Gaps in speed, precision, and privacy diminish the power of contact tracing against COVID-19. A holistic approach to filling these gaps can help contact tracing live up to its potential, enabling communities to safely reopen.

“You’ve had contact with someone who has tested positive for COVID-19.”

The information in these words will save lives for the foreseeable future. It will help protect employees, colleagues, customers, children, family, partners, and friends.

• When shared with individuals, this information will allow them to self-manage. They’ll be able to adjust their behavior to mitigate risks to the health and safety of others.
• When shared with organizations, this information will allow them to act. Decision makers will be able to allocate resources and enact safety protocols that help contain outbreaks.

Collecting and sharing this information is a process called contact tracing. It’s increasingly seen as ​one of the biggest hopes for safely reopening workplaces and communities. Officially speaking, contact tracing is the process of identifying where, when, and how a person has contracted an illness and whom they may have exposed.

But it’s not without its shortfalls, especially against COVID-19. Experts across Slalom examined the gaps in contact tracing, then looked at how to fill them.

Here is what we found:

Gap #1: Speed

Manual contact tracing methods alone are too slow for COVID-19 — at least for the level at which the virus is present in many U.S. states and counties. With manual tracing, professional contact tracers identify, interview, counsel, and provide support for those impacted by COVID-19, usually over the phone. Manual tracing has successfully slowed the spread of COVID-19 in some parts of the world, but factors including U.S. conditions, workforce scalability, and the nature of the virus mean it’s simply not enough right now.

• Contact: When “contact” can mean any time spent less than 6 feet away from someone within the last two to three weeks, remembering when and with whom you have had contact becomes harder and more time-consuming. Add to that unfamiliar contacts outside the home and follow-up becomes exceptionally more difficult than it is with diseases with lower rates of communicability. The more that communities and workplaces reopen, the more contact there will be, and so it will continue to take more time to investigate individual cases. The more time between being exposed and being notified, the greater risk that the virus is spreading unchecked.
• Scalability: Manual tracing methods aren’t easy to scale. Germany has 400+ contact tracing call centers with far fewer cases to investigate than the U.S. In April, experts at Johns Hopkins estimated that an extra 100,000 contact tracers were needed across the U.S. to make manual contact tracing work. In June, only seven states and the District of Columbia were adequately staffed. Staffing up demands time (and money). Beyond that, a human workforce that is both growing quickly and dealing with sensitive information can lead to mistakes that leave that information exposed.

Filling the speed gap: To accelerate traditional forms of contact tracing, we see a lot of potential in combining manual and digital tracing methods.

With digital — or automated — tracing, people opt into sharing location and/or Bluetooth data, usually from their mobile devices. That data is used to estimate the proximity and duration of their exposure to COVID-positive patients, who must also opt into sharing their data. Also known as proximity tracing, this approach can alert users of possible exposure to COVID-19 with no manual intervention, as well as allow contact tracers to quickly narrow their scope of manual outreach, saving time and resources.

Can automated contact tracing solve all problems related to speed? Unfortunately not. For example, testing delays — which are common in the U.S. — can make any contact tracing effort less effective because the solution doesn’t start until a positive test result is returned. To speed up testing, we’re hopeful about an emerging technique that “pools” samples from asymptomatic individuals and tests them in batches. Machine learning could make it even more efficient.

Gap #2: Precision

The automated tracing solutions currently available, while fast, are not as smart as they need to be. The reason lies again in the nature of the virus but also in limitations around common automated methods.

• Exposure: Experts say that COVID-19 is airborne. When a COVID-positive person passes through an area, that would mean that the virus may linger in the air for a little while in the form of aerosol microdroplets. This is a type of indirect exposure. So is exposure via surfaces. Many automated tracing methods still don’t track indirect or transient exposure, leaving users more vulnerable to it.
• Altitude: Imagine an employee is working on the third floor of a building. Directly below them, on the first floor, is an employee who later tests positive for COVID-19. Many automated tracing methods still don’t measure altitude. The first employee will be told they’ve had close, prolonged contact with a COVID-positive person, a “false positive” that can drain limited resources and cause unnecessary fear.

