Technical
Methodology
A transparent look at how we process 2 million+ records into helpful, comprehensive healthcare insights.
Protocol Version 2.1
"Transparency is the cornerstone of trust in healthcare. We publish our methodology to ensure patients, providers, and regulators understand the logic behind every data point on our platform."
01. Data Sourcing Pipeline
Our primary data source is the National Plan and Provider Enumeration System (NPPES), managed by the Centers for Medicare & Medicaid Services (CMS). This is the only "source of truth" for provider identity in the U.S.
Weekly Synchronization
We perform a full delta-sync every 7 days. This captures new NPI enumerations, practice location updates, and deactivation events in near real-time.
Multi-Source Validation
Beyond basic NPI data, we cross-reference the OIG LEIE (List of Excluded Individuals/Entities) to ensure patient safety signals are accurate.
02. Atomic Content Algorithm
To solve the "Helpful Content" challenge at scale, we developed the Atomic Content System. This is not generative AI, but a deterministic grammatical engine that ensures every provider profile provides a unique, human-readable summary.
How it works:
- A
Deterministic Hashing
We use the provider's NPI as a seed for a hashing algorithm. This ensures that a specific NPI always produces the same text, satisfying Google's requirement for stable, indexed content.
- B
Combinatorial Diversity
Content is assembled from four "slots" (Hook, Core, Value, Trust). With dozens of linguistic variations per slot, the system generates over 17,000 unique semantic paths.
- C
Semantic Logic
The engine uses conditional branching based on real clinical data (e.g., Early Career vs. Established, High Medicare Volume vs. Private Practice) to ensure the summary is contextually accurate.
03. Quality Benchmarking
Raw numbers (like a MIPS score of 75) mean very little to a patient without context. Our methodology focuses on Relative Performance within a provider's specific geography.
State-Level Percentile Calculation:
Percentile = (Count of Peers with Lower Score / Total Peers in State) * 100By comparing a provider only against peers in the same state, we account for regional variations in healthcare delivery and reporting standards. This allows us to issue labels like "Top 10% in State" or "Above Average," which are far more helpful for patient decision-making than raw clinical scores.
04. Objectivity & Integrity
Unlike many "Top Doctor" lists, NPI Telehealth has zero pay-to-play involvement.
- Rankings are purely based on clinical relevance, proximity, and quality data.
- Advertising does not influence the presentation of federal credentials or Medicare statistics.
- Patient reviews are screened for fraud but are never deleted at a provider's request unless they violate terms of service.
Open Data Policy
We believe in the power of open health data. If you have questions about our calculations or wish to report a discrepancy in our percentile logic, please reach out to our Data Integrity Team.