Digital epidemiology leverages data generated outside the public health system to better understand how disease is spreading and how it can be contained. These include cellphone geolocation data and information from wearables, video surveillance, social media posts, internet searches and news reports, as well as crowdsourcing apps that collect self-reported symptoms.

Although promising, these technologies raise a number of ethical issues, write Mello, a professor of medicine and law, and Wang, an associate professor of pediatrics and director of the Center for Policy, Outcomes, and Prevention.

“For instance, some have voiced concern that trust and participation in such approaches may be unevenly distributed across society; others have raised privacy concerns,” they write. “Yet counterbalancing such concerns is the argument that sometimes it is unethical not to use available data; some tradeoffs may be not only ethically justifiable, but ethically obligatory.”

They note one of the most controversial applications of digital epidemiology during the pandemic has been the Chinese government’s requirement that citizens in 200 cities install an Alipay app on their smartphones. After collecting personal health data, the app assigns each of them a COVID-19 risk code — red, yellow, or green — that determines how they are permitted to move around their communities.

“The coding algorithm reportedly incorporates information on time spent at risky locations and frequency of contact with other people,” they write. “Public dissatisfaction with the app arose from lack of transparency about the reasons people were classified into particular groups and mismatch with individuals’ own believes about their risk level.”

Tough Love

Yet governments with massive troves of their citizens’ personal data at their disposal have been able to quickly track down those at elevated risk of infection and help them get tested. The authors note that the Taiwanese government, for example, linked immigration and customs data on travelers to National Health Insurance data on hospital and clinic visits to identify individuals whose symptoms could be due to contracting the coronavirus.

New Zealand, Thailand and Taiwan use cellphone location data to monitor the movement of people subject to quarantine or isolation orders. The authors note China, Poland and Russia have gone even further, using facial-recognition software to monitor compliance with orders.

“Such measures, though intrusive, help reduce the need for labor-intensive, in-person monitoring,” they write. “Location data from cellphones and social media apps can also be used to monitor population-level adherence to social distancing orders.”

Digital contact tracing has garnered the most attention, however. Mello and Wang note that a real-time experiment is underway in Singapore, where the government in March requested that its citizens install a government-developed smartphone app called TraceTogether. It uses Bluetooth technology to exchange identifier numbers with the phones of other TraceTogether users within 6 feet, sharing data with the government only if the user becomes subject to contact tracing due to a COVID-19 diagnosis. As of late April, the authors note, similar apps have been rolled out in nearly 30 countries.

Israel and South Korea have gone even further than Singapore, using geolocation data without seeking consent and texting people who come into contact with COVID-19 cases that they must immediately quarantine.

Silicon Valley Rivals Team Up

The United States has the largest number of COVID-19 cases and deaths, far surpassing China, where the coronavirus first took flight. That brings the authors to consider a high-profile experiment in digital epidemiology: the Google-Apple partnership.

The companies are working to enable smartphone users to download updates to their operating systems that will make it possible to track the physical proximity between phones. If a user later tests positive for the coronavirus, they can report it through the app and any users who have been in contact with those patients will receive a notification.

Navigating the Ethical Thicket

Mello and Wang recommend two policy “lodestars” be adopted when considering the ethics of digital surveillance during future pandemics. 

1. The wisdom of adopting a digital surveillance measure should be evaluated by considering what and how they likely would be used instead of the technology, such as shelter-at-home and business closure orders. Though intrusive, technologies offer the prospect of expediting the end of more burdensome public health orders imposed on the entire public and avoiding them in future outbreaks.
2. The use of digital technology must be judged the least burdensome alternative that would accomplish the public health objective. When digital technologies are used, this principle can also help minimize privacy intrusions by identifying the minimum necessary data elements and duration and scope of use.

The researchers endorse the use of electronic monitoring to support individuals subject to isolation, quarantine, and shelter-at-home orders, such as virtual visits from public health workers to check symptoms and ask if people need help fulfilling basic needs like food. But the use of electronic monitoring to enforce these orders is more problematic, they write. 

Though more effective than relying on police to detect violations, for “the benefits of stringently enforcing mass shelter-at-home orders are not entirely clear, and the potential for strict enforcement — particularly through electronic eyes — to undermine trust in government and stoke resistance is troublesome.”

Despite the privacy concerns, Mello and Wang lean toward favoring digital contact tracing.

“Because the virus is transmissible through casual contact for at least a few days before onset of symptoms,” they write, “people are unlikely to be able to recall all those they may have exposed. Even if they could, the number of public health workers needed to perform contact tracing grossly exceeds the available supply. The most likely counterfactual is failure.”

Although about 20% of the U.S. population lacks smartphones, “using the technology can conserve scarce human resources for working with those who don’t.” The researchers endorse an “opt-out” approach, in which smartphone users who object to the technology can choose to uninstall it. And they urge that data about COVID-19 cases and their contacts, including geolocation data, be shared with public health officials.

To ensure that these technologies are deployed responsibly, Mello and Wang argue they must be implemented through transparent processes with public input. An oversight body and carefully crafted data-use agreements would help assure a trustworthy system and lay down new rules-of-the-road for future pandemics.

“There has been much talk of harnessing the power and ingenuity of the tech sector to fight disease outbreaks, but `harnessing’ implies carefully placed constraints and firm direction by a driver,” they conclude. “We have yet to craft that yoke.”