We want to fix healthcare! 48% of physicians in the US are burned out, which is a driver for higher rates of medical error, lower patient satisfaction, higher rates of depression and suicide. Three graduate students at Stanford have been applying design thinking to the burnout epidemic. A CS grad from USC joined us for TreeHacks!
We conducted 300 hours of interviews, learned iteratively using low-fidelity prototypes, to discover, i) There was no “check engine” light that went off warning individuals to “re-balance” ii) Current wellness services weren’t designed for individuals working 80+ hour weeks iii) Employers will pay a premium to prevent burnout
And Code Coral was born.
What it does
Our platform helps highly-trained individuals and teams working in stressful environments proactively manage their burnout. The platform captures your phones’ digital phenotype to monitor the key predictors of burnout using machine learning. With timely, bite-sized reminders we reinforce individuals’ atomic wellness habits and provide personalized services from laundry to life-coaching.
Check out more information about our project goals: https://youtu.be/zjV3KeNv-ok
How we built it
We built the backend using a combination of API's to Fitbit/Googlemaps/Apple Health/Beiwe; Built a machine learning algorithm and relied on an App Builder for the front end.
Challenges we ran into
API's not working the way we want. Collecting and aggregating "tagged" data for our machine learning algorithm. Trying to figure out which features are the most relevant!
Accomplishments that we're proud of
We had figured out a unique solution to addressing burnout but hadn't written any lines of code yet! We are really proud to have gotten this project off the ground!
i) Setting up a system to collect digital phenotyping features from a smart phone ii) Building machine learning experiments to hypothesis test going from our digital phenotype to metrics of burnout iii) We figured out how to detect anomalies using an individual's baseline data on driving, walking and time at home using the Microsoft Azure platform iv) Build a working front end with actual data!
Note - login information to codecoral.net: username - test password - testtest
What we learned
We are learning how to set up AWS, a functioning back end, building supervised learning models, integrating data from many source to give new insights. We also flexed our web development skills.
What's next for Coral Board
We would like to connect the backend data and validating our platform with real data!