NOVID: Personal Pandemic Network Radar

New Test-and-Trace approach driven by self-interest: anonymously see disease from afar in your interaction network, act preemptively and prevent exposure.

Po-Shen Loh

The current pandemic demonstrated that behavior is a major factor in pandemic control. Before vaccines and therapeutics are available, contact tracing and quarantine are workhorses for control. However, they have an incentive misalignment problem: directly involved participants (infected individuals and their contacts) are asked to protect the rest of society from themselves. Their quarantine compliance does not directly help them avoid infection (they already were in contact with infected individuals, or positive).

Our solution focuses on the incentive alignment problem, providing a new type of intervention with the distinctive property that even selfish behavior contributes to community control. This addresses the “tragedy of the commons”. We analyze interaction data in a novel way for Category 2a: “Identify and incentivize preventive interventions.”

The scale of this problem is massive. We focus on populations where most households have a smartphone. While some cultures’ citizens align with government directives (often aided by technology) and cooperate with scalable contact tracing and enforced quarantine, most of the world is not that way. For example, in a paper by Smith et al. studying the UK’s Test, Trace, and Isolate system, only 11% of quarantined individuals declared proper adherence. Our solution would directly benefit billions of people.

Our solution serves all households with smartphone access, addressing their need to avoid getting infected, while still interacting with the outside world. Many households cannot isolate for the entire pandemic, and we provide them with more precise information that they can use to budget their caution. It is not sufficient simply to know how many positive cases are in a city, because when there are few cases and the city opens up, people will get infected until cases grow out of control, and the city shuts down again. In contrast, when there are few cases, our solution dynamically alerts the socially-nearby people, helping them avoid getting infected, and helping the economy remain open.

In countries with high smartphone penetration, this solution serves everyone from schoolchildren (parents’ apps can link to their children’s classrooms) to the workforce. To understand needs, our team’s core has a User Experience Design group, which drives all feature development. They are frequently on discovery calls with institutions ranging from schools to employers to cities to nursing homes, where we seek to understand their needs. We also spoke with hundreds of users in our early experimental deployments, and have also used our app itself to survey users.

Since we’ve invented a new approach for pandemic control which complements and works together with standard isolation and manual/digital contact tracing, the resulting public knowledge base from research on this approach could have significant ramifications. The conceptual ingredients are so simple that with this knowledge, future groups anywhere could adapt the concept on their own. Consequently, our work could affect all future pandemics, permanently adding a new non-pharmaceutical intervention that complements all other tools. Indeed, our innovation even could synergize with the Test-Trace-and-Isolate paradigm by nudging towards early testing, faster self-isolation, and better adherence to quarantine. We will document our findings through peer-reviewed publications.

We will also provide public goods relating to how to practically deploy. We learned through experimentation that one must educate decision makers and populations on the difference between our new approach and standard digital contact tracing apps. By creating mass awareness around this new approach, we will contribute a public good that makes it easier for anyone to deploy this type of approach. We also learned the impacts of different ways to communicate the direct value proposition of our innovation. We are collaborating with established researchers to contribute this knowledge in the form of peer-reviewed publications.

We primarily have anecdotes now, but our extensive informal observations indicate that people are much more interested in apps that deliver direct (self-interested) value to them. Multiple people using NOVID took extra caution after learning that COVID-19 struck within 2 relationships of them. We also informally observed that when people considered the hypothetical scenario of a new deadly hemorrhagic fever, most were emphatically interested in installing an app to warn them of disease coming from many relationships away, and would strongly reduce their interactions if disease got close. In contrast, there was much less enthusiasm for participating in interventions whose goal was to isolate you after your exposure to the disease.

This informal evidence indicates that our solution likely brings a new powerful tool for mass behavioral alignment. To achieve impact, we will continue partnering with other researchers, and pursuing multiple deployments, which we work closely with while continually refining the system and its surrounding communication. We will publish research documenting best practices, and engage media to drive widespread awareness of this new approach. Then, when the next pandemic strikes, the general public will demand the pandemic network radar for their own self-defense, thereby more efficiently containing the disaster.

