ClearView

ClearView lets patients search for a medical procedure, compare real costs across nearby hospitals, and estimate their out-of-pocket expenses — all powered by CMS Medicare data.

Comment

ClearView — Healthcare Cost Transparency

Know what you'll pay before you go.

Built at ViTAL Hacks 2026 under the Healthcare Cost Transparency track. ClearView lets patients search for a medical procedure, compare real costs across nearby hospitals, and estimate their out-of-pocket expenses — all powered by CMS Medicare data.

Features

  • Procedure search with live autocomplete from the CMS Medicare database
  • Hospital price comparison — sticker price, average total paid, Medicare portion, and negotiated savings side by side
  • Interactive map with color-coded pins by cost tier (MA hospitals)
  • Out-of-pocket estimator — calculates your estimated cost based on deductible and coinsurance
  • Sort by average paid, billed price, or hospital name
  • No fake data — every number comes directly from CMS Medicare Inpatient 2023

Tech Stack

Layer Technology
Frontend React + Vite + Leaflet (react-leaflet)
Backend Python 3.13, FastAPI, pandas, SQLite (FTS5)
Data CMS Medicare Inpatient Provider Service 2023
Geocoding Nominatim (OpenStreetMap)

Project Structure

ViTAL-Hacks/
├── backend/
│   ├── main.py              # FastAPI app — CORS, routes
│   ├── cms.py               # CSV ingestion, SQLite FTS5, search
│   ├── oop.py               # Out-of-pocket cost calculator
│   ├── geocode.py           # Address → lat/lng via Nominatim
│   ├── geo_cache.csv        # Pre-geocoded hospital coordinates (MA)
│   ├── cms.db               # Auto-generated SQLite database
│   ├── requirements.txt
│   └── scripts/
│       └── geocode_all.py   # One-time offline geocoding script
│
└── frontend/
    └── frontend/
        ├── src/
        │   └── App.jsx      # Full React app (single file)
        ├── package.json
        └── vite.config.js

Data Source

CMS Medicare Inpatient Provider Service Data — Release Year 2025, Calendar Year 2023

Downloaded from data.cms.gov. Public domain.

Key fields used:

Field Description
Rndrng_Prvdr_CCN Provider ID
Rndrng_Prvdr_Org_Name Hospital name
Rndrng_Prvdr_City/State/Zip5 Location
DRG_Cd / DRG_Desc Procedure code and description
Avg_Submtd_Cvrd_Chrg Average billed (sticker) price
Avg_Tot_Pymt_Amt Average total actually paid
Avg_Mdcr_Pymt_Amt Average Medicare portion paid

Prices are Medicare averages. Your actual cost will vary by insurance plan.

Setup

Prerequisites

  • Python 3.11+
  • Node.js 20+
  • CMS CSV file at /Users/zaidjilla/Downloads/MUP_INP_RY25_P03_V10_DY23_PrvSvc.csv (or set CMS_CSV_PATH environment variable)

Backend

cd backend
pip install -r requirements.txt
uvicorn main:app --reload --port 8000

The first startup ingests the CSV into cms.db (~1-2 seconds). Every subsequent restart is instant.

Check it's running:

curl http://127.0.0.1:8000/health
# {"status":"ok"}

Frontend

cd frontend/frontend
npm install
npm run dev
# → http://localhost:5173

API Reference

Base URL: http://localhost:8000

Endpoint Method Description
/health GET Health check
/search GET Search hospitals by procedure + location
/suggest GET Autocomplete DRG procedure names
/oop POST Calculate out-of-pocket cost
/plan-presets GET Insurance plan templates

GET /search

Param Type Description
q string Procedure name or DRG description
zip_code string 5-digit ZIP for radius filtering
state string 2-letter state abbreviation
radius float Search radius in miles (requires zip_code)
limit int Max results (default 100)

POST /oop

{
  "avg_submtd_cvrd_chrg": 85432.10,
  "avg_tot_pymt_amt": 15022.55,
  "deductible_remaining": 1000.00,
  "coinsurance_rate": 0.20,
  "oop_max_remaining": 5000.00,
  "use_negotiated_rate": true
}

Geocoding

Map pins are currently available for Massachusetts hospitals only. To add more states:

# Edit STATE_FILTER in backend/scripts/geocode_all.py
# Change: STATE_FILTER = "MA"
# To:     STATE_FILTER = "NY"  (or None for all states — takes hours)

cd backend
python3 scripts/geocode_all.py
rm cms.db
uvicorn main:app --reload --port 8000

Deployment

Frontend → Vercel

  1. Connect GitHub repo to vercel.com
  2. Set root directory to frontend/frontend
  3. Framework: Vite (auto-detected)
  4. Deploy

Backend → Render

  1. Connect GitHub repo to render.com
  2. Set root directory to backend
  3. Build command: pip install -r requirements.txt
  4. Start command: uvicorn main:app --host 0.0.0.0 --port $PORT
  5. Deploy

After deploying the backend, update API_BASE in src/App.jsx:

const API_BASE = "https://vital-hacks.onrender.com";

Team

Built at ViTAL Hacks 2026.

Role Work
Backend (Zaid Jilla, Roonak Thapa) FastAPI, data pipeline, SQLite, OOP calculator, geocoding
Frontend (Bivan Prajapati, Zaid Jilla) React UI, Leaflet map, search, cost comparison
Frontend (Bivan Prajapati, Roonak Thapa) Design, OOP estimator, component wiring

License

Built for hackathon purposes. CMS data is public domain.

Share this project:

Updates