By mid-February with the virus getting closer to Spain we come up with a solution that would be anonymous for the users, easily adoptable by anyone with an smartphone and cheap by using the mobile phones sensors. By 6th of March we had a MVP working version ready and we wrote to different Spanish administrations to let them know, but got no response. We have continued developing this solution for some more weeks adding improvements and working on moving it from a prototype to a market-ready version. In early April we discovered the PEPP-PT and DP-3T initiatives and checked them, we can say our design is very similar to initial one proposed by DP-3T on his white paper. A decentralized design that fully respects privacy and doesn't uploads any data to any cloud.
What it does
It helps stop the spread by making every citizen aware of his risk of being infected with covid19, therefore helping them to take preventive actions, like calling a doctor for a test based on exposure to confirmed cases provided by the app, and thus helping the administrations to do more selective tests and lock-downs. It works as follows: It constantly scans nearby devices and stores their anonymous IDs. Everyday downloads from a server a lists of anonymous IDs that have been confirmed as positive by an authorized medical institution. It searches for matches of any of those anonymous IDs with any of the ones you have on your local encounters database. If anyone found, it computes the encounters startTime and endTime and provides user with the amount of time of exposition to confirmed cases over the last 20 days. Everyday the encounters from longer than 20 days are removed (the app is data-driven so those values can be modulated based on new clinic evidences).
We have developed a secondary app for the medical authorized users, that allows, only to a pre-autorized personnel, register users as "confirmed with covid-19". That app doesn't take any personal data of the patient, just scans the code and stores it on the server, so that anonymous ID appears on the list of "confirmed cases". During the process of scanning for confirmation, a new anonymous ID is given to the user, so no one can relate him with an active case.
All the numbers are configurable. We expect the public administration and researches to provide us with the correct numbers (like minimum distance for contagion, minimum time to be at risk, quarantine days, messages with indications to display to the users, etc).
How I built it
We are videogames and apps development startup, so far the company has paid some of his workers to develop these two apps (in total has expend around 5k-6k€, not looking to recover it, but is worth mention to show how much implicated we have been on that).
Challenges I ran into
We run into some problems with Bluetooth detection technology, so we have used different approaches. Getting the app not to be killed and be resilient also took us quite a lot of time. And testing a tracing app during a lock-down has not been easy either, luckily family and friends has helped a lot. Administration collaboration: This app can't work/can't be used without the collaboration of the administrations. We need them in order to better understand the medical requirements, to polish the app according with whatever restrictions they need to be in place, to conduct testing groups, to know if they want just one app or they want to have an app that integrates with other public services apps or any other requirements they might have. Initiatives might come from private sector or individuals, but without public administration collaboration there is nothing we can do.
Accomplishments that I'm proud of
To have had working version ready by 6th of March and having notified it, much earlier than the local administrations figured out the importance of that. To have created a protocol and an architecture design that is almost the same as the one proposed by many doctors on white papers. To have been focus on the solution and worked hard on it just because we know something like this is needed, without thinking about his profitability.
What I learned
That reaching the administration is very difficult and even in situations like this one they don't pay attention to you if you are not friend of someone or you write a paper with several "Dr." title on it, your proposal unlikely will be attended.
What's next for Heroes
See if we can get attention from public administrations to conduct tests and check if it's really worth for us to continue investing on that development. We have an app which with few close group tests could be ready to hit the market and help to deescalate the situation. In Spain for instance the economic damage performed by the quarantine is enormous and we have started the deescalate without a fundamental tool, which might lead to new lock-downs and severe economic impact. The app can run itself or can be easily integrated on other apps. But on top of that we have started working on a super simple SDK with just two calls:
- OnExposureDetected; So anyone can create their fancy UI/UX or tools, and add up the tracing logic by just implementing those two functions.
TO BE ABE TO DOWNLOAD THE APP FIRST YOU NEED TO JOIN THE PRIVATE TESTING GROUP: https://groups.google.com/forum/#!forum/heroes-test-group TO TEST THE APP: USE INVITATION KEY: 01234