Linked Federal and State Sanction Lists Identify More Excluded Medicaid Providers Than Federal Checks Alone
Linking paid Medicaid claims with federal exclusion files and state lists to check active providers.
State and county Medicaid teams audit provider directories and sanction lists to check clinic capacity and program integrity. However, single registries record only one status snapshot and lag behind active billing by years. Linking paid Medicaid claims with both federal exclusion files and California suspension lists gives a complete provider picture.
Linking 488,032 Medicaid billing providers (2018 to 2024) across paid claims, federal exclusions, and California suspension lists finds 487,384 active non-excluded providers and 648 sanctioned providers.[1] A federal-only check leaves out 199 state-suspended providers (30.7%), while 95 later-excluded fraud billers took in $9,694.90 per month unflagged.[2] Because formal exclusion trails final billing by a 50.0-month median gap, teams pair federal and state list checks with paid-claims volume reviews.[3]
Table 1. Linked Billing and Sanction Records Across Four Provider Groups
Observed 2018 to 2024 Medicaid paid claims linked to federal OIG and California DHCS sanction lists.
| Linked Measure |
Group 1: Baseline
Active Non-Excluded Controls
|
Group 2: In-Panel
Sanctioned in 2018-2024
|
Group 3: Delayed
Sanctioned in 2025 or Later
|
Group 4: State File
California State-Only List
|
|---|---|---|---|---|
| Group Size | 487,384 billing providers99.87% of all | 204 billing providers0.042% of all | 95 billing providers0.019% of all | 199 billing providers30.7% of 648 |
| List Recording Status | No LEIE or S&I match | OIG LEIE fraud2018-2024 | OIG LEIE fraud2025+ | DHCS S&I list onlyno LEIE |
| Monthly Paid (Median) | $3,820.94base | $6,721.781.76x base | $9,694.902.54x base | Not reported in source tables |
| Patients & Total Paid | 33.42 patients$83,867 paid | 50.76 patients$81,474 paid | $242,405 paid64.8th pct | Not reported in source tables |
| Billing Timing | No sanction datebase group | Sanction dated in 2018-2024 | 50.0-month gap to exclusion | No date on state S&I list |
| Published Benchmark | Zhu 202358% low-volume | No matched external study | No matched external study | Single lists omit linked cases |
Table 1 Notes & * Modeled Estimates
Expand notes
Table 1 Notes (Mode B Observed Linked Records): Observed 2018-2024 HHS Medicaid Provider Spending claims linked to OIG LEIE (February 2026) and the California DHCS S&I list. Each column states its own group total. Cells report medians within that column. Post-panel gap is 50.0 months median (mean 50.7) from final billing month to LEIE exclusion. External benchmark: Zhu et al. (2023) found 58% of listed Oregon Medicaid mental health providers saw few or no Medicaid patients.
Column Scope. Active Non-Excluded Controls. N = 487,384 NPIs; unit: billing NPI, 2018-2024 claims. Sanctioned in 2018-2024. N = 204 financial fraud NPIs (LEIE Spec A); unit: billing NPI. Sanctioned in 2025 or Later. N = 95 NPIs billing unflagged in 2018-2024; unit: billing NPI. California State-Only List. N = 199 of 648 sanctioned NPIs with zero LEIE match; unit: billing NPI.
Model Crosswalk. Plain table labels map to standard study models. Raw Yearly Average / Simple Avg is an unadjusted OLS model where routine and outbreak budgets mix. County & Year FE is a Two-Way Fixed Effects (TWFE) panel model. Routine-Budget Analysis / 21-Year County Budget Model is a Two-Stage Least Squares (2SLS / Lewbel 2012) model that isolates routine base budget shifts. County Board Budget Rules uses Brown (2014) county board rules.
Two-Ruler Benchmark Summary. Mode B Observed Linked Records (HHS Medicaid Provider Spending 2018-2024, OIG LEIE February 2026, California DHCS S&I List). Ruler 1: Single-List Count: 488,032 Billing NPIs; 648 Sanctioned NPIs Counted as One Total. The single-list count shows 488,032 Medicaid billing NPIs, with 648 (0.13%) on a sanction list. Read as one total, it does not show which cases involve fraud. It also does not show state-only bans or how long providers billed before action. Ruler 2: Linked Claims & Multi-List Check: 299 Fraud Exclusions + 199 State-Only Suspensions + 150 Other. Linking claims to both lists splits 487,384 non-excluded controls from 299 financial fraud exclusions (204 in 2018-2024 and 95 in 2025 or later). It also finds 199 California-only suspensions with no federal match. A federal-only check sees 449 of 648 sanctioned NPIs (69.3%).
