As the U.S. contemplates routes to economic recovery in response to Covid-19, it is hurtling toward a fork in the road. A sustainable approach to contact tracing is becoming the central issue in economic recovery. The U.S. still has two choices: a light-weight, human-centric approach or a heavy-weight, machine-centric approach. The clock is ticking; the sand is draining from the hourglass as we avoid both options in futile hope of some third possibility.
The light-weight, human-centric approach is the hermetically-sealed perfection of the proposed Apple/Google app-based solution for non-intrusive contact tracing in a distributed architecture. The Apple/Google initiative, positioned as “exposure notification,” would enable us to restore social and economic activity without unleashing a biological catastrophe. We’ve become familiar with the need for testing to accompany a relaxation in social distancing; testing is futile without contact tracing. Without contact tracing, we have no way of knowing if and when individuals who test negative are subsequently exposed. We would literally need to test continuously. Likewise, contact tracing must be accompanied by testing in order to trigger notifications when individuals test positive.
The heavy-weight, machine-centric approach is the oppressive, almost draconian location-based solution for extremely intrusive contact tracing that we are seeing in Singapore and South Korea — as those regimes respond to the infection spikes that correspond to premature relaxation. We can say that such extreme measures would never fly in the U.S. — let’s see what policymakers and their constituencies say when we relax prematurely and, like other regimes, we hit a spike that takes us back to the future.
The options are clear:
- Relax while avoiding contact tracing, hit a spike, revert to extreme social distancing, and deploy a heavy-weight approach to contact tracing; Or
- Relax while deploying a light-weight approach to contact tracing and avoid a spike.
A poll conducted by The Washington Post and the University of Maryland indicates that only 40% of Americans are willing and able to implement the Apple/Google solution. We need 60% minimum — and even 60% is a bare minimum. While the Apple/Google initiative has gained traction in Germany and Italy, France and the U.K. are still in limbo and public-health agencies in the U.S. have generally disregarded it. Probably rooted in distrust of technology and loss of control, U.S. agencies typically envision a manual approach implemented by “armies” of contact tracers. Other perspectives suggest that employers, rather than public-health agencies, will be the organizational locus of contact tracing. While the various parties deliberate, the option of a light-weight approach recedes.
The evaluation of a manual approach seems to consider volume, not speed and accuracy. While manual approaches could probably manage volume over time, we don’t have time. By the time that a manual process traces the first tier of contacts, those contacts will have spread the infection to the next tier and beyond, as the spread grows exponentially. A manual approach is feasible at a small scale; it is not feasible at a large scale. The accuracy of the manual approach is also an issue — historical experience has shown that human recall, especially over a two-week period, is problematic. Infected individuals aren’t always cooperative with the investigation. As we attempt to scale the manual approach, infected individuals will find that the perceived intrusiveness of the automatic approach would be a relief in comparison.
An employer-centric approach would be location-based; employers can’t enforce adoption off-site. There is no way of tracking exposure that occurs off-site/off-hours. It’s a leaky system. In order to be effective, contact tracing needs to be comprehensive. Even if employers could enforce off-site use of a particular app, their systems wouldn’t integrate with the interactions that their employees are having off-site. Their apps won’t have interoperability with the apps of other employers. Employer-centric proliferation of apps may seem practical as a site-level response, but it fails as a system-level response.
While the Apple/Google solution solves the integration and scalability issues, users resist adoption. Their privacy concerns are justified. In a typical example of Silicon Valley myopia, Apple and Google produced a brilliant technical solution that ignores the behavioral and social context. The overall fabric of contact tracing is a socio-technical system — the technical system operates in the context of a social system. Every system operates in the context of a super-system. How will we get to 60% adoption — or even 80%? While the technical solution is reasonably convincing, even to a skeptic, the organizational system isn’t slightly convincing, even to an optimist. No trusted intermediary has visibility into the Apple/Google “machine room”. The Apple/Google initiative should not have a free hand — it requires third-party supervision. With third-party supervision by a formidable privacy advocate like American Civil Liberties Union (ACLU), a 60% adoption rate becomes conceivable. Skeptics argue that even Singapore has achieved adoption by only a third of its population; in the U.S., Utah launched an app-based approach and has achieved only a 1% adoption rate. With third-party supervision, we can reverse that momentum — we can turn our open and balanced social system to our advantage and out-perform even Singapore.
Like the vendors, the public-health agencies also must make concessions: they need to provide visibility into configuration and administration of the solution by a trusted intermediary; and they need to abandon their visions of location-based tracking and global visibility into consumer behavior.
ACLU has indicated its expectations for app-based contact tracing: a voluntary and distributed implementation. It is prepared to take legal action in order to enforce its expectations. The Apple/Google initiative has conveniently delivered a solution that potentially satisfies those expectations. Privacy advocates in Germany and the U.K. have endorsed the Apple/Google solution. ACLU and public-health agencies make strange bedfellows; ACLU and Apple/Google also make strange bedfellows. The Apple / Google initiative itself is surely an un-conventional alignment… until the age of Covid. Strange times call for strange combinations. All of the parties need to “give a little bit”. If consumers resist the Apple / Google option, the situation will devolve into an employer-centric and location-based approach that is fragmented and oppressive; agencies and employers will be able to legitimately enforce it without respect for privacy.
|Configuration||Speed / Capacity||Accuracy||Integration||Intrusiveness|
|Automatic / Location-Based||High||High||Low||High|
|Automatic / App-Based (Aggregated / GPS)||High||High||High||High|
|Automatic / App-Based
(Distributed / Bluetooth)
The path to freedom: Public-health agencies should adopt the Apple / Google initiative on the condition of third-party supervision (and their constituencies should encourage that direction); the agencies should relax social distancing on the condition of sufficient adoption (and their constituencies should embrace that contingency). We need an integrated, scalable, and non-intrusive approach to contact tracing. It’s an obstacle course.
This approach to the organizational context of contact tracing solves the tension between effectiveness and intrusiveness; it solves the tension between private rights and the public interest. Agencies, vendors, and consumers all need to “give a little bit”. Europe is trending in the human-centric direction while Asia is trending in the machine-centric direction. The U.S. has a choice; the U.S. needs to make a sustainable decision. That fork in the road still approaches; that sand is still draining from the hourglass.
Photo: LeoPatrizi, Getty Images