Clinix

Clinix is an AI assistant that streamlines hospital front-desk tasks—scheduling, records, and insurance—quickly and securely while protecting patient privacy.

Comment

Inspiration

Clinix was inspired by the growing capabilities of large language models and their potential to streamline complex tasks. One of the fields that would benefit the most is healthcare. We wanted to harness the power of AI and make the healthcare experience better without compromising privacy. We want to use LLMs, but design a system such that it can do all the tasks without sacrificing integrity.

What it does

Clinix helps with three tasks in the front end of hospitals or clinics. It helps make appointments, find medical records, and estimate costs based on the service and insurance. For appointments, Clinix automatically finds available doctor times using only a generic calendar. No names, personal details, and no sensitive information are ever used, keeping scheduling fully private. For medical records, Clinix uses the patient ID to locate records on a skeleton spreadsheet. Only the location data from this spreadsheet is then used to access the real databases, so the AI never sees sensitive patient information. For costs estimates, Clinix uses only generic information about the provider, insurance company, plan, and service. This allows it to quickly estimate coverage and costs without accessing any personal patient data. We also developed a chatbot to help with general health questions and concerns. It has an escalation system in cases requiring immediate professional care.

How we built it

For the front-end, we used JavaScript and HTML, with a React framework. It runs on Vercel, and the backend is powered by Render. We used Python with FastAPI and JavaScript, and Tailwind for the Backend.

Challenges we ran into

We struggled a lot with implementing the AI chatbot. We also had to learn how to use the different APIs, including Gemini, Google Sheets, Google Drive, and more. We had to design a system that keeps private information truly private and not passed into LLMs.

Accomplishments that we're proud of

As all first-time hackers, we are proud to get a product by the deadline. I think we chose something quite ambitious, and we were able to pull it off. We used a lot of different tools, and getting them to work in the end was great.

What we learned

We learned a lot. We had to learn a lot of tools like APIs, how to incorporate outside tools, and get a better sense of how a project goes through the development cycle. We gained experience working with a lot of technology we don't usually work with.

What's next for Clinix

Clinix can be tested in local clinics and hospitals. Additionally, we can add more features. For example, the medical records could generate a PDF in line with a provider's format. The formatting for our insurance and medical records page could use work. The chatbot is imperfect, and we could exapnd its functionality. We could use a better database system, as Google Sheets may not be powerful enough for the scale of a large hospital.

Built With

Share this project:

Updates