Filling the precision gap:There is at least one way to track indirect exposure to airborne COVID-19. It’s a technique called ray tracing, and it’s commonly used in video games and computer graphics to determine how light illuminates a scene. Ray tracing and the principles behind it work with people if you treat them like light sources. When someone passes through an area, their trail is illuminated. The trail fades over time, as would droplets and/or particles in the air.

No single tracing technique can provide the full picture, so we recommend ray tracing as a supplement to technology that helps track more direct exposure, such as Bluetooth. We also recommend collecting altimeter data to keep track of users’ relative heights and adjust their exposure accordingly.

Gap #3: Privacy

The more people who use contact tracing apps, the more effective those apps are. Thankfully, not as many people as you’d think need to be using contact tracing apps for them to have an impact. Models from Oxford University indicate that even 14% app adoption can slow the epidemic to some extent. Still, few countries have that level of adoption, and privacy concerns — especially around data and transparency — are a major reason why.

• Data: The co-author of a book on the ethics of contact tracing and her colleagues found that experts agree on one thing: “For any large-scale digital contact tracing system to work, people need to trust that their data will be handled safely.” That trust is tenuous, as evidenced by digital contact tracing controversy in countries including India, Norway, and the UK.
• Transparency: Many providers of digital tracing solutions don’t invite third-party reviews of their security and privacy protections. We understand their fears but challenge them to play the long game. They’re giving up the opportunity to fix vulnerabilities and improve user trust.

Filling the privacy gap: How can you implement digital tracing solutions while protecting users’ privacy? For one, you can retain users’ data only as long as it’s useful for contact tracing. As soon as location and/or Bluetooth data ages beyond the window of possible transmission, purge it. Second, we propose a sliding scale of location accuracy to keep users from being identified. When there aren’t enough users in an area to protect an individual, automatically purge the data of those who are present. Users can still receive automatic exposure notifications in anonymity, and risk data can still be shared in aggregate with individuals and organizations.

For their part, technology providers can ensure that their contact tracing tools don’t use any personally identifiable information (PII) — and assure users of that. They should also invite third-party reviews of their security and privacy protections, as well as consider making their technology open source. But ultimately, we agree with this assessment from the Brookings Institution: “The more common uses and abuses of contact-tracing apps aren’t defined by the architecture of the technology, but by the political context into which they’re deployed.” To the best of their ability, governments and other organizations must work together to depoliticize contact tracing.

If you build it…

Governments, organizations, and technology providers may overcome speed, precision, and privacy gaps only to discover that no one actually wants to use their apps. In the end, contact tracing solutions must be useful not just to organizations and decision makers, but to individual users. That’s why our final recommendations are about making the user experience a good one.

• Make it visual. From risk heatmaps to local testing sites, pull in and visualize the data that allows users to self-manage. For decision makers, you can augment visualizations with dashboards displaying global views of office exposure, employee quarantine, and sick leave.
• Make it personal. Personalize and even gamify the experience with symptom self-assessments and tools to support handwashing, mask wearing, community service, and more.
• Make it clear. Provide evidence-backed tips and information that are validated by public health authorities.

Slalom focuses on the intersection of people, process, and technology. It’s never felt more necessary to consider that intersection than it does now, as we face down a pandemic and the future rides on the strength of our spirits, the coordination of our efforts, and the capabilities of our technology. If your organization could use some help navigating this complex new landscape — through contact tracing or something else — we are here. Together, we can reopen safely.

Tune in for our contact tracing webinar

On August 19, join us for a virtual discussion about the dos and don’ts of adoption of contact tracing. You’ll learn more about how Slalom is filling critical gaps in contact tracing, and how you can leverage this work at your organization. Register now.

Maarten Sundman is Principal Architect for Slalom’s Technology Enablement practice, Northern California.

Special thanks to:

Chris Collier, Data & Analytics Consultant, Slalom Austin

Slalom Technology is created by IT industry leaders and practitioners from Slalom, a modern consulting firm focused on strategy, technology, and business transformation.