Home
Open-source Python toolkit for Medicaid claims data analysis
medicaid-utils is a Python package for constructing patient-level analytic files from Medicaid claims data published by the Centers for Medicare & Medicaid Services (CMS). It implements validated cleaning routines, variable construction methods, and public-domain clinical algorithms for both MAX (Medicaid Analytic eXtract) and TAF (Transformed Medicaid Statistical Information System) file formats.
Built on Dask for scalable, distributed processing of large-scale claims datasets.
- Documentation: https://uc-cms.github.io/medicaid-utils/
- PyPI: https://pypi.org/project/medicaid-utils/
- Source: https://github.com/uc-cms/medicaid-utils
- Installation — How to install and set up medicaid-utils
- Data Layout — Required folder structure for MAX and TAF Parquet files
- Quick Start — Load claims, clean them, and extract a cohort in minutes
- Preprocessing — What cleaning and variable construction routines do
- Cohort Extraction — Build patient-level analytic files from diagnosis and procedure codes
- Risk Adjustment Algorithms — Elixhauser, CDPS-Rx, BETOS, PQI, NYU/Billings, PMCA, low-value care
- MAX vs TAF — Key differences between MAX and TAF file formats and how the package handles them
- Common Recipes — Frequently needed operations with code examples
- Scaling with Dask — Configuring Dask clusters for local, HPC, and cloud environments
- Working with Geographic Data — RUCA, RUCC, PCSA, and ZIP code crosswalks
- Glossary — CMS terminology, acronyms, and column name conventions
- Publications — Peer-reviewed papers built with medicaid-utils
- FAQ — Frequently asked questions
- Contributing — How to contribute to the project