{ "study": { "slug": "excluded-providers-still-enrolled-2026", "title": "Barred but billable: excluded providers still enrolled in Medicare", "standfirst": "19 providers barred from all federal health programs by the OIG still hold an active Medicare enrollment record in PECOS — out of 6,880 in-force NPI-identified federal exclusions. Most trace to a single refresh cycle's lag, but two have stood for over a year, one excluded since 2015.", "desk": "financial-distress", "article_type": "Original Research", "published": "2026-06-14", "issue": 55, "doi": "10.5072/fonteum/excluded-providers-still-enrolled-2026", "url": "https://fonteum.com/research/excluded-providers-still-enrolled-2026", "methodology_version": "excluded-enrollment/v1" }, "data_as_of": "2026-06-14", "datasets": [ { "slug": "oig-leie", "name": "OIG LEIE", "publisher": "HHS OIG — List of Excluded Individuals/Entities", "upstream_url": null }, { "slug": "cms-pecos", "name": "CMS PECOS", "publisher": "CMS — Provider Enrollment, Chain & Ownership System", "upstream_url": null } ], "key_findings": [ { "number": "19", "finding": "providers on the in-force federal OIG exclusion list still carry an active Medicare enrollment record in PECOS — matched on NPI, never on name", "dataset": "oig-leie" }, { "number": "6,861", "finding": "of 6,880 in-force NPI-identified federal exclusions correctly carry no active Medicare enrollment record — the screening machinery mostly works", "dataset": "oig-leie" }, { "number": "8", "finding": "of the 19 sit on mandatory §1128(a) conviction-based exclusion bases — the most serious category, not administrative license actions", "dataset": "oig-leie" }, { "number": "2", "finding": "have stood for more than a year — one excluded since 2015 — too old to be explained by a refresh-cycle lag", "dataset": "cms-pecos" } ], "faqs": [ { "q": "How many excluded providers are still enrolled in Medicare?", "a": "Nineteen. Of 6,880 in-force, NPI-identified providers on the federal OIG List of Excluded Individuals and Entities, 19 still carry an active Medicare enrollment record in the CMS PECOS Ordering & Referring file. The match is made on the National Provider Identifier only, never on a name." }, { "q": "What does it mean for a provider to be on the OIG LEIE but still in PECOS?", "a": "An OIG exclusion bars a provider from billing — directly or indirectly — any federal health program. A live PECOS enrollment record means that provider still holds an approved Medicare enrollment eligible to order and refer. Holding both at once is the compliance contradiction this study counts: barred on one list, billable on the other." }, { "q": "Are these 19 a fraud problem or a data-lag artifact?", "a": "Mostly lag. 17 of the 19 were excluded within the last 90 days, consistent with CMS revocation trailing the OIG's 2026-05-20 exclusion release. But two have stood for more than a year — the oldest excluded in 2015, roughly eleven years ago. Those two cannot be explained by refresh timing." }, { "q": "Why is the number so small?", "a": "Because the screening mostly works: 6,861 of 6,880 in-force NPI-identified exclusions correctly carry no active enrollment record. The finding is not that the system is broken — it is that the residual is not zero, and each non-zero case is a 'knew or should have known' civil-monetary-penalty exposure for any organization that bills on that enrollment." }, { "q": "What about exclusions with no NPI?", "a": "61,030 of the 68,055 records on the federal LEIE carry no NPI at all. They cannot be matched to a PECOS enrollment record by identifier, so NPI-based enrollment screening cannot reach them. They are excluded from the matchable denominator and reported separately rather than guessed at by name." }, { "q": "What counts as an active enrollment record here?", "a": "Presence in the CMS PECOS Ordering & Referring extract, federal release 2026-05-28. Every row in that file is an approved Medicare enrollment record eligible to order and refer. A single NPI can hold more than one record across provider types or states, so enrollment records (20) slightly exceed