Indiana Medicaid: Notable Patterns & Detection Methodology Validation

What this report establishes

This memorandum synthesizes findings from a surface analysis of three public Indiana data sources: the 2012-2017 Indiana Medicaid provider claim universe (63,569 provider-year records, 7,211 distinct National Provider Identifiers, $20.0 billion in total claims); the 2024-2026 Indiana PeopleSoft Financials general ledger expenditure detail (2.08 million transactions, $141 billion in net state spending); and the FSSA Social Determinants of Health screening dataset for returning citizens released from Indiana Department of Corrections facilities between January 2019 and March 2022 (approximately 6,000 respondents).

The lead analytical finding is a methodology validation. Applying an unsupervised pattern detector to the historical claim universe, we independently identified fourteen out-of-state laboratories exhibiting a "post-enforcement collapse" signature — billing that peaked, then declined by sixty percent or more within one to three years. Cross-referencing these labs against federal Department of Justice and HHS-OIG records after the statistical identification, we confirmed that six of the fourteen are documented federal enforcement cases: PremierTox 2.0 ($15.75M False Claims Act settlement plus criminal convictions of five owners), True Health Diagnostics (CMS payment suspension, Chapter 11 bankruptcy, CEO indictment, $6M settlement), Ameritox, Physicians Choice Laboratory, Calloway Laboratories ($20M federal settlement), and Millennium Health ($256M DOJ/OIG settlement). The detector surfaced these cases from public Indiana data alone, with no prior knowledge of the federal proceedings.

The same detector identifies a forward-looking surveillance list: nineteen out-of-state laboratories that entered Indiana Medicaid billing during 2014-2017 with the inverse signature — sudden appearance and rapid growth. These are the candidates most likely to exhibit the post-enforcement collapse pattern in subsequent years, and the most defensible targets for continuous monitoring against current managed-care encounter data.

Beyond the laboratory findings, the analysis surfaces several additional patterns worth FSSA attention:

• The genetic and molecular testing subcategory grew 767% over the six-year period — by far the fastest-growing lab subcategory — with per-recipient costs concentrated in the $2,000-$7,000-per-test range characteristic of specialty genetic panels. Several named genetic testing laboratories appear in the high-cost outlier list.
• FSSA's optional ten-question SDOH screening of returning citizens shows two domains at 76% YES (of respondents who answered), substantially above general-Medicaid population benchmarks of 17-30% for any single SDOH domain. The dataset is archived as of July 2025 and the original question text is not publicly available; restoring both would substantially increase research utility, particularly in support of any Section 1115 reentry Medicaid demonstration Indiana pursues.
• One named mental health provider — Meridian Health Services Corporation, Muncie — exhibits an extreme claims-per-recipient pattern (156-189 service encounters per recipient per year, $30,000-$56,000 per recipient annually, versus a peer median of approximately 15 claims and $1,000 per recipient). The provider remains active in the current state expenditure ledger ($5.5M FSSA flow in the 2024-2026 window).
• Two same-address laboratory clusters outside obvious hospital campuses (Chicago and Cincinnati locations) warrant ground-truth verification to distinguish legitimate shared facilities from shell-network patterns.

How the analysis was conducted

This analysis was conducted entirely from public data — no protected health information, no FSSA-internal claim data, no managed-care encounter data. The methods applied are standard healthcare-fraud-detection techniques: peer-outlier analysis of dollars per recipient and claims per recipient within provider type; year-over-year growth and decline detection; geographic distance analysis between recipients and out-of-state providers; vertical concentration analysis in known high-fraud laboratory subtypes (urine drug testing, genetic and molecular testing); and Benford's Law verification on first-digit distributions of claim amounts. The population-level Benford check passes (chi-square 9.87 against critical value 15.51), confirming no broad fabrication signal — individual outliers can still hide in a Benford-compliant population, as Millennium Health did.

Throughout the report, statistical patterns are described as signals worth investigating rather than findings of wrongdoing. Where federal enforcement actions are part of the public record, they are cited explicitly. Where allegations are unproven — including the September 2024 whistleblower lawsuit naming Indiana managed-care entities and hospital systems — the data patterns are described without imputing intent, and the lawsuit is noted as unproven with defendants not having been found liable.

Three categories of pattern that this analysis is structurally unable to assess:

• Service-code-level details, patient diagnoses, and medical-necessity criteria, which would require claim-level data with appropriate data-use agreements.
• The bulk of current Medicaid spending, which flows through managed-care organizations and is visible only in CMS T-MSIS encounter data (accessible through the Research Data Assistance Center application process). Of the approximately $19.6 billion in net annual Medicaid cash flow visible in the 2024-2026 state ledger, the four contracted managed-care entities — Anthem, MDwise, MHS/Centene, CareSource — do not appear as identified vendors; the spending is booked through internal accounting placeholders.
• Durable medical equipment, hospice, and home health categories, which are essentially absent from the historical Indiana provider dataset (one DME record, one home health agency record across six years) and are known high-fraud verticals in other state Medicaid programs nationally.

Why these patterns matter beyond Indiana

Luminary AI Technologies is conducting parallel analyses of Medicaid program performance and provider-claim integrity across multiple state Medicaid programs. The patterns visible in the Indiana data — sudden-appearance and post-enforcement-collapse signatures concentrated in out-of-state laboratories; vertical concentration in genetic testing and toxicology subcategories; SDOH needs profiles for justice-involved populations that exceed general-Medicaid benchmarks by factors of two to three — are consistent with what we observe in other state Medicaid programs we have examined.

The federal enforcement environment is intensifying. Recent CMS communications and HHS-OIG work plans indicate expanded program integrity audits of state Medicaid programs, with attendant care providers, behavioral health services, and managed-care encounter data quality emerging as 2026 priority areas. A representative recent action: in April 2026, Indiana FSSA itself announced it is seeking the return of approximately $200 million in improper payments to attendant care providers — a vertical that is essentially invisible in the public dataset analyzed for this report.

State Medicaid program integrity teams nationally are operating with detection infrastructure that has not kept pace with the sophistication and scale of provider-side schemes, particularly in the post-2018 managed-care-administered environment where MCO encounter data quality and timeliness vary widely across states. The detection patterns demonstrated in this report — applied to the right data sources, on a continuous-monitoring basis — would have surfaced the disappearing-lab cases visible here within months of their growth phase rather than years after the federal enforcement actions resolved.