Inspiration
You know how when you get a prescription, you go to the pharmacy, hand over your insurance card, and the pharmacist can instantly tell you if your drug is covered and what it’ll cost? That’s the pharmacy benefit, it’s standardized, electronic, and pretty well organized. Now imagine a completely different world. Your doctor says you need a medication that has to be injected or infused in a clinic, something like a cancer drug, an immunotherapy, or a biologic for rheumatoid arthritis. These drugs don’t go through the pharmacy at all. They’re covered under your medical benefit, the same part of your insurance that covers surgeries and doctor visits. And here’s the problem: there’s no clean formulary list for these drugs. Instead, every health plan writes its own medical policy , a lengthy document that spells out whether the drug is covered, what diagnosis you need, what other drugs you must have tried first, whether prior authorization is required, and where you’re allowed to receive the infusion. These policies are different at every health plan. They’re published in different formats. They change constantly. And there’s no single place to look them all up and compare them. So our project helps such patients view and understand their policies and benefits.
What it does
We have built a visualization dashboard enabling users to see details like number of policies, each policy details, the prior authorizations, etc. such important information. It also has a chat bot enabling patients to interact with the AI agent so that they can ask their questions helpful for those users who are unable to understand the visualizations.
How we built it
We have used RAG for the chatbot and dashboard visualizations, web crawling and information extraction for training the RAG model and some use of LLM for pdf parsing in case the pdf is made of images. We have created the dashboard and the whole UI using React.
Log in or sign up for Devpost to join the conversation.