distinct providers (19)." }, { "q": "Can I reproduce these numbers?", "a": "Yes. Every figure is a direct NPI equijoin between the public oig_leie_exclusions and pecos_providers tables. The exact SQL is published in the reproducibility block below; each count resolves to specific rows in specific frozen snapshots, and no match is ever inferred from a name." } ], "citation": { "apa": "Fonteum Research. (2026, June 14). Barred but billable: excluded providers still enrolled in Medicare. Fonteum Research, Issue 55. https://doi.org/10.5072/fonteum/excluded-providers-still-enrolled-2026", "url": "https://fonteum.com/research/excluded-providers-still-enrolled-2026" }, "reproducible_sql": "-- Barred but billable: excluded providers still enrolled in Medicare —\n-- fully reproducible query.\n--\n-- Question: how many providers the OIG has EXCLUDED from all federal health\n-- programs still carry an ACTIVE Medicare enrollment record in PECOS?\n--\n-- Sources:\n-- public.oig_leie_exclusions — OIG List of Excluded Individuals and Entities,\n-- federal monthly bulk download, release\n-- 2026-05-08 (latest exclusion 2026-05-20),\n-- 68,055 active records (7,025 carry an NPI).\n-- public, read-only.\n-- public.pecos_providers — CMS PECOS Ordering & Referring / provider-\n-- enrollment extract, federal release 2026-05-28,\n-- 2,981,799 enrollment records / 2,556,656\n-- distinct NPIs. Presence in this file is an\n-- active, approved Medicare enrollment record\n-- eligible to order and refer. public, read-only.\n--\n-- Join key: NPI only (10-digit, btrim). We never match on name — a name match\n-- is not a defensible identity assertion, so LEIE rows with no NPI are excluded\n-- from the matchable denominator and reported separately (see no-NPI query).\n--\n-- \"In force\" mirrors the production exclusion lookup (src/lib/exclusions): a row\n-- is in force when reinstatement_date IS NULL OR reinstatement_date > today.\n-- The OIG drops reinstated parties from the published LEIE, so this set equals\n-- \"any LEIE row by NPI\". Date basis: current_date (2026-06-14 at publish).\n--\n-- Every headline figure in the study resolves to one of the rows below.\n\nWITH leie_inforce AS ( -- distinct in-force, NPI-identified federal exclusions\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.oig_leie_exclusions\n WHERE nullif(btrim(npi), '') IS NOT NULL\n AND (reinstatement_date IS NULL OR reinstatement_date > current_date)\n),\npecos_npi AS ( -- distinct NPIs with an active PECOS enrollment record\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.pecos_providers\n WHERE nullif(btrim(npi), '') IS NOT NULL\n)\nSELECT\n (SELECT count(*) FROM leie_inforce) AS leie_inforce_npi, -- 6,880\n (SELECT count(*) FROM pecos_npi) AS pecos_distinct_npi, -- 2,556,656\n (SELECT count(*) FROM leie_inforce l JOIN pecos_npi p USING (npi)) AS excluded_and_enrolled, -- 19\n (SELECT count(*) FROM leie_inforce l\n LEFT JOIN pecos_npi p USING (npi) WHERE p.npi IS NULL) AS excluded_not_enrolled, -- 6,861\n (SELECT count(*) FROM public.oig_leie_exclusions\n WHERE nullif(btrim(npi), '') IS NULL) AS leie_rows_no_npi, -- 61,030\n (SELECT count(*) FROM public.oig_leie_exclusions) AS leie_total_rows; -- 68,055\n-- leie_inforce_npi pecos_distinct_npi excluded_and_enrolled excluded_not_enrolled leie_rows_no_npi leie_total_rows\n-- 6,880 2,556,656 19 6,861 61,030 68,055\n-- => 19 / 6,880 = 0.28% of in-force NPI-identified federal exclusions still hold an active enrollment record.\n\n-- Enrollment records vs distinct providers among the matched set (one NPI can\n-- hold multiple enrollment records; one match is an organizational record):\nWITH leie_inforce AS (\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.oig_leie_exclusions\n WHERE nullif(btrim(npi), '') IS NOT NULL\n AND (reinstatement_date IS NULL OR reinstatement_date > current_date)\n),\nmatched AS (\n SELECT p.npi, p.state_cd, p.provider_type_desc, p.org_name, p.last_name\n FROM public.pecos_providers p\n JOIN leie_inforce l ON l.npi = nullif(btrim(p.npi), '')\n)\nSELECT\n count(*) AS enrollment_records, -- 20\n count(DISTINCT npi) AS