No items found.
Blog /

Datavant Future of Healthcare Hackathon: Best Public Health App

Author
Publish Date
Read Time
Datavant
September 22, 2022

From September 8–11, Datavant hosted its first annual Future of Healthcare Hackathon. Over 200 attendees spent the weekend developing innovative solutions to improve the future of healthcare. This week, we are excited to announce the various winners of the Hackathon.

The Winner of the Best in Category — Public Health App is: Exquisicare. Read on to learn more about this fantastic project.

Development Team:

Edward Gent — Technology Associate at Morgan Stanley / Founder at Health Haven Technologies

SubbaRao Bellamkonda

Project Summary: Exquisicare — Using chat to treat opioid addiction

Since moving to the U.S., Edward Gent has been struck by the severity of the American opioid crisis. Bringing his broad awareness of comorbidities between different types of addiction, he aimed to leverage current approaches to cell phone addiction as a means to treat more serious and sinister opioid addictions. To do this, Gent built an automated Instagram chatbot, Exquisicare, that responds to DMs and advises people concerned about their risk of addiction on their actual risk. His model is built on a CDC data-trained classification model, and after engaging with a user, the chatbot can then refer that person to medical professionals if they so wish.

Project Inspiration

Gent is a qualified nutritionist and via his startup, Health Haven, he assists people at developing nutritional optimization in partnership with the GBS/CIDP Foundation. He is deeply passionate about improving people’s overall health, and hopes to inspire others to develop a similar passion. While working on this project, Gent came to appreciate the magnitude of the opioid crisis in the U.S., which is currently growing 30% YoY, a growth rate he hasn’t observed in other medical afflictions or causes of death. This was personally relevant to him, having recently undergone surgery and been given an opioid prescription, which came with a letter of warning regarding post-recovery addiction risks.

Approaching the Problem

Gent wanted to leverage existing technology to build a scaled approach capable of moving a large number of people to take action. After seeing a clear correlation between opioid related deaths in the US and the number of Americans who admit to spending over half of their day on their phone (>7 hours on average!), he researched open source Instagram libraries that facilitate automated messaging.

Exquisicare chatbot

Using Python, Gent trained the classifier to make predictions based on patient input and embedded the app with a Firebase backend to securely store conversation metadata and intents. In his choice of Python, Gent traded-off quality of model performance (little/no chatbot model confidence scores or classifier precision/r1/recall) in return for a pipeline that worked end to end. Because of the time constraints of the Hackathon, he was not able to set up a cloud deployment or incorporate a multiprocessing toolbox to allow multiple conversations to be had at once, both of which, he acknowledges, would be necessary for the project to scale. But because he used open source Instagram libraries, he was able to spend less time on UI/UX considerations and more on HIPAA compliance, data encryption, and implementation.

Implications

The goal of this app is to help reduce the >91,000 annual opioid related deaths in the US. Gent notes that in the face of such numbers, “Virtually any positive steps in this direction would be a huge win.”

Future Steps

Gent used CDC data provided by Datavant, which meant that, in his words, “the model code was a little brittle.” He also noted that the hyper parameters are probably overly biased and not suited to fitting predictions outside of the provided input feature space. The data largely considered addiction rate as a function of features that most people are unaware of and that are not “personal attributes,” which made it a challenge to develop useful questioning of the end user.

A big next step would be to expand the data sources used to train the app. In the longer term, the chatbot could also be used to gather patient data, including demographics and experience with addiction, which could be used to provide insights into the broader non-patient population.

Exquisicare was developed by:

Edward Gent, Health Haven Technologies and SubbaRao Bellamkonda

Congratulations to Edward for developing this project!

Considering joining the Datavant team? Check out our careers page and see us listed on the 2022 Forbes top startup employers in America. We’re currently hiring remotely across teams and would love to speak with any new potential Datvanters who are nice, smart, and get things done and want to build the future tools for securely connecting health data and improving patient outcomes.

Achieve your boldest ambitions

Explore how Datavant can be your health data logistics partner.

Contact us