Covid-19: Contact Tracing Apps for Proximity Alerts
Covid-19: Contact Tracing Apps for Proximity Alerts
Let's see how developers are making contact tracing apps and how GPS and Bluetooth technologies are helping countries fight coronavirus.
Join the DZone community and get the full member experience.Join For Free
One of the important tools China used to deal with the Coronavirus pandemic attack is contact tracing. A similar system is being deployed in India. Europe and the USA are also attempting to do this but privacy-related matters make it hard for Europe and the USA to roll this out. In this article we will dig into various ways to implement contact tracing.
Contact Tracing During Health Pandemic
Contact tracing involves identifying individuals who may have come into contact with an infected person. For example, in the figure below, the person inside the circle is infected. He has come into contact with others. The key to contact tracing is to identify who has potentially come into contact with this person.
In this example, 8 people have come into contact with this one infected person. Next step is to identify which one of those has been infected. For example, two people have been confirmed. Three others don’t have symptoms and they may be tested or put in quarantine for the recommended period. Not everyone who comes into contact can be identified/reached. Here the two people in blue appear to be at risk for themselves and others because they may not be reached.
For those who are infected, the process continues to identify who may have been in contact with them until all the people who have been in contact downstream have been contacted and confirmed to be not infected.
The Technology Behind Contact Tracing
Let’s assume that the GPS data is readily available. In this case, once we know the GPS data of the phone of the infected person, the next step is to identify all the phones that were within the vicinity of that phone. Vicinity could be based on distance or it could be based on a location. For example, if the person was in a restaurant or a mall or a religious service, then everyone who was in the same location can be suspected to be infected. If it's an open area, then it could be based on distance affinity. When defining this distance affinity one has to be very cautious about the population density. The higher the population density, the more care needs to be taken for identifying the potential suspects.
What Type of Data Is Needed?
As mentioned earlier, for this to work, one needs to have access to GPS data from the phone of the affected people. In some countries, the government can request this kind of data during emergencies to tackle a health pandemic. In countries like the USA, getting access to this data is hard because of privacy issues.
In the absence of this data, what options do we have for contact tracing?
Data Access and its Limitations
In many situations, GPS data may not be available because of laws in different countries. For example, in Europe and the USA, because of privacy laws, one cannot access the GPS data to be able to do the required analysis for contact tracing.
In the next few sections, we will discuss a few options for building out applications that work around these limitations and discuss their pros and cons.
Temporary Updates to Law in Affected Areas
Is it possible for our government to pass a law to have access to location data in affected areas for the past 30 days to the next 60 days? My immediate reaction is, this is not going to happen. I will be pleasantly surprised if this happens.
Crowdsourcing the Location Data
The idea here is users can request their provider to volunteer their location data to a 3rd party service. Say one of the people who is registered is infected. If another person volunteers their location data, the app will be able to provide guidance if one should be tested based on possible contact with the affected person.
Downside with this approach is, there is no guarantee many will volunteer their data. So if there are only some volunteers to share the data, we still have the risk of those who are affected but did not volunteer the information could potentially spread without knowing especially if they are asymptomatic.
Data From a Custom Coronavirus App
Imagine a contact tracing app that keeps track of the GPS location of the phone for every 5 mins. The location data about the phone is captured locally on the phone and retained on the cloud for further analysis and tracking.
If the same application is installed by many, if the data is available for who is infected, the app can do analysis and send notification to all other app users if they were in close proximity to the infected person. The downside to this is, it works only for people who installed the app and it only works for the future contact. In earlier contacts it would not know.
This is the area where many startups are / will be focused on to create proximity apps. Proximity or location data of the user is used to notify someone who has recently been near to an infected person, so that they can take preventive action. How much long this data should be saved is still an open question.
Tagging Data Using Bluetooth Enabled Apps
This is based on apps talking to each other and recording their signature based on proximity. Each phone will record the app signature of all the phones in its vicinity. This time stamped data can be stored in the cloud for further analysis. If a person is infected, this person can alert the app. Once this is done, the app can identify & notify all the users who were within the vicinity of the infected person based on the app proximity location signature captured. This would work as long as there is mass adoption of the app. If the person who is infected does not have the app, the whole system fails.
Challenges With GPS-Based Data
Another challenge with the GPS data is, one can be on a different floor in a mall and the infected person may be on a different floor. It's not an exact science.
They could be passing each other on two different floors. Using GPS it may appear they were too close but in reality they were on two different floors.
Countries like China, India and South Korea are using health emergencies to bring in real time surveillance. This has shown clear positive impact to reduce the cases. We have to figure out ways to do this by preserving privacy which is a broad challenge. That is when these kinds of proximity apps are very useful if it gets mass adoption. Key is mass adoption. Only some use it will not work.
This is especially very useful after relaxing social distancing so we have a mechanism to defend against second wave of infections.
In the next article, we will review more details of the analytics that goes beyond the app.
Opinions expressed by DZone contributors are their own.