Luminary AI Technologies Insight at Scale
Indiana Medicaid
Indiana Medicaid · system studies & briefings

Our work on Indiana Medicaid.

One place for the maps, structural studies, and workflow and fraud-pattern work Luminary has done on the program — how it's built, where it breaks, and what that costs the people inside it.

We build the working picture of a system before anyone proposes changing it. These maps and studies are that picture for Indiana Medicaid: how the program is put together, where it strains, and what the people inside it — providers, members, and the caseworkers holding it together — actually run into.

This is work we've already done, on our own. We're looking for the partner or the contract that puts it to use: an agency, an MCE, a county, or a team inside one of them with a problem shaped like this one. If that's you, write to us.

Luminary AI Technologies · Indianapolis ben@luminary-tech.ai

Indiana Medicaid · Structural Systems Maps

4 in this set
In depth
Deep Dive

Indiana Medicaid · Deep Dive

The narrative companion to the System Map — every tier explained in full, top to bottom.

The detailed read behind the map: each of the seven tiers explained in depth, with the policy mechanics, dollar flows, statutory authority, and operational dependencies behind every node. Where the System Map shows what connects to what, the Deep Dive explains why, how, and where it strains — the working reference for anyone who needs the full picture, not just the diagram.

Open the Deep Dive→
Wall-poster
System Map

Indiana Medicaid · System Map

The full diagram — seven tiers, every node and connector, the eleven core process flows. The at-a-glance source of truth, designed to print at wall-poster scale.

Open System Map→
Print set
Posters

Indiana Medicaid · Posters

A printable poster set built from the Deep Dive — each tier on its own sheet, designed for war-room walls and briefing rooms where the system needs to be in front of decision-makers.

Open Posters→
Risk register
Fail Points

Indiana Medicaid · Fail Points

Where the system breaks today — the documented and pattern-detected failures in eligibility, claims, MCE handoffs, and provider operations, ranked by impact and where intervention is feasible.

Open Fail Points→

Indiana Medicaid · Studies & Briefings

5 in this set
Study
Luminary Study · Enrollment Workflow

The Workflow Factor

An analysis of Indiana's Medicaid enrollment workflow — and what the screens themselves contributed to a coverage decline of more than 400,000 people.

Fifteen screens across five workflows, captured live: the member keeping coverage, the provider enrolling in IHCP, the provider answering a PERM audit, the First Steps provider in the EI Hub, and the DFR worker inside IEDSS. Each screen is read for what it asks, what it assumes, and where an eligible Hoosier or a current provider drops out — then set against the state's own 2024 PERM error profile, which is predominantly administrative rather than fraud or ineligibility. Ends with costed interventions in four phases.

400,000+
Enrollees lost, Jan 25–Apr 26
20%
Children's drop, highest in US
7.57%
2024 PERM blended rate
46%
Of error: document routing
Open the study→
Data study
Luminary Study · Money Flow

Indiana Medicaid Money Flow

Where Indiana's $19.2B Medicaid dollar goes by fiscal year — and the growing share that leaves as a managed-care premium, with no public record of what it bought.

A decade of CMS-64 expenditures, Indiana's spend per enrollee against the national average, and the FY2016 split where 40.7% of the dollar leaves as capitation — a premium, not a traceable service. Then the national drug layer, where price rather than volume is driving spend, and the three data traps that silently inflate every Medicaid total. Public data only.

$19.19B
Total computable, FY2024
70.8%
Federally funded
40.7%
Left as capitation, no service detail
81.6%
In managed care, 2024
Open the study→
56 states
Luminary Study · Observability

The Medicaid Observability Gap

A 56-jurisdiction study of how much of each state's Medicaid dollar can still be traced to a service, a provider and a price — and what happened to the rest. Indiana runs through every finding.

The program shifted into managed care almost everywhere while the encounter data didn't keep pace — half the country can't reliably identify a Medicaid provider in its own federal submission. It tests what actually drives cost (demographics explain almost nothing; three policy choices explain ten times more) and ends with exactly what's missing and where to ask for it. Public data only, no data-use agreement.

$370.4B
In heavily capitated, data-weak states
26 of 54
Can't reliably ID a provider
78.2%
Of enrollees in managed care
47 of 53
States improving since 2019
Open the study→
Federal data
Luminary Study · Integrity

Where Medicaid Actually Leaks

A reading of the federal record: $37.39B in Medicaid improper payments, and why more than three-quarters of it is missing documentation rather than fraud.

The FY2025 PERM measurement, read against GAO's own findings: the managed-care channel that covers three in four enrollees is scored at 0.00%, and CMS attributes 77% of the measured error to missing documentation — eligibility evidence and provider records, not misdirected dollars. Indiana's free Oracle claims-anomaly pilot sits in the same frame. Public sources only, checked against the primary documents — and no finding of fraud against anyone.

$37.39B
Improper payments, FY2025
77%
Of it: documentation, per CMS
0.00%
Managed-care rate, ~3/4 of enrollees
2003
On GAO's high-risk list since
Open the study→
Preliminary
Medicaid Fraud Pattern Study

Preliminary Medicaid Fraud Pattern Study: Indiana

A surface read of Indiana Medicaid claims and provider data, structured for a program-integrity audience — not duplicating findings already in federal enforcement records or active litigation.

Combining the 2012-2017 provider claim universe, the 2024-2026 state expenditure ledger, and the FSSA SDOH screening dataset. Disappearance patterns, sudden-appearance lab candidates, recipient-distance outliers, address clusters, and an MCE-level read on Meridian — with six pattern-detected labs that independently surfaced into documented federal enforcement cases as the methodology's working validation.

6 of 14
Pattern → fed case
19
Sudden-appearance labs
+766%
Genetic lab 6-yr
6,057
SDOH respondents
Open the briefing→