Inspiration
The Biotech prompt stood out to us, we felt that we could build something with true purpose if developed well and something that potentially make an impact in the medical field by catching some medication errors when prescribing drugs to patients. Even regular users of medication can use the app to check if their prescriptions can cause negative interactions (like two drugs could potentially negate each other's effects)
What Chill Pill does
In healthcare, medication errors often happen when past prescriptions are copied incorrectly or incomplete information is carried forward. Our application uses historical patterns to guide current entries, flag risks, and prevent repeated mistakes by showing warnings when dosages outside the usual range are prescribed or when two drugs with potential negative interactions are prescribed.
How we built it
The backend portion uses python to merge CSVs to create a large file full of drug-drug interactions. JSON files were made to store data and a local list of common drugs prescribed so that when the API key fails or is slow, at least the local data can be used. JavaScript/React was used to get the API key from RxNorm to match the user-entered medication to the database in NIH.
Previously: frontend uses primarily Figma to showcase the app and manages the code internally using various plugins (like Kernel)
Now: Frontend is made with AI Studio to better accommodate the fact that multiple source files are needed and Figma's Kernel plugin did not work for us.
Challenges we ran into
Finding the right url addresses to the right websites and databases took a while and the most difficult part was integrating the code to our frontend; there were a lot of hiccups with getting plugins and figuring out where to add code and how the final product should look like. At the end, we made a quick pivot to AI Studio. Though it may seem trivial, it took us a solid minute deciding which color palette and fonts to use and we hope that time spent paid off!
Accomplishments that we're proud of
The user interface and logo feel recognizable and memorable, as we feel that the frontend is the first thing users see. We also hope that the smoothness of the matching data pays off because the backend has two databases (local and external), not just one.
What we learned
As a team, we collectively learned from each other's skills. Each one of us now has a GitHub repo up and going, figured out the basics of AI Studio and we learned to collaborate even though we don't know each other well, because teamwork is all about communication.
What's next for Chill Pill
Given more time, there could be a voice agent added to the medication pages, where it reads out summaries of the data given, highlights the potential problems audibly, and might even automatically read out the drug name with a generic name (ex: instead of acetaminophen, it would read out Tylenol). This would be particularly useful to the elderly since it can be difficult to read
Log in or sign up for Devpost to join the conversation.