How smartphone apps could save lives (and the economy)

The economy headed for recession. Companies are going out of business. Employees are being laid off. Social activities have ended. Everyone is hiding and waiting in their homes. All because of the coronavirus, right?

Wrong.

The reason all this is happening is because of ignorance.

Specifically, we don’t know who’s got COVID-19 and who does not have it. Because we don’t know, everybody has to stay indoors, wear a mask when outside, avoid going to work and even avoid buying food at the store. Because people can transmit the disease even when they have no noticeable symptoms, it’s impossible to know who’s infectious and who isn’t.

The overwhelming majority of people do not have COVID-19, but have to self-quarantine as if they do. Because we don’t know.

The knowledge we need — knowledge about who is infected and who is not infected — could be provided by smartphone apps. That’s why hundreds of projects around the world are working around the clock to develop such apps.

The COVID-19 diagnosis

Apps that could diagnose before symptoms appear would enable containment more effective than social distancing rules, according to experts. In theory, AI-based apps could find a way to diagnose before symptoms appear.

Cambridge University’s Covid-19 Sounds project is aiming to use machine learning to identify people with the coronavirus by the sound of their speaking and coughing. Users breath, cough and talk into their PC microphone at a website — plus give details about their age and other variables, as well as whether or not they’ve tested positive for the coronavirus. The AI may be able to reliably learn to detect when people are positive even if they haven’t been diagnosed. The ultimate success would be if people could be diagnosed over the phone. They’re using a website for now until Apple and Google approve their app.

Carnegie Mellon University’s Covid Voice Detector also uses machine learning, based on previous voice-profiling work done at the university. They took it down, however, because they found people took their very experimental results as a diagnosis. They hope to bring it back exclusively for data gathering, rather than for use as a self-diagnosing tool.

How contact tracing works

Contact tracing involves apps that tell you who you came in contact with, so that later, if …

Read the full 1,829-word column.

 

The economy headed for recession. Companies are going out of business. Employees are being laid off. Social activities have ended. Everyone is hiding and waiting in their homes. All because of the coronavirus, right?

Wrong.

The reason all this is happening is because of ignorance.

Specifically, we don’t know who’s got COVID-19 and who does not have it. Because we don’t know, everybody has to stay indoors, wear a mask when outside, avoid going to work and even avoid buying food at the store. Because people can transmit the disease even when they have no noticeable symptoms, it’s impossible to know who’s infectious and who isn’t.

The overwhelming majority of people do not have COVID-19, but have to self-quarantine as if they do. Because we don’t know.

The knowledge we need — knowledge about who is infected and who is not infected — could be provided by smartphone apps. That’s why hundreds of projects around the world are working around the clock to develop such apps.

The COVID-19 diagnosis

Apps that could diagnose before symptoms appear would enable containment more effective than social distancing rules, according to experts. In theory, AI-based apps could find a way to diagnose before symptoms appear.

Cambridge University’s Covid-19 Sounds project is aiming to use machine learning to identify people with the coronavirus by the sound of their speaking and coughing. Users breath, cough and talk into their PC microphone at a website — plus give details about their age and other variables, as well as whether or not they’ve tested positive for the coronavirus. The AI may be able to reliably learn to detect when people are positive even if they haven’t been diagnosed. The ultimate success would be if people could be diagnosed over the phone. They’re using a website for now until Apple and Google approve their app.

Carnegie Mellon University’s Covid Voice Detector also uses machine learning, based on previous voice-profiling work done at the university. They took it down, however, because they found people took their very experimental results as a diagnosis. They hope to bring it back exclusively for data gathering, rather than for use as a self-diagnosing tool.

How contact tracing works

Contact tracing involves apps that tell you who you came in contact with, so that later, if someone tests positive, you can self-quarantine based on the real possibility that you got the virus. These apps could also help you avoid the infected.

Contact tracing holds out the promise that by knowing you haven’t been exposed to people who test positive, you can live and work like you normally would.

In general, some Asian countries were first to “market” with contract-tracing apps, and these are part of the success of containment in those countries.

In China, where the coronavirus first began, peaked and got under control, the government uses an app that color-codes everyone. Called the Alipay Health Code, the system uses Alibaba’s Alipay ecommerce platform to assign everyone a color code — green, yellow or red — to indicate whether each person is required to stay quarantined or is allowed to move freely in society. Data from the app is also reportedly shared with police. Before entering a market, store, public transportation or other public place, people have to scan their QR code and only those believed to be immune to coronavirus are allowed in.

South Korea, which reported its first coronavirus case on the same day as the United States, now maintains a new daily case report to about 50 people. The country achieved this through a variety of aggressive measures, including smartphone-based systems. One South Korean app tracks everyone who has tested positive and warns others when they are nearby via a required SMS alert system. (One downside is that residents are starting to suffer from “alert fatigue”).

Singapore launched a Bluetooth-based contact-tracing app called TraceTogether. The government claims that personal data is encrypted and deleted after 21 days, although privacy advocates are skeptical of that claim.

While these Asian apps are partly responsible for relatively low new infections in those countries, they’re also blamed for increasing state control and surveillance and further eroding the privacy of citizens in those countries. In the West, there’s a far greater emphasis being placed on developing apps that offer privacy or even anonymity.

Who’s up for apps

The Italian government has called for proposals for a smartphone app that will help trace contacts by people who test positive for the coronavirus. Users would be volunteers, whose anonymity will be “guaranteed.” The app will record when the phone came close to another phone and for how long. When the owner of one of the phones tested positive, everyone they came in contact with would be notified. This would enable selective, pre-emptive self-quarantining of those exposed, and also enable the non-exposed to work, travel and go about their lives.

Poland launched an app for enforcing compulsory self-isolation.

Experts from 18 European countries, including Germany, France and Switzerland, are working on a technology platform that would support Bluetooth-based contact tracing apps and is based on tough European Union data privacy rules.

Ireland is expected to launch in the next two weeks that citizens would voluntarily use.

The UK is thinking about launching an app.

The Swiss Federal Institute of Technology is creating a Bluetooth-based proximity monitoring app.

Germany intends to launch a new app for contact tracing. It’s currently being tested by German soldiers.

Despite all the national projects in Europe, The EU’s data protection watchdog has called for a single, European app that everyone in the EU would use instead of individual apps in development in every European nation. The zone’s General Data Protection Regulation explicitly allow the collection and processing of private information for the purpose of public health.

Many different states, universities and companies in the United States are also working on contact-tracing apps.

North Dakota this week rolled out an iPhone app called Care19 created by ProwdCrowd that records contacts lasting ten minutes or more, and informing those contacted when someone tests positive. An Android version is expected soon. Each user is identified in the system with a randomly assigned ID number. If they test positive, they can consent to giving information about who they were in contact with to the state. The app was repurposed by Microsoft software engineer named Tim Brookins, from an app called Bison Tracker he created for football fans.

A company called COVID Watch worked with Stanford University to create a contact-tracing app also called COVID Watch.

Researchers at Boston University have invented but not yet built an app that uses short-range transmission technologies, including Service Set Identifier broadcasts, near-filed communication (NFC) and Bluetooth to generate random ID numbers for contact tracing, to enable the app while preserving anonymity. The app would broadcast a random token number that would change every few minutes to maintain privacy and anonymity. Each app would collect tokens of anyone else using the app who came near. Then, the data and tokens would be uploaded to the CDC to track cases. It would also alert users if they came in contact with someone who tested positive.

A GPS app called Chirp GPS has been given specific Covid-19 functionality. Normally, Chirp GPS is used for tracking friends and family members. But now the app asks you questions about your own coronavirus and health status. Users can then see the location of people who have tested positive.

Crowdsourcing processing power and knowledge

Vodafone launched an app that contributes the processing power of each user’s phone to tackle the pandemic. The so-called DreamLab app has existed for years to provide processing power used to for all kinds of disease research. Now they’re working on the coronavirus.

Pinterest CEO Ben Silbermann worked with scientists at Harvard, MIT, University of Pennsylvania, Stanford University and Weill Cornell Medicine and other schools to create a non-profit and app called How We Feel for people to self-report their symptoms daily. Users also add their age, gender and other details. The app helps researchers spot outbreak locations or coronavirus clusters.

UC Davis researchers launched their own tracking app, which is designed to work globally. Its purpose is to share real-time information in the form of “graphs on demand.” As with some other projects, the UC Davis app has been repurposed from an existing app, which was created to track outbreaks of mosquito-borne diseases.

Stanford Medicine worked with Apple to create a COVID-19 app for first responders. Called the First Responder COVID-19 Guide, the app helps police, paramedics and others to monitor their own symptoms to get early testing if necessary. The app takes advantage of Apple’s ResearchKit and CareKit frameworks, and offers up-to-date information from Stanford researchers.

There’s just one problem

If you haven’t spotted the problem with this global movement toward coronavirus-fighting apps, I’ll spell it out: There are far too many of them.

This is especially true in the United States.

For example, if I’m using the Stanford app, and my friend is using the Boston University app and our co-worker is using an app developed by the state, then we might as well not be using any apps at all.

The ideal solution would be a single, global app that both accurately uses AI for self-diagnosis and also performs high-quality contract tracing and provides detailed data to researchers, all with perfect anonymity and privacy.

That’s never going to happen. So the best case scenario is that we have open standards and interoperability for data and information exchange so that different apps can work together toward the same goal. And it’s not clear what the best way forward is toward that goal.

An anti-coronavirus app could save lives and save the economy and enable every healthy person to live and work normally. But this is possible only if everyone is using either the same app or using an open app that shares data with other apps.

Reason for hope in beating the pandemic

That’s why I’m optimistic about the news announced yesterday (Friday) that Silicon Valley smartphone superpowers Apple and Google have announced a joint effort for a global standard for contact tracing functionality.

The Apple/Google solution will include APIs both companies promise to release in May, which any company, nation, state or university can use to make their own contact-tracing apps work with everybody else’s.

Also: Both Apple and Google announced their intention to bake Bluetooth-based contact tracing functionality into their respective mobile operating systems — iOS and Android.

Google and Apple also plan to release draft technical documentation so that other organizations can get cracking on interoperable solutions. 

I can’t overstate how good this news is. Interoperable smartphone contact tracing could bring about the end of quarantines, company closures, layoffs and recession.