High-volume Medicare billers absent from NDH (directory completeness)
The NDH is meant to be the federal source of truth on provider identity. Material billers absent from NDH are a directory-side failure of the federal system, distinct from H10 (NPPES match rate). **Result: 2 of 1,259,343 Medicare Part B billing NPIs in CY 2023 (0.00016%) are absent from NDH (practitioner ∪ organization tables, 2026-05-08 release) AND billed ≥ \$10K in Medicare Part B.** Both are individual practitioners (entity_type=I) with material billing — \$764K and \$77K respectively. The 99.99984% NDH completeness rate against the Medicare Part B universe is a strong positive finding for the federal directory: when a provider bills Medicare materially, they are almost certainly in NDH.
Headline
2 of 1,259,343 Medicare Part B billing NPIs in CY 2023 (0.00%) are absent from NDH (practitioner ∪ organization tables, 2026-05-08 release) AND billed ≥ $10,000 in Medicare Part B. Combined paid amount: $840,826. The NDH is the federal source of truth on provider identity; material Medicare billers absent from NDH are a directory-side completeness gap distinct from H10's NPPES-match-rate signal. 0 of the material absences bill from a VA-state address. Entity-type breakdown in the material cohort: {'I': 2} (O=organization, I=individual).
2 / 1.3M = 0.00%
unit: count
Null hypothesis
The NDH is exhaustive — no NPI with material Medicare Part B billing in CY 2023 is missing from the pinned NDH bulk export.
Denominator
1,259,343 NPIs in the CMS Medicare Physician & Other Practitioners by Provider file for CY 2023. Numerator = NPIs absent from NDH (cms_npd.practitioner ∪ cms_npd.organization, 9,440,285 NPIs) AND paid amount ≥ \$10K in CY 2023.
Data source
CMS NDH bulk export (2026-05-08 release) × Medicare Physician & Other Practitioners by Provider (MUP_PHY_R25_P05_V20_D23_Prov.csv, CY 2023, ~270 MB, NPI-aggregated). NDH NPI set = UNION of practitioner._npi and organization._npi — joining against just the practitioner table would over-count organizations (labs, IDTFs) that ARE in NDH-organization. See `analysis/claims_sources/ndh_completeness.py`.
Notes
NDH NPI set = UNION of cms_npd.practitioner._npi and cms_npd.organization._npi. Part B billing NPIs can be either individuals (entity_type=1, NDH practitioner table) or organizations (entity_type=2, NDH organization table — big labs, IDTFs, group practices billing in aggregate). Joining against just practitioners would over-count absences with entity_type=2 NPIs that ARE in NDH-organization. The union join produces the honest completeness gap. Per-state CSV at /api/v1/states/va/h36-billers-absent-from-ndh.csv carries the VA subset for state PI triage; national rollup at /api/v1/findings/ndh-completeness-gap-detail.csv. Verify individual NPIs at the NPPES Registry to distinguish true directory-completeness gaps from out-of-scope billers.