It’s been over a year since teams around the world began launching exposure notification systems (EN). Given this milestone, it is an opportune time to look back over what we’ve learned from advising numerous public health authorities (PHAs) around the world about this technology, and reflect on how the US has done.
Effective against COVID-19
EN apps have proven to slow the spread of COVID-19. While several studies have been published on this topic, the most impressive is the study out of the UK, published in Nature today. There they estimate that the NHS COVID-19 app has averted nearly 600,000 cases and achieved over 50% adoption levels. In the US, we have received anecdotal evidence that in the US people who receive a notification change their behavior as it serves as a “wake up call” to alert them to how prevalent COVID is in their community.
EN has also shown to be an extremely cost-effective non-pharmaceutical intervention. Most states have spent under $1.00 per user and many of the expenses have been related to traditional and digital marketing costs to increase downloads and adoption.
It’s unlikely that any US state is seeing benefits to the same level as the UK due to our lower adoption rates and the fact that most states are on Exposure Notification Express (ENX). The UK, which built a custom app, has more ability to fine-tune their risk scoring in order to get the results they’re seeing. However, especially in states with high levels of adoption, there is still a high likelihood that these apps are making a tangible difference in the fight against the pandemic. Because these systems are designed to be privacy-preserving, it’s been very challenging to get clear data on the efficacy of the apps so far.
Challenges of building a brand new technology
The biggest barrier to adoption by US states was the lack of common deployment guidelines and policies at the federal level. Development was delayed as states waited for guidance from the executive branch. The lack of coherent messaging and the fact that every state needed to solve the same problems over and over again meant that there was a lot of duplicative effort spent on states figuring out how to launch these tools. Federal leadership, in terms of guidance, sample code, a national app, shared infrastructure, a unified marketing campaign, or any other actions would have made a meaningful difference.
Talking to a number of states, we heard the same challenges come up across jurisdictions:
- Many public health authorities don’t have large technical staffs. The states that were most successful had departments of innovation or digital service teams taking a strong leadership role. For those that went with custom apps, choosing good vendors was critical to their relative success.
- Convincing people to upload their keys after testing positive has been an uphill battle for every state. There was a lot of marketing time given to increasing downloads but very little addressing what to do after testing positive. This, combined with limiting code distribution to case investigators slowed down the upload of keys dramatically. As states automated more of their systems, it improved these key metrics but it’s still not at the +80% we see in the most successful systems in Europe.
- Analytics proved a challenge. The design of the system made it initially hard to get good data about usage of the tool. With later updates those analytics tools are now available, only a small number of states have started making good use of this data.
- The amount of planning, work, and money that needed to be set aside to drive adoption was more than the first states anticipated. Later-launching states learned from this to improve their launch, messaging, and PR. The push notifications sent out by ENX systems and ease of opting in to those systems also had meaningful effects.
Overall, the lack of time, staff, and attention were the primary bottlenecks. No surprise considering the massive budget cuts that these departments have seen over the past 10+ years. At the beginning of the pandemic, they were generally focused on scaling up manual contact tracing systems and by the time November came around they had to focus on vaccine distribution. There was very little time for them to dedicate to EN. While the cost was a barrier at the beginning, open source codebases being ready-to-deploy brought the cost down dramatically, and ENX took it down yet another notch.
In a future epidemic, the federal government could choose to launch a national EN solution rather than having each state need to figure everything out themselves. If that were not an option, then the federal government could at least provide infrastructure and guidance – running the backend servers like APHL has been doing, sponsoring the development of an open source app, creating the marketing campaigns and playbook, and making it as easy as possible for each state to deploy. A federal, coordinated marketing campaign would go a long way, along with surrogates supporting the technology in press appearances and creating national support for this type of solution. Having concerted, early support of technology like this could help slow the spread of a disease early on in the lifecycle of the pandemic and would have exponential benefits moving forward.
It is not too late for the federal government to support this technology today – while a federal app no longer makes sense, having trusted spokespeople from the CDC, NIH, or executive branch express their support for and use of an EN solution would increase downloads and ease the burden on beleaguered states.
We’re winning the fight against COVID-19, but it isn’t over yet. We still need to use every tool in our toolbox, and EN remains a cost-effective software that can make a difference in lowering case counts around the US and world.