Since CMS issued the Transparency in Coverage (TiC) final rule, many organizations have focused on solving the engineering challenge of downloading and parsing the massive volume of machine-readable files (MRFs) published by health insurers. However, like most raw materials, raw data must be transformed to be valuable. Without an accurate provider directory and claims-based utilization data, it is impossible to use the MRFs to answer the fundamental question: Who is being paid what, for which services, where and how often?
An accurate provider directory is essential to interpret the files. There are numerous limitations to the raw MRFs, the most important of which are these:
On a standalone basis, MRFs can be informative, revealing the minimum and maximum negotiated rate that an insurer pays. Transforming the health plan price transparency (HPPT) data into actionable insights requires enriching that raw data with provider and utilization data.
| Challenge | Alternative Approaches | Our Approach |
| Massive file sizes with disorganized data | Partial file downloads, or relying on smaller files posted by hospitals | Data infrastructure to ingest and standardize terabytes of TiC files every month |
| Missing provider data | Dependence on outdated NPPES data | Linked to our proprietary provider directory for real-time specialties, locations and practice status |
| Missing organizational hierarchies | No link between NPIs and parent organizations | Hierarchical mapping of parent-child relationships for organization analyses |
| Zombie rates | Heuristic-based filters remove rates for services a provider would not reasonably render | Claims-based filtering to identify the rates for services that providers have actually billed |
| No locations | Oversimplified and often-outdated primary practice locations from NPPES | Claims-enriched location data for market-level analysis |
| No utilization context | Rate comparisons in isolation | Analysis of rates and volumes to show revenue and market share |
To address data limitations, our methodology is based upon six core principles:
One month of MRFs from the “BUCA” plans – Blue Cross Blue Shield, UnitedHealthcare (UHC), Cigna and Aetna – totals over 100TB even when compressed, roughly equivalent to two-thirds the amount of data collected by the Hubble Telescope over the last 28 years. These MRFs are also structured differently, even within the CMS schema: some payers post thousands of files with tens of millions of data points each, while others post millions of small files with just hundreds of entries. The size and variation make accessing and managing the MRFs a formidable technical challenge.
Our team of data engineers built a pipeline designed to handle petabytes of TiC data, with built-in de-duplication, systematic storage and metadata mapping across all major plans. This has allowed us to create an organized dataset that supports efficient, accurate comparisons across plans, providers and markets.
Trying to extract insight from HPPT files without a provider directory is like trying to analyze an outdated, inaccurate roster with no names, just numbers that are disconnected, ambiguous and often wrong.
The raw MRFs do not distinguish clearly between Type 1 (Individual) and Type 2 (Organization) NPIs. Providers are frequently misclassified or inactive. Even if analysts can overcome the technical hurdles of accessing the data, they are left guessing which providers are represented in the files – and which of those providers actually deliver those services in their markets.
We resolve this ambiguity by linking the TiC files to our proprietary provider directory, which allows us to classify NPIs based on recent billing activity. We can tie rates to specific providers for specific procedures, which enables precise, defensible analysis of reimbursement patterns and provider performance.
This sample data demonstrates how we contextualize NPIs from the payer MRFs with our provider directory data:
| TiC File | Trilliant Health Provider Directory Data |
||
| NPI | NPI Type | Name | Classification |
| 1265524847 | Individual | Christina P Hitchcock MD | Obstetrics & Gynecology |
| 1801184296 | Individual | Saroj Neupane MD | Cardiovascular Disease |
| 1396882205 | Organization | Vanderbilt University Medical Center | Short Term Acute Care Hospital |
| 1134262868 | Organization | North Shore Surgical Center | Surgery Center |
The TiC files list rates at the NPI or EIN level but do not explain how those identifiers connect to a broader organization. Because a single hospital may bill under dozens of NPIs across sites, different payers often use different NPIs to represent the same provider organization. For example, Anthem may use one NPI for an organization while UHC uses another, which makes it difficult to make meaningful comparisons across payers.
We use our provider directory to map parent-child relationships between NPIs. We group billing entities under their parent organization, so we can identify the provider regardless of which EIN or NPI is used across the payer files. Without a provider directory to map these relationships, it is impossible to make meaningful comparisons of negotiated rates across provider organizations. Assigning organization hierarchies makes these comparisons possible, transforming a flat list of rates into a structured view that can inform pricing and competitive strategies.
