Editorial Policy
Last updated: June 2026
CareCost Explorer publishes negotiated-rate data derived from federally-mandated Machine-Readable Files. This page explains how we source and verify what we publish, how AI is and is not used, our independence from commercial interests, and how we handle errors.
1. Source hierarchy
- Federal Transparency in Coverage Machine-Readable Files (45 CFR Part 147) — payer-published in-network rate files. Primary source for all negotiated-rate data.
- CMS administrative files — HCPCS Level II and CPT code descriptions, CMS ASP payment limits, NPPES provider registry. Used for code labels, benchmark comparisons, and provider identification.
- Our normalization layer — CareCost's pipeline maps each payer's MRF schema to a common format (payer, provider, billing code, negotiated rate, rate type, geography). The methodology is published at /methodology.
No figure in CareCost Explorer is modeled, estimated, or derived from third-party benchmarks. Every rate traces back to a payer-published source file.
2. AI use
Artificial intelligence is used in engineering (code generation, pipeline development, infrastructure tooling) and in data analysis (quality scoring, anomaly detection, normalization assistance). AI tools do not author public-facing content autonomously. All data figures, editorial claims, and public copy are reviewed and approved by a human before publication. We do not use AI to generate synthetic rate data or to fill gaps in the source corpus.
3. Conflict of interest
CareCost Explorer accepts no money from payers, manufacturers, or pharmacy benefit managers. Revenue comes from practices, researchers, and organizations that purchase access to the product. We have no financial relationship with any insurer whose rates appear in the product, and no financial relationship with any pharmaceutical manufacturer or PBM. This independence is non-negotiable: the numbers answer only to the source files.
4. Corrections policy
We publish our errors. When a figure in CareCost Explorer is found to be incorrect — whether due to a source-file error, a pipeline bug, or a documentation mistake — we correct it, note it on our corrections log, and where material, notify affected users. To report a suspected error, email hello@carecostexplorer.com with the billing code, payer, and the figure you believe is wrong. We confirm receipt within two business days and resolve confirmed corrections within five business days.
See the public corrections log.
Questions about our editorial standards? Email hello@carecostexplorer.com. For data questions and corrections, see the corrections log. For methodology details, see /methodology.