distinct_providers, -- 19\n count(*) FILTER (WHERE last_name IS NOT NULL AND last_name <> '') AS individual_records, -- 19\n count(*) FILTER (WHERE org_name IS NOT NULL AND org_name <> '') AS org_records -- 1\nFROM matched;\n\n-- Excluded-yet-enrolled providers by active-enrollment state (distinct NPI):\nWITH leie_inforce AS (\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.oig_leie_exclusions\n WHERE nullif(btrim(npi), '') IS NOT NULL\n AND (reinstatement_date IS NULL OR reinstatement_date > current_date)\n)\nSELECT p.state_cd AS state,\n count(DISTINCT p.npi) AS providers\nFROM public.pecos_providers p\nJOIN leie_inforce l ON l.npi = nullif(btrim(p.npi), '')\nGROUP BY p.state_cd\nORDER BY providers DESC, state;\n-- NY 3 · CA 2 · FL 2 · NC 2 · PA 2 · TX 2 · AR 1 · CO 1 · MI 1 · MT 1 · NV 1 · OH 1 · OK 1 (13 states)\n\n-- Excluded-yet-enrolled providers by enrolled specialty (distinct NPI):\nWITH leie_inforce AS (\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.oig_leie_exclusions\n WHERE nullif(btrim(npi), '') IS NOT NULL\n AND (reinstatement_date IS NULL OR reinstatement_date > current_date)\n)\nSELECT p.provider_type_desc AS specialty,\n count(DISTINCT p.npi) AS providers\nFROM public.pecos_providers p\nJOIN leie_inforce l ON l.npi = nullif(btrim(p.npi), '')\nGROUP BY p.provider_type_desc\nORDER BY providers DESC, specialty;\n-- FAMILY PRACTICE 4 · CLINICAL SOCIAL WORKER 2 · INTERNAL MEDICINE 2 · then 11 specialties at 1 each\n-- (SNF, Anesthesiology, Cardiology, CNS, Clinical Psychologist, Hematology/Oncology, Medical\n-- Oncology, Mental Health Counselor, PT in private practice, Podiatry, Vascular Surgery) = 14 specialties.\n\n-- Statutory basis of the matched exclusions (mandatory §1128(a) vs permissive §1128(b))\n-- and the exclusion-date span — separates one-cycle refresh lag from standing failure:\nWITH pecos_npi AS (\n SELECT DISTINCT nullif(btrim(npi), '') AS npi\n FROM public.pecos_providers\n WHERE nullif(btrim(npi), '') IS NOT NULL\n),\nper_npi AS (\n SELECT e.npi,\n min(e.excl_date) AS first_excl,\n bool_or(trim(e.exclusion_type) LIKE '1128a%') AS any_mandatory\n FROM public.oig_leie_exclusions e\n JOIN pecos_npi p ON p.npi = nullif(btrim(e.npi), '')\n WHERE (e.reinstatement_date IS NULL OR e.reinstatement_date > current_date)\n AND nullif(btrim(e.npi), '') IS NOT NULL\n GROUP BY e.npi\n)\nSELECT\n count(*) AS matched_providers, -- 19\n count(*) FILTER (WHERE any_mandatory) AS mandatory_basis_providers, -- 8\n count(*) FILTER (WHERE first_excl >= current_date - INTERVAL '90 days') AS excl_within_90d, -- 17\n count(*) FILTER (WHERE first_excl < current_date - INTERVAL '365 days') AS excl_over_1yr, -- 2\n min(first_excl) AS oldest_standing_exclusion, -- 2015-06-18\n (current_date - min(first_excl)) AS oldest_age_days -- 4,013\nFROM per_npi;\n-- By statutory basis (distinct NPI): §1128b4 license action 10 · §1128a1 program-related conviction 6\n-- · §1128a3 felony health-care fraud 1 · §1128a4 felony controlled substance 1 · §1128b1 ownership/control 1.\n-- Mandatory §1128(a) conviction-based bases account for 8 of 19; the §1128(b) permissive bases, 11.\n-- 17 of 19 were excluded within the last 90 days (consistent with revocation lag behind the 2026-05-20\n-- LEIE release); 2 have stood for over a year, the oldest since 2015-06-18 (~11 years) — not lag.\n\n-- Rows with NO NPI — excluded from the matchable denominator above and reported\n-- separately. These exclusions cannot be matched to PECOS by identifier at all,\n-- so NPI-based enrollment screening cannot reach them.\nSELECT\n count(*) AS total_leie_rows, -- 68,055\n count(*) FILTER (WHERE nullif(btrim(npi), '') IS NULL) AS rows_no_npi, -- 61,030\n count(*) FILTER (WHERE nullif(btrim(npi), '') IS NOT NULL) AS rows_with_npi -- 7,025\nFROM public.oig_leie_exclusions;", "license": "U.S. Government Works (federal sources; 17 U.S.C. §105)", "generated_by": "Fonteum — https://fonteum.com", "notes": "Aggregate, source-traced figures frozen to the snapshot above. Reproduce by running reproducible_sql against the cited federal dataset; no per-entity records are included." }