Methods & data vintages
Methods appendix · build v34-20260928T202228Z
A one-page, citable record of what these numbers are, how they are attributed, and which CMS and CDC files they were built from.
What these numbers are
- What these numbers are: exact Medicare fee-for-service counts from CMS public files — disclosed, auditable, never modeled. On-site market totals anchor on a labeled all-Medicare estimate scaled from this observed foundation; exported rows are always the observed counts. Part D prescribing counts are drug-plan claims and include Medicare Advantage plans.
- What they are not: all-payer market estimates. The all-Medicare estimate is Medicare only — never commercial, Medicaid, or cash-pay. If you use modeled all-payer tools, expect their totals to run higher; the named providers and the rank order are typically consistent. Figures based on fewer than 11 beneficiaries are suppressed by CMS at the source and excluded.
How volume is attributed (D-062)
Group aggregates sum single-membership clinicians only. A clinician who appears on more than one group's roster is excluded from every group's totals and shown as a dash ("—") — volume that cannot be unambiguously attributed to one group is never guessed or split. Each such clinician is still counted exactly once in the market (state) totals. Shares are computed over this attributable volume and can therefore sum to less than 100% of a market.
Markets are billing-state markets (D-065)
A group's market is the state its claims are billed from, not the state where it is registered. A group appears in every state its attributable members bill in, carrying that state's volume; the registered location rides along for display only. Shares are ≤ 100% by construction — numerator and denominator share the same billing-state basis.
Small-count suppression (the <11 floor)
CMS suppresses provider-by-code figures below 11 beneficiaries at the source; suppressed volume is excluded and cannot be estimated from this file. Nevvi applies the same 11-unit floor to its own city- and specialty-market benchmarks. Absence is suppression, never a zero.
Payer-mix and the MA-scaled estimate
Observed procedure and hospital figures count Medicare fee-for-service claims only — not Medicare Advantage, commercial, Medicaid, or cash-pay — and all-Medicare figures are labeled estimates scaled from them, as described below. Prescribing counts are the exception: they are Medicare Part D drug-plan claims, which include Medicare Advantage plans and cover no commercial or cash purchases. The payer-mix strapline states how much of Medicare in a state is fee-for-service so the coverage of each number is explicit. The estimated all-Medicare volume (FFS + estimated Medicare Advantage) divides each state's FFS total by that state's FFS share for the claim year queried, falling back to the nearest earlier published vintage when that year's enrollment is not yet published — the share year divided by is disclosed beside the estimate. Scaling is per state before summing; the result is shown as a service count only (payments are not scaled, because Medicare Advantage is capitated) and is labeled scaled estimate — assumes MA utilization mirrors FFS; not an observation on every number, never presented as an observed count.
One exception governs the national figure. Some codes book most of their national volume to a single state under the billing-state basis — central-lab tests billed from one facility (the motivating case), specialty- pharmacy drugs, and thinly billed codes whose volume concentrates by small numbers alone. Scaling such a market per state before summing would let that one state's fee-for-service share set the whole national estimate. When a code's national volume concentrates in a single billing state (roughly half or more, with a stability band so a borderline code does not switch methods release to release), the national figure is instead scaled by the blended national fee-for-service share, and the estimate carries the line billing for this code concentrates in one state — scaled by the national payer mix. This is the only place a blended national share is used; every state figure, and every other national figure, still scales per state before summing.
Disease-burden context
CDC BRFSS prevalence is survey data covering all-payer adults 18 and older; Nevvi utilization counts Medicare fee-for-service only, largely 65 and older. The two appear side by side as context — Nevvi never combines them into a score, rating, or ranking.
Sources & data vintages
| Source | Vintage | Loaded |
|---|---|---|
| asc_fs | 2026q3 | 2026-09-05 |
| cdc_brfss_state_adj | 2024 | 2026-07-23 |
| cdc_brfss_state_crude | 2024 | 2026-07-23 |
| cdc_places_county | 2025 | 2026-07-05 |
| cms_inpatient | FY2021_2023-09-05 | 2026-08-31 |
| cms_inpatient_geo | FY2023_2025-05-22 | 2026-08-31 |
| cms_msdrg_table5 | FY2027 | 2026-09-18 |
| cms_outpatient_hosp | DY2023_2025-08-22 | 2026-08-31 |
| cms_partb | CY2024_2026-05-21 | 2026-07-09 |
| cms_partb_provider | CY2024_2026-05-21 | 2026-08-19 |
| cms_partd | CY2024_2026-05-21 | 2026-08-31 |
| cms_partd_display | CY2024_2026-05-21 | 2026-09-24 |
| cms_partd_geo | CY2024_2026-05-21 | 2026-08-31 |
| cms_rbcs | 2025 | 2026-07-10 |
| cpsc | 2026-08 | 2026-08-31 |
| dac | 2026-07-31 | 2026-09-03 |
| dac_affiliations | 2026-07-31 | 2026-09-03 |
| dac_snapshot | CY2026 | 2026-07-04 |
| hospital_general_info | 2026-07-22 | 2026-09-03 |
| ma_penetration | 2026-03_2026-06-23 | 2026-07-18 |
| mcd_article_codes | 2026-07-23 | 2026-07-28 |
| mcd_article_contractors | 2026-07-23 | 2026-07-28 |
| mcd_articles | 2026-07-23 | 2026-07-28 |
| mcd_contractor_jurisdiction | 2026-07-23 | 2026-07-28 |
| mcd_contractors | 2026-07-23 | 2026-07-28 |
| mcd_lcd_codes | 2026-07-23 | 2026-07-28 |
| mcd_lcd_contractors | 2026-07-23 | 2026-07-28 |
| mcd_lcds | 2026-07-23 | 2026-07-28 |
| mcd_state_lookup | 2026-07-23 | 2026-07-28 |
| medicare_enrollment | 2026-03_2026-06-23 | 2026-07-04 |
| mmd_prevalence | cy2023p | 2026-07-18 |
| nppes_providers | 2026-08 | 2026-08-15 |
| pfs_gpci | 2026a | 2026-09-19 |
| pfs_rvu | 2026a | 2026-09-19 |
Current build covers 69,855 billing groups, 1,528,560 clinicians, 6,260 procedure codes across 63 states, through CY2024.
Citing this appendix
Please cite as: "Nevvi (nevvi.app) methods appendix, build v34-20260928T202228Z, from the CMS Medicare Physician & Other Practitioners PUF and CDC PLACES / BRFSS." The build identifier pins the exact data vintage behind any figure you quote. Full method narrative: nevvi.app/methods.
Rights & attribution
An independent product, not affiliated with, or endorsed by, CMS or HHS. The five-digit numeric codes are CPT®, maintained by the American Medical Association; their descriptions reach Nevvi inside CMS's public files and remain the AMA's work. CPT copyright 2025 American Medical Association. All rights reserved. Fee schedules, relative value units, conversion factors and/or related components are not assigned by the AMA, are not part of CPT, and the AMA is not recommending their use. The AMA does not directly or indirectly practice medicine or dispense medical services. The AMA assumes no liability for data contained or not contained herein. CPT is a registered trademark of the American Medical Association.