Medicare Stars Measure Analytics Platform
End-to-end simulation of CMS Star Ratings calculation engine
Project Overview
This project simulates a complete Medicare Advantage Star Ratings calculation engine. It generates synthetic member, claims, and pharmacy data, processes it through a BigQuery pipeline to calculate HEDIS-like measures (e.g., Medication Adherence, Statin Use), and visualizes performance gaps in Looker Studio.
The Challenge
Healthcare payers struggle to calculate Star Ratings in real-time due to data latency and complex measure logic. The goal was to build a scalable, transparent calculation engine that allows for daily performance tracking.
Key Results
- Reduced measure calculation time from days to minutes using BigQuery optimization.
- Enabled drill-down capabilities from contract level to individual member gaps.
- Created a reusable framework for adding new measures with minimal code changes.
Technical Solution
- 1
Designed a Star Schema data model in BigQuery with Bronze (Raw), Silver (Clean), and Gold (Measure) layers.
- 2
Wrote complex SQL procedures to calculate Proportion of Days Covered (PDC) for medication adherence measures.
- 3
Implemented a Python-based synthetic data generator to create realistic healthcare datasets for testing.
- 4
Built an interactive Looker Studio dashboard to track measure performance against cut points.