Joanna Mazel –
Joanna Mazel Quickly realizes that the contact tracing apps built to track COVID-19 cases are not as good as they could be. Apps built on the platform offered by Apple and Google are not used by many people in the United States and do not seem to have a big impact on the spread of the virus.
But Mazel, an evolutionary biologist at the University of Arizona and an advisor to the Arizona exposure notification app development team, believes there are ways to make these apps more useful. In an article published last week with colleagues, she described one possible improvement: tweak apps so that they can better recognize situations where people exposed to the virus have a lower risk of contracting COVID-19. Then they may not need to be quarantined for the full 14 days.
“The capability of this technology is much better than the best implementation at the moment,” she says.
Most of the apps used in the US warn anyone of a potential COVID-19 infection if they are in the same area with a potentially infectious person for more than 15 minutes. This is a rough metric for determining how risky the interaction was, and the recommendations from the app tend to be just as broad: they instruct people to isolate themselves.
More nuances can be added to this process. Mazel’s paper describes a method of including information about where someone was in relation to a sick person, which can be an indication of how much virus they might have been exposed to. Someone closer to or directly in front of a person with COVID-19 will inhale more of the virus than someone a few steps back and to the side. When combined with other information (such as how long the sick person had symptoms before the interaction, or whether the exposed person had symptoms), the app could provide more specific guidance on how long a person should be isolated.
“The idea is personalized public health,” says Mazel. Rather than treating interactions as equally risky, it could help people understand how serious their risk of contracting COVID-19 from the interaction is. “Then you have the opportunity to simplify things to very simple things, like getting tested on day x and staying home on day y,” she says.
The way Google and Apple are building the foundation for an exposure notification system makes it difficult, but not impossible, Mazel said. This can be done with a privacy-preserving framework. Developers will have to combine various pieces of available data (such as Bluetooth readings and how long someone was sick) to figure out which interactions are more likely to lead to infection.
Providing personalized risk assessments and recommendations is where exposure notification apps can be better suited than traditional manual contact tracing. “It’s too difficult for phone trackers to use extremely complex algorithms,” she says.
Improving an app’s ability to detect interaction risk could also make it more effective in slowing down the spread of COVID-19. In the United Kingdom, where about a quarter of the population used an exposure notification app, one analysis found the program prevented hundreds of thousands of cases. Most of the benefits were concentrated in the months following the update to the app’s Risk Assessment feature.
Masel believes that impact notification technology with changes could be useful even after a pandemic. COVID-19 will not go away and people will continue to get sick after the pandemic phase of the outbreak. Outside of an active pandemic, apps may not ask people exposed to the virus to isolate – they might just encourage people to get tested or direct them for treatment, Mazel says.
But the decision to leave the technology in place is not a public health challenge. Last year, Google and Apple said they plan to close the exposure notification program after the pandemic ends, shutting off access to resources for government health authorities. Massel says this is not their call. Why is this solution owned by Apple? she said. “Apple and Google don’t have to make public health decisions. I’m not saying that decisions are lazy or that my views are the only ones, but I don’t think it should be a decision-making process. ”
Instead, she said, experts must find ways to improve the system. Even the UK app development team does not consider themselves well done, Mazel said, because it was built in a rush. “We had to learn lessons on how to increase reach, better assess risks and design user interfaces – we can improve everything.”