Currently a lot of people who show symptoms of corona, who are not tested, rather asked to go home, self-isolate and rest, only visit hospitals once their conditions get worse. By which, it means emergency rooms. This has led to wild mortality rates especially at nursing homes or at home deaths. What if we could create a system that detects and notifies such people, who feel they have symptoms but are not tested based on known biomarkers.
What it does
Biomakers such as temperature, heart rate, blood oxygenation are easy to monitor and good markers of COVID infection. The idea is to use sensors such as PPG and temperature sensors to detect anomalies in physiological markers communicated directly to an app through IoT/Cloud based systems. Based on regular measures, the system can notify your primary GP and on whose recommendation further testings can be taken.
How I built it
The system is composed of a hardware module, a mobile/web application front end and a cloud analytics backend. The hardware module, of the size of an euro coin, records temperature, heart rate, respiratory rate and blood oxygenation in real-time and at a regular basis. Physiological data are transmited through IoT technology to a Smartphone app, which automatically communicates it to the cloud server. Subsequently, an analysis based on machine learning algorithms is performed through the Cloud app to detect alerting COVID infection. If a COVID infection is suspected, the smart app creates an alert and send a report to the GP. Our system is user-friendly, totally open-source and can be produce at low cost.
Challenges I ran into
We started from scratch. We performed an extensive literature review. We developed:
- a PCB design
- a first functional device prototype
- a Cloud app
- a mock up smart app
- a business plan
Accomplishments that we are proud of
The idea came from a simple exchange of emails between two strangers looking for a team on Friday evening. After 48h, we are proud of presenting Vitalcatch, an innovative project based on the collaboration of a great multidisciplinary team. We succeeded to develop a first functional prototype able to record physiological data, a Cloud app able to collect and analyze data and a bussiness plan based on litterature review.
What's next for IoT sensors to detect infection
We still need to develop the machine learning analysis and validate our prototype. Over the next three months, we plan to launch a pilot study involving 100 nursing homes across Europe on a quarterly subscription based model. In the meantime, we want to expand partnerships with research institutes and exposure tracking solutions such as Ohioh or Safepaths.