Healthcare Price Advocate Agent

Gives users easy-to-understand estimates for hospital procedures, sees if the price on a bill is fair, and helps users negotiate their medical bills. Makes healthcare billing easy and accessible.

Comment

Inspiration

The US has the highest healthcare costs in the developed world, yet it has lower life expectancy than other developed nations. As someone whose family has been affected by high medical bills, I wouldn't wish it on my worst enemy. I want to show people truly how expensive and overcharged healthcare is in the country, and arm people with knowledge so they can stand up for themselves and fight back.

What it does

It makes healthcare billing data accessible and dynamic for users. If a user asks about a price, it tells users if they're being overcharged and what price to expect in their area. If a user asks what the expected price of a procedure is, the agent will tell users the range of prices in their area and how much they should expect to pay. If a user is paying too much for a bill, it generates a transcript so the user can contact the respective billing department and provide them with the necessary information to push for a lower bill.

How we built it

Used Codex and Antigravity, respectively, to build the MVP. It uses the following sponsored services:

  • Nimble: To scrape the hospital's pricing MRF data from online, based on the user's location. It also scrapes CPT information from the AMA
  • Clickhouse: Organizes the scraped data from Nimble so the agent can easily parse the data
  • Datadog: Observes and organizes the responses from the agent for optimization and improvement

Challenges we ran into

Parsing the MRF data from different hospitals was tricky due to inconsistent formatting, especially with the limited time. Also, ensuring the answers provided by the agent were accurate while being readable and accessible to users.

Accomplishments that we're proud of

Being able to pull pricing data and the acceptable range of rates, especially in a limited time, and making the information accessible to users.

What we learned

While there is healthcare price transparency data online, it's too technical and not accessible for users to easily access. Demand is there (and needed); when talking to other people at the hackathon, several people asked if they could personally use the product. When given structured data and information, the AI agent is good at parsing and analyzing the data. Otherwise, it can be hit or miss.

What's next for Healthcare Price Transparency Agent

Integrate insurance and payer information into the agent's knowledge, set it up so the agent can contact the respective billing department on behalf of the user, further UI/UX polish, and features

Built With

Share this project:

Updates