We currently have the only app in the world which delivers this new pandemic network radar paradigm, and we are gaining collaboration interest from other researchers. We will use the next year to pursue as many high-quality pilot deployments as possible, positioning NOVID not only as an app to contain COVID-19, but also to facilitate stable reopening of society. (That is uniquely facilitated by our ability to provide precise alerts even when there are relatively few positive cases.) In each pilot, we will focus effort on ensuring success, encompassing not only technical support, but also communication support.

Since our app is already downloadable anywhere, we expect that as we find success in pilots, demand will snowball. We will bring large populations around the world into our common global network. Although this pandemic will hopefully be under control in many countries within the next year, many other countries unfortunately will not be fully vaccinated so soon. We will seek to use our wide reach to conduct epidemiological research on the real-time data coming from our global interaction network. So, over the next three years, we will help to observe for new variants emerging from regions which do still have COVID-19 cases.

In each community where we deploy, our app measures not only the number of raw downloads, but also the robustness of the interaction network. We specifically track the total number of users within interconnected clusters of at least 100 users, where each connection corresponds to two users spending a significant amount of time near each other. In measuring installation, we seek to optimize the percentage of people in the community who are counted by this metric. This would validate that our incentive-aligned solution is indeed compelling.

We also measure the raw number of signals of positivity voluntarily entered by app users, so that we can track what fraction of positive users actually end up reporting their status.

Finally, unlike all contact tracing apps based on the Apple-Google infrastructure, we are able to keep track of the frequency of interactions between arbitrary users. For example, once we have a fully controlled pilot, we seek to observe the strength of correlation between a user’s relationship distance to the nearest positive signal, and the user’s frequency of interactions.

Our primary barriers are financial and reputational. We are applying to the Trinity Challenge as it could help in both dimensions. That said, we do not place all bets on any single approach, and so we have several strategies running to address these.

Financially, this work was supported thus far through a combination of philanthropy and contributions from the existing profitable social enterprise run by NOVID’s founder. That proved sufficient to establish our innovation’s current world-leading position, but significant funds are required in order to provide close support for deployments, and in order to improve the codebase for larger-scale deployment. We are applying for funds from many sources, ranging from philanthropy to governments.

Reputationally, this work represents a new paradigm, and it has not yet been publicly recognized by trusted voices in the public health space as a potentially powerful new intervention. That makes it more difficult to gain enthusiastic buy-in from a regional administrator, which in turn makes it more difficult to achieve a successful deployment. We are actively building collaborations with researchers to raise awareness of this approach.

Carnegie Mellon University

Expii, Inc.

Our primary barriers are financial and reputational. We have been engaged in significant fundraising efforts, which, if successful, would cover approximately half of our expenses over the next 12 months. A large financial award from The Trinity Challenge would play a major role in establishing our financial security to drive forward with this innovation.

The reputational benefits from a major award would also be transformational. We have introduced a new and unfamiliar paradigm, which many people initially dismiss at first glance because it is an app, and many people have determined that apps do not work. Indeed, there are issues with the existing app paradigm, which is why we innovated a fundamentally new app paradigm. Recognition from a highly prestigious judged award like The Trinity Challenge would drive scientific and media attention towards this innovation, making it far easier to convince administrators and constituents in pilots to take the few minutes to understand how our solution delivers a value proposition to them even from a purely selfish behavioral perspective. That will vastly improve the success rate in our deployments, and help this new approach snowball into a research-backed behavioral intervention mechanism to combat future pandemics.

We would be thrilled to partner with the listed universities (HKU Med, Imperial College London, London School of Economics and Political Science, Nanyang Technological University, National University Singapore, Northeastern University, Tsinghua University, University of Cambridge, the University of Melbourne, Johns Hopkins University School of Public Health) because of the research angle that our work provides. We would be able to provide researchers with anonymized interaction network data, and would be thrilled to pool expertise. Our vision is to combine product-driven interventions which collect and process that network data, together with insights from public health researchers.

We would also be thrilled to partner with the Bill & Melinda Gates Foundation, as our work could provide a cost-effective way to control pandemics (it operates on existing smartphones, which will become increasingly common over time in the developing world).

From the industry side, we would be happy to partner with Facebook and Tencent because they have no competing interests in the sense of active involvement in the Apple-Google contact tracing app architecture, while they also operate on networks (social instead of NOVID’s physical). Cross promotion across social networks driving awareness of this new approach would be very valuable.