Data disclaimer.
What's in our reports, what isn't, and how to read the numbers. Effective April 1, 2026.
Scope.
Pouls MedTech Insights reports and the live sample deliverables gallery are derived from publicly available CMS, HHS, and NPPES datasets. We do not handle Protected Health Information and we do not source proprietary or multi-payer claims data.
Vintage.
Utilization figures are from the most recent public release of the CMS Medicare Physician & Other Practitioners by Provider and Service file. CMS publishes a new vintage approximately twelve months after the close of each data year, and we update our deliverables as each new vintage lands. Facility metadata is from the contemporaneous CMS Hospital General Information release. Provider names and specialty are from the NPPES NPI registry. Procedure-code crosswalks come from the latest HCPCS catalog release.
What our scope excludes.
- Medicare Advantage utilization. CMS publishes Medicare Advantage data on a delayed and aggregated basis; it is not in the Part B fee-for-service file we use.
- Medicaid utilization. State Medicaid claims are not in the federal Part B file. State APCDs and the CMS T-MSIS are separate datasets and outside our scope.
- Commercial-payer utilization. Private-payer claims are proprietary; we do not redistribute them and we do not estimate them.
- The full multi-payer market. Combining Medicare fee-for-service, Medicare Advantage, Medicaid, and commercial would yield the full multi-payer view. Our reports cover Medicare fee-for-service only. The full multi-payer market is materially larger.
- Inpatient detail beyond the public file. Some MS-DRG and ICD-PCS detail in inpatient reports is sourced from the CMS Medicare Inpatient Hospitals public file.
How facility joins work.
Each facility row on our maps and in our reports is anchored to a CMS Certification Number (CCN). Provider-to-facility linkage uses the CMS Facility Affiliation file. A small fraction of providers do not have a published facility affiliation and are therefore not tied to a CCN; they are still counted at the state level.
How provider names work.
Provider names and specialty taxonomy come from NPPES, the federal NPI registry. Names are last-then-first as registered; we render them in the conventional first-last order on the maps. NPPES addresses are the provider's primary practice location of record and may be roughly twelve months stale. Specialty taxonomy is self-reported; some providers report multiple taxonomies and we use the primary.
How geocoding works.
Distance and catchment analyses use a public-domain ZIP centroid set. Five ZIP codes (territories and Puerto Rico) failed geocoding in the precompute step and are excluded from map distance calculations.
HAI and quality data.
When a report layers in CMS Healthcare-Associated Infections (HAI) data, we use the same public file CMS uses to calculate hospital-acquired-condition penalties. HAI metrics are reported with a 12–18 month lag and small-N facilities can be statistically noisy; we surface CMS confidence intervals where CMS publishes them.
What this disclaimer is not.
It is not legal or compliance advice. It is not a privacy notice under any health-information regulation. We do not handle Protected Health Information, so health-information privacy frameworks do not apply to our work, and we make no compliance claims either way. It is also not a representation that the underlying CMS files are themselves error-free; CMS occasionally republishes corrected vintages, and we update reports on request when that happens.