Method & Data Audit (TooEarlyToSay.com): OIG Exclusion List Label Audit → · Medicaid Provider Spending File Guide →
Table 2. Reading Each Provider Group Before Drawing Access or Audit Rules
Why one list count needs a link to paid claims and to both federal and state lists.
| Choice Factor | Active Non-Excluded Controls | Sanctioned in 2018-2024 | Sanctioned in 2025 or Later | California State-Only List |
|---|---|---|---|---|
| What the Linked File Sees | Controls bill $3,820.94 per month to 33.42 patients and serve as the base group. | In-panel fraud billers take in $6,721.78 per month, or 1.76 times the base median. | Post-2025 exclusions bill $9,694.90 per month, or 2.54 times controls, with no flag. | 199 providers sit on the state list with no federal LEIE record. |
| Data Gap the Group Shows | A directory headcount treats every listed provider the same, while claims show real volume. | A sanction date inside the window mixes pre-action billing with post-action stops. | A 50.0-month median lag passes between final billing and formal federal exclusion. | Checking only the federal list leaves out 30.7% of sanctioned providers (199 of 648). |
| Planning Risk From One List | Counting all 487,384 controls as equal capacity equates one-time billers with full-time clinics. | Treating 204 excluded providers as never active leaves out claims paid before action. | Directories built in 2024 list all 95 post-panel fraud billers as clean. | Audits that check only the federal LEIE list 199 state-suspended providers as clean. |
| Best-Fit Data & Audit Fix | Weight Lists by Paid Claims. Report billing providers next to monthly pay and patient counts to measure real access. | Track Sanction Dates by Month. Store the sanction date with the last billed month to split pre-ban claims. | Pair List Checks With Claims. Monitor high-volume billing during the 50-month lag and relink past claims when bans post. | Match Federal and State Lists. Check every provider against both OIG LEIE and the state list before reporting totals. |
Table 2 Notes: Reconciliation Math, Payment Ratios & External Benchmarks
Expand notes
• Reconciliation Math (Mode B): 488,032 billing NPIs equal 487,384 non-excluded controls plus 648 sanctioned NPIs. The 648 sanctioned NPIs split into 299 Specification A financial fraud exclusions (204 dated 2018-2024 and 95 dated 2025 or later; statutes 1128(a)(1), 1128(a)(3), 1128(b)(1)). Another 199 are California S&I-only suspensions. The remaining 150 are other LEIE exclusions (patient abuse, license actions, controlled substances, and pre-2018 cases). The state-only share is 199 / 648 = 30.7%. The federal share is 449 / 648 = 69.3%.
• Payment Ratios & Enforcement Lag (Mode B): Median monthly paid compared to controls is 6,721.78 / 3,820.94 = 1.76x for in-panel cases. For post-panel cases, it is 9,694.90 / 3,820.94 = 2.54x. The enforcement lag is the historical median gap from a provider's final billing month to formal LEIE exclusion (50.0 months, mean 50.7, IQR 32.5-68.0). Refreshing registry linkages updates historical labels once exclusions post, but does not shorten OIG's investigation-to-exclusion timeline.
• External Benchmark: The Zhu et al. (2023) Oregon Medicaid number (58% of listed mental health providers) is cited in public/tools/access-explorer/method.html. It describes directory lists in another state and is not subtracted from any count here.
• Data Provenance: All counts and medians come from Mcaid/05_fraud_detection/paper/output/results.json (controls_total_n, in_panel_fraud_n, prospective_fraud_n, ctrl_paid_per_month_median, during_paid_per_month_median, post_paid_per_month_median, post_pct_rank_total_paid, gap_median_months) and output/tables/tab_a2_exclusions.tex (199 S&I-only, 648 total, 488,032 universe).
Why Counting Names on a List Differs From Measuring Active Billers
Provider directories and sanction files are both static lists. Each records a status on the day it was last updated. Read alone, the HHS Medicaid file lists 488,032 billing NPIs and 648 sanctioned providers as one sum. Linked to paid claims and to both federal and state lists, those 648 split into 299 fraud exclusions, 199 California-only suspensions, and 150 other cases.
