MediBill

Inspiration

We were inspired by the situation our team members grandmother is facing as she is constantly in the hospital and is burndened by the overpriced medical bills. We are determined to put an end to the medical overcharging epidemic that plagues Americans day in and day out.

What it does

Our project analyzes a user-inputted fully itemized medical bill and flags each inflated price, explained what an average price would be compared to what you were charged. It then calculates the estimated amount that you were overcharged and how much you can get back. To assist in resolving the user's situation, our program will generate a well-written letter directed to the hospital in order to dispute these expenses.

How we built it

We created MediBill using the experiences that each of our group members have obtained from past Hackathons and projects. The first thing we did was use Streamlit to create a clean interactive web interface. We created a data prossesing system that allows the AI to analize the costs of each type of machinary or prosedures used during the users hospital visit in order to clean inconsistant formatting or normalize prosedure names so they can be analyzed accurately. We also created a price comparisan engine which estimates the average costs of each medical service and compares that to the amount that the users were charged which then caculates the differences and flags the items that appear significantly inflated. We integrated Groq API in order to generate a clear explenation for each inflated item. Finally, the aplication attamatically generates a professional, well written letter addressed to the hospital bill department, which users can download and send.

Challenges we ran into

One of the biggest challanges we faced was the inconsistancy of the medical bill formatting. Different hopitals use differnet abreviations and formatting which made categorizing line items difficult. Another challange we faced was finding reliable benchmark prices for our AI to use becuase of the lack of tranparency many healthcare services provide.

Accomplishments that we're proud of

We are proud that we were able to turn a complex healthcare billing issue into a functional and user-friendly tool. We successfully built a working web application that interprets medical bills, analyzes charges, and calculates estimated overpayments. We are especially proud of the automated letter generator, which provides users with a ready to send document tailored to their specific bill. We also demonstrated how AI can be combined with financial analysis to solve a real world consumer problem. Most importantly, we built a solution that has the potential to help people save significant amounts of money.

What we learned

Through building MediBill, we learned how inconsistent healthcare pricing systems are. We also learned that with data processing and AI we can produce more reliable and useful outputs than relying on AI alone. We gained experience in prompting and discovered how important it is to clearly define tone and constraints when generating formal documents. Additionally, we learned how to rapidly create and run a full AI application under time constraints.

What's next for MediBill

In the future, we plan to add support for PDF uploads and optical character recognition so users can scan physical bills directly. We also plan to make MediBill into an app that users can download on their personal devices to make it much easier and quicker to access.