The Problem With the NPI Registry #
Navigating healthcare is complex enough without incorrect, outdated provider information adding frustration to the mix. Yet, the primary national database of U.S. healthcare providers—the NPI (National Provider Identifier) Registry—is notoriously messy, outdated, and hard to use. It’s time we created an open-source solution.
While the NPI registry, maintained by the U.S. government’s CMS, is incredibly comprehensive (with millions of entries), its accuracy leaves much to be desired:
- Providers frequently change practices without updating their records
- Location and contact information often becomes outdated
- Duplicate and incomplete records are rampant
- No effective mechanism exists for reporting or correcting errors
This outdated data causes headaches for patients, providers, and healthcare innovators alike. Patients struggle to find in-network doctors, providers miss referrals, and developers building healthcare tools face significant data quality challenges.
Why We Need an Open-Source Solution #
An open-source NPI mapper would allow communities—patients, providers, tech developers, and civic-minded citizens—to collaboratively enhance provider data accuracy. Imagine:
- Matching NPI records directly to real-world Google Maps locations
- Allowing contributors to correct or suggest updates easily
- Keeping a transparent audit trail of changes and validations
- Building machine learning models to identify and flag potential data inconsistencies
By distributing the work of maintaining accurate provider information across a community of stakeholders, we can create a resource that benefits the entire healthcare ecosystem.
How the Open NPI Mapper Could Work #
The system could be simple, secure, and scalable:
1. Canonical Dataset #
- Store a regularly updated canonical file (like Parquet format) that represents the trusted, verified set of data
- Each month, CMS releases a “diff” CSV showing changes—this could automatically integrate into the canonical dataset
- Maintain historical versions to track changes over time and allow for rollbacks if needed
2. Crowdsourced Contributions #
- Users authenticate via GitHub OAuth to provide accountability and transparency
- Community edits get stored separately, reviewed by maintainers, and merged into the canonical dataset upon approval
- Implement a reputation system for contributors, allowing trusted users more latitude in making changes
3. Versioning and Auditability #
- Monthly snapshots would be stored using GitHub Releases—lightweight, easily accessible, and free
- Every change and its contributor would be clearly tracked, providing full transparency
- Regular data quality reports would highlight areas needing attention
4. Matching NPI to Google Maps #
- Contributors match NPI business names to Google Maps Place IDs
- This doesn’t violate Google Maps terms since you’re linking identifiers rather than caching proprietary information
- Additional data sources like provider websites and phone directories could further enhance accuracy
Technical Implementation #
A basic technical approach could look like:
- A Next.js or Flask web interface to collect contributions, authenticated via GitHub
- JSON-formatted contributions stored in Git, processed monthly into the canonical dataset
- Automated scripts and GitHub Actions to manage updates, diffs, and versioning
- API endpoints allowing developers to query the enhanced dataset for their applications
The Road Ahead #
An open-source NPI mapper can become the healthcare data backbone we’ve been missing—helping patients access accurate provider information, aiding healthcare innovators, and empowering community collaboration.
At Care Carta, we’re exploring how such a solution could dramatically improve the healthcare experience for all stakeholders. The technology exists today—what’s needed is a coordinated effort to implement it.
Interested in contributing, collaborating, or discussing ideas? Let’s connect and make healthcare data open, accurate, and useful for everyone.
Colin Atkinson is a medical coding expert and co-founder of Care Carta, with over a decade of experience in hospital administration systems.