This view of a health system’s various NPIs illustrates how our provider directory represents parent-child relationships:
| Parent Organization | Organizations | Select NPIs |
| Vanderbilt Health | Vanderbilt University Medical Center | 1396882205 1427447697 |
| Vanderbilt Medical Group | 1235154972 1013932946 1679614812 |
|
| Vanderbilt Wilson County Hospital | 1306889597 1215979190 |
|
| Vanderbilt Imaging Services Hillsboro | 1144241985 | |
| Vanderbilt Pediatric Associates | 1356936744 |
MRFs are littered with zombie rates – contracted amounts for procedures a provider would never render. For example, a 2025 UHC MRF includes what they would pay a physical therapy group in Washington if they were to perform colonoscopies or start taking psychotherapy appointments, services they do not – and likely will never – perform:
| Organization, Name, NPI and Code | Code and Description | UHC Negotiated Professional Rate | Rendered Service? |
| Olympic Sports & Spine; 1033622311; Physical Therapy |
CPT 97161 - Physical therapy evaluation: low complexity | $187.40 | Yes |
| CPT 98941 - Chiropractic manipulative treatment (CMT); spinal, 3-4 regions | $88.70 | Yes | |
| CPT 45385 - Colonoscopy, flexible; with removal of tumor(s), polyp(s), or other lesion(s) by snare technique | $977 | No | |
| CPT 90834 - Psychotherapy, 45 minutes with patient | $202 | No |
To identify zombie rates, we identify active providers and the services they actually bill, based on recent activities in our all-payer claims database representing more than 300 million American lives. We can flag inactive providers and identify each provider’s practicing specialty based on their practice patterns. With this context, we eliminate irrelevant rates from our HPPT data, including only the rates for procedures that providers actually perform.
The MRFs do not contain any geographic data, but most practical applications of the data are executed at the local market level. Strategic use cases, like negotiating contracts, typically require knowing not only who was reimbursed and for which services – but also where they practice.
We resolve this gap by identifying where care was rendered at the site of service level, enabling visibility to rates across facilities, markets and regions.
This example shows providers whose NPPES-listed state does not match the states where they actually practice:
| Provider Name | Specialty | NPPES Primary Practice Address |
Trilliant Health Facility Address |
| Brendan J Cavanaugh, MD; 1013947639 | Cardiovascular Disease |
502 Elm St. NE Albuquerque, NM 87102 | 5177 McCarty Ln Lafayette, IN 47905 |
| Irakli Todua, MD; 1003301144 | Internal Medicine |
901 Heartland Rd. Ste 3800 Saint Joseph, MO 64506 | 260 Tremont St Boston, MA 02116 |
An unlabeled list of negotiated rates offers limited strategic value. What matters most is revenue, which is the product of negotiated rate and service volume. A provider reimbursed at a lower rate may still generate significantly more revenue than a peer with a higher rate, simply by performing more services. Without service volume, price benchmarking can be misleading, making an organization with high rates but low volume appear more financially significant than a competitor with lower rates and greater market share.
We combine utilization data with the TiC files to calculate provider- and system-level revenue across service lines. This approach transforms raw price benchmarks into actionable information that can be used to assess market share, inform pricing strategy and improve financial performance.
Here is an example of the negotiated rate and market share for MS-DRG 470 at two short-term acute care hospitals near Seattle, WA:
| Organization Name | Aetna Negotiated Rate | Market Share |
| Evergreen Health Medical Center | $40,010 | 11.7% |
| St. Anthony Hospital | $49,296 | 1.7% |
Without market share context, St. Anthony Hospital’s higher negotiated rate may appear more favorable. However, utilization data reveals that Evergreen Health Medical Center, with a lower rate but higher market share, generates more revenue.
Trilliant Health’s approach to processing and contextualizing the HPPT data unlocks actionable information about negotiated rates that are not accessible from public MRFs:
Trilliant Health’s enhanced methodology equips stakeholders to understand not just what providers could be paid, but what they are paid – by whom, for what, where and at what volume. That clarity is essential in a health economy increasingly defined by transparency, accountability and value.