a variety of languages and accommodate varying levels of health literacy is paramount to providing equitable and representative access.
In addition, mobile contact tracing apps must be compatible between and across different operating systems. Apple and Google recently partnered to develop the Exposure Notifications System, which allows apps that support public health endeavors to appear on either Android or iOS devices. Further steps should be taken to ensure compatibility across technologies, including Bluetooth, GPS, and low tech options.
Regulation
Contact tracing tools will need to work the way people do, including across state borders, which means the US government will need to develop federal rules to regulate them. At present, most tools are regulated by individual states. As interstate travel begins to reopen, the lack of federal regulations will be especially problematic for states with highly porous borders (e.g., New Jersey and New York). Interstate data sharing or a central data repository could provide a more flexible structure for digital contact tracing, but jurisdictional concerns may prove challenging. A successful centralized system would likely require federal involvement to coordinate between state public health agencies—an effort that necessitates the swift formulation of a legal framework.
Contact tracing tools will need to work the way people do, including across state borders, which means the US government will need to develop federal rules to regulate them. At present, most tools are regulated by individual states.
Furthermore, claims made by digital contact tracing platforms should be regulated to make clear that such tools can identify only potential contacts. The technology is just not advanced enough to give 100% certainty. For instance, GPS-based mobility data can be notoriously inaccurate at pinpointing users’ movements, especially when attempting to differentiate between whether a user is in one building versus another. Although Bluetooth-based approaches can offer more precision, they, too, struggle with ascertaining whether a user is one meter or ten meters away from someone else. The former might be considered an epidemiological contact while the latter would not. Finally, voice survey and text-based approaches to digital contact tracing are both prone to human error because they rely on users entering location information manually.
Privacy and security
Not enough people will download or use contact tracing apps to make the effort worthwhile if the software does not adequately protect their private information. The risk that their health or location data could fall into the wrong hands is just too great. And yet the problem of being able to identify individuals in anonymized data is a major concern with both GPS- and Bluetooth-based apps.
Not enough people will download or use contact tracing apps to make the effort worthwhile if the software does not adequately protect their private information.
Generally, GPS-based data directly involve users’ location information and are thus likely to be more easily identifiable than Bluetooth data, which describe whether two users were in near proximity of each other irrespective of location. That said, identifiability of Bluetooth data via other means—such as IP addresses—remains an issue. Furthermore, Bluetooth technology is uniquely vulnerable to security breaches, so much so that people who own smartphones are typically advised to turn off that capability when not in use. However, with digital contact tracing tools that depend on this technology, users would be expected to keep Bluetooth running in the background whenever they’re outside their homes, thus making them perpetually susceptible to cyberattacks.
One alternative to mitigate some of these privacy and security concerns—implemented successfully in New Zealand, South Korea, and Taiwan—uses check-in points equipped with Quick Response (QR) codes situated in high-traffic areas such as restaurants or other indoor spaces. Users would be encouraged to check in whenever they came across a QR code but would not have to reveal everywhere else they have been. In addition, software that more securely anonymizes phone numbers can be developed for digital contact tracing tools (including low tech ones)—potentially assuaging privacy concerns as they relate to sharing sensitive health data.
Developing digital contact tracing tools that address these considerations is a formidable challenge. But it will be essential if the United States is to limit the spread of COVID-19 before a safe and effective vaccine is available and adopted widely by the public. And even if a vaccine becomes available, it will take time to ramp up production and to distribute it widely enough to contain the pandemic—time that could also be used to adopt measures that limit further spread. Developing the right technology now could also save lives in future pandemics.
Beyond these technical challenges, however, lies the legal thicket into which most of these tools now fall. Many of the usability, regulatory, privacy, and security concerns we’ve raised present particularly potent legal challenges in the United States. The COVID-19 Dispersed Volunteer Research Network undertook a law review that examines these and other issues in greater detail, and it was recently accepted for publication. So far, we have found the potential obstacles to be daunting but not intractable. What is lacking is the political will and the necessary coordination to tackle them.