DECLASSIFIED -- FOR PUBLIC DISTRIBUTION

EP01 | THE HOUSE ALWAYS WINS | EVIDENCE LOCKER | CASE FILE 01 OF 06

DECLASSIFIED

EVIDENCE LOCKER

Claim-by-Claim Sourcing -- Every Receipt Documented

EP01_01_Evidence_Locker_v3_0 | Updated 2026-04-15

[P] Primary Source -- lawsuit, government document, internal records
[J] Quality Journalism -- ProPublica, STAT News, major investigation
[S] Secondary Source -- aggregated, contextual, academic
19Claims
15+Sources
3Segments
-1Unverified

For how these mechanisms work, see the Decoder Ring. For who profits, see Follow the Money.

The Appeal Gap

The Appeal Gap

Sources: KFF/CMS 2024-2025, AMA 2024 Survey, Senate Finance Committee 2023

% / Rate values (right axis) Volume (left axis, millions)
COLD OPEN
Three thesis-level statistics. Each sourced under the segment where it is developed:
- "They deny claims every 1.2 seconds." -> Claim #12
- "Eighty percent of denials get overturned -- when someone actually fights." -> Claim #17
- "Point-two percent of people fight." -> Claim #18
SEGMENT 1: MORNING -- "WELCOME TO THE CASINO" [[1:21-5:45]]
##1 [[1:21]] [S] Secondary ALLEGATION

Your body got stapled to someone else's spreadsheet in 1942.

Employer-sponsored health insurance originated during WWII wage controls. The 1942 Stabilization Act froze wages; employers offered health benefits as compensation instead. IRS later codified the tax exclusion via Revenue Ruling 54-264.

IRS Historical Records: Revenue Ruling 54-264.

Standard historical account, widely documented.
#2 [[1:35]] [S] Secondary

An accident that's been running for about eighty-four years.

Math check: 2026 - 1942 = 84 years. Accurate as of recording date.

Math check: 2026 – 1942 = 84 years. Accurate as of recording date.

Arithmetic verification only.
#3 [[1:52]] [J] Quality Journalism

Christopher McNaughton was a college student with ulcerative colitis. His account was flagged "HIGH DOLLAR."

ProPublica investigation (Feb 2023) based on internal UnitedHealthcare documents obtained through litigation. McNaughton v. UnitedHealthcare (lawsuit, settled Feb 2023). Internal documents revealed the "HIGH DOLLAR" patient classification used to flag costly accounts.

ProPublica: UnitedHealthcare Tried to Deny Coverage to a Chronically Ill Patient (Feb 2023)
McNaughton v. UnitedHealthcare (lawsuit, settled Feb 2023)

Internal documents from litigation revealed the "HIGH DOLLAR" classification.
#4 [[2:15]] [J] Quality Journalism ALLEGATION

"Why not just deny them all and see which ones come back? From a cost perspective, it makes perfect sense."

Quote attributed to a former insurance executive. Referenced in investigative reporting on industry denial practices. Exact original source document pending verification.

Quote attributed to a former insurance executive. Referenced in investigative reporting on industry denial practices.

Quoted on-air. Source attribution pending. Marked as allegation.
#5 [[3:10]] [S] Secondary

Humira. Seventy-two thousand dollar a year drug.

Humira (adalimumab) list price widely reported at approximately $72,000-$80,000/year before biosimilar competition. At the time of Joey's case, $72,000/year was the standard list price for the brand-name product. Manufacturer pricing data.

Manufacturer pricing data. Multiple industry sources and pharmacy benefit analyses.

Manufacturer pricing data. Verified against multiple industry sources.
#6 [[4:30]] [S] Secondary

Every prescription runs through a pharmacy benefit manager. A PBM. Every prescription generates