Health economics treats list records as a data system that needs cross-checks on both sides. Directories keep listing providers who see few Medicaid patients. For example, Zhu et al. (2023) found 58% of listed Oregon Medicaid mental health providers saw few or no patients. Sanction lists also lag behind billing by years. The 95 providers excluded in 2025 or later billed a median $9,694.90 per month through 2024 without a flag. Formal exclusion arrived a median 50.0 months after final billing, so list lookups alone cannot flag active billing during that pre-exclusion window and must be paired with paid-claims volume checks.
Data Sources, Model Inputs, Notes & Endnotes
(Numbered endnotes [1]–[4], state data sources & math rules)
Expand details
Case Study Focus: System Dynamics & Feedback · Single-Registry Counts vs. Multi-Registry Reconciliation (Mode B Linked Claims & Exclusion Records)
Numbered Endnotes & Main Sources
- [1]Linked Medicaid Claims & Sanction Universe (2018-2024): HHS Medicaid Provider Spending claims cover 488,032 billing NPIs and 487,384 non-excluded controls. These link to the OIG List of Excluded Individuals/Entities (LEIE) with 449 federal matches, including 299 financial fraud exclusions.
- [2]California State-Only Suspension Records: California DHCS Suspended & Ineligible Provider List holds 199 of 648 sanctioned NPIs (30.7%) with zero federal LEIE match.
- [3]Enforcement Lag & Directory Benchmarks: Post-panel financial fraud exclusions cover 95 NPIs dated 2025 or later ($9,694.90/mo median paid vs. $3,820.94/mo controls, 50.0-month median gap from final billing to exclusion). Zhu, J. M., et al. (2023), Health Affairs, 42(1), 49-57 reports a 58% inactive-directory benchmark.
- [4]OIG List of Excluded Individuals/Entities (LEIE)
Mode Classification (Mode B Observed Linked Records & Published Literature): All provider counts (488,032 total; 487,384 controls; 204 in-panel and 95 post-panel fraud exclusions; 199 S&I-only; 648 sanctioned) are observed Mode B records in output/results.json and output/tables/tab_a2_exclusions.tex. Median monthly payments ($3,820.94, $6,721.78, $9,694.90), patient counts (33.42, 50.76), and the 50.0-month lag come from the same files. The 58% Oregon figure (Zhu et al., 2023) is external context in public/tools/access-explorer/method.html.
Primary References & Data Sources: Zhu, J. M., Charlesworth, C. J., Polsky, D., & McConnell, K. J. (2023), Health Affairs, 42(1), 49-57; HHS Medicaid Provider Spending; OIG List of Excluded Individuals/Entities (LEIE); and the California DHCS Suspended & Ineligible Provider List.
Companion Method & Code Guides (TooEarlyToSay.com)
- The Label That Isn't: Why "Excluded" Doesn't Mean "Fraudulent". Section 1128 exclusion authorities and why LEIE exclusion is a contaminated fraud label; the 40% license-revocation share and matched-provider counts are article-reported.
- What 238 Million Medicaid Billing Rows Can Support. Field definitions of the HHS Medicaid provider-spending file, including billing and servicing NPIs, suppression rules, and the 238,015,729-row artifact.
Common Questions & Method Notes
- What does a single sanction list leave out in Medicaid provider data?
- A single list records one status per provider as of its last update. In the linked 2018-2024 Medicaid file, 199 of 648 sanctioned NPIs (30.7%) appear only on the California DHCS Suspended and Ineligible list. A check of the federal LEIE alone does not return them.
- Why do some sanctioned providers show up as active billers in claims data?
- Formal exclusion lags behind billing. Among providers excluded for financial fraud, the median gap from the final billing month to the LEIE exclusion date is 50.0 months. The 95 providers excluded in 2025 or later billed a median $9,694.90 per month during 2018-2024 with no list flag.
- How should analysts link directory, claims, and sanction records?
- Match every NPI against paid claims, the federal LEIE, and the state suspension list. Record exclusion dates next to the last billed month, relink past claims when new bans post, and monitor high-volume paid claims during the 50.0-month window before formal exclusion.