One of our members’ personal experience with a family member.

What it does

Connects patients with one another to form accountability relationships and improve medication compliance. We combine numerous studied components which have been studied to show improvements in compliance when integrated into a single solution.

How we built it

On the front end we would use Swift for iPhone and Go for Android app development. On the backend, the platform with the API endpoints and Jupyter machine learning modules would be running on Intel Nervana AI Cloud Compute. The users are initially matched randomly according to their initial preferences. As they are using the application, their voices are transcribed by Google Speech API. The transcriptions are then combined into single file. Each person's conversation history and the compilation are separately sent to the IBM Watson Personality Insights API, where the outputs are assigned.

Challenges we ran into

The need is incredibly broad, it was hard for us to pinpoint exactly which area we wanted to design our solution around. It was also disheartening to see so many solutions already on the market that were not actually improving outcomes. By spending more time on the research side, we managed to both narrow down our solution and discover more meaningful methods for improving outcomes.

Accomplishments that we're proud of

There are many solutions already on the market that aim to increase medication compliance/adherence, however none of these solutions work on their own. We managed to create a solution which seamlessly integrates these simple solutions into a comprehensive application which addresses the key components that have actually been shown to work.

What we learned

A strong focus on need-finding, brainstorming and early research leads to stronger ideas which have a greater chance of positive impact for patients and the healthcare system as a whole. Going down the rabbit hole with your first idea could lead to large amounts of wasted time on a product that you later discover is infeasible.

What's next for MedBud

We will be continuing development and raising seed funding to complete a MVP and then conduct a pilot study within the VA health system. We will use the data to reiterate the design and make the application ready for an official road out. Meanwhile, we will be expanding and developing a more complete business model.

Built With

Share this project: