Every day, health professionals risk their lives while saving ours. 25% of current COVID-19 cases in Rhode Island are health professionals, while 20% of NYC healthcare workers are projected to contract the virus over the course of the pandemic. While there are several creative solutions being created to combat coronavirus in patients, we wanted to focus on the heroes that are most at risk: frontline healthcare workers.
Every person who tests positive for coronavirus is projected to go on to infect multiple people. A doctor can see upwards of 20 people on a 12-hour shift, making the likelihood of exposure and ultimately infection in healthcare workers extremely high.
After interviewing 4 hospitals and a dozen healthcare leaders, doctors, nurses, and volunteer testers, we realized that many hospitals do not have a system in place to keep their healthcare workers safe, despite them being the most at-risk and the most important population.
Our team was inspired to combine our backgrounds in engineering, product, & research to create a tech solution that can create security & awareness for these heroes.
That's why we created Harbor Health.
What it does
Harbor Health is a team health monitoring platform that:
1) Gathers health data from different biometric sensors (the software is sensor agnostic). Currently we are working with LifeSignal to implement their FDA-Approved wireless disposable sensor into our platform. Check out the sensor here.
2) Allows individual healthcare workers to track their own health data via a mobile app
3) Helps administrators effectively track their teams' health (exposure and symptom intensity) through a consolidated web dashboard.
4) Texts users if they start to show any symptoms for infection.
Try out our dashboard here.
How we built it
We built a Mobile app using:
- Kotlin, Java, XML
- Volley, Firebase, LifeSignal SensorProc library
- Volley allowed for an asynchronously executed request queue to handle our custom-built API calls
- LifeSignal SensorProc library converted raw ECG data into filtered signal (mV) and heart rate (bpm)
We built the server using:
- A Python flask server architecture with endpoints for front-end and back-end API calls
- Heroku to deploy the server.
- Firebase to allow for data to be streamed quickly and in realtime from the backend to both the app and the web platform
- Twilio for text messaging
- Google Cloud Functions to track trigger events given user data creations and updates (huge shoutout to Wesley and Ryan from Google for helping us out with this and giving us feedback on our project)
We built also built a web app using:
- React.js deployed via Netlify as a single page application which communicated with our Flask API server
Challenges we ran into
- Asynchronous transmission of data (utilized Volley library)
- Realtime data visualization (utilized Firebase for data streaming)
- Generating useful high level metrics for the large amounts of data our sensor collects
- Balancing conversations with healthcare workers (because their priority is helping other people).
What we learned
On the engineering side, we had to learned how to:
- develop an Android app
- perform asynchronous requests using Volley
- work in Kotlin and XML
- how to use Netlify
- Automated text messaging through Twilio
- Serverless Function Calls & Database On-trigger Events with Google Cloud Functions
In general, we also learned that:
- The general lack of testing in the US means that hospitals cannot afford to test their own workers for Coronavirus unless they display symptoms, despite repeated exposure.
- For what little COVID-19 symptom testing there is, there is little to no reliable data tracking
- Hospital materials (PPE, ventilators, ICU beds, etc) are extremely sparse. As more healthcare professionals are infected, hospitals will face a lack of frontline workers too.
What's next for Harbor Health
Check out our slide deck here.
Short term: Harbor's immediate entry point is frontline healthcare workers. We would like to first test our product in hospitals near the LA area, where most of our team is located. Our goal would be to implement this system in 1-3 hospitals in LA by July 2020. After making sure our systems work well in Southern California and that our system is scalable, we would like to expand to 3 hospitals where we are most needed: NYC.
Long term: Apple, Facebook, and Amazon have all recently announced initiatives for temperature checking, virus testing, and limiting the scope of employee events. We believe that these are only corporations' first steps, as warehouse, rideshare, and factory workers are severely at risk of COVID-19.
We also believe that there is a large home consumer market. Due to the relative unavailability and inelasticity of current Coronavirus tests, any service that focuses on providing household consumers reliable health tracking methods to show possible indications of Coronavirus could help streamline the flow of patients most in need, while providing a sense of security for those who aren't able to go outside.
EDIT: For any questions about the code/repo, please contact us! For now we've made our repos private as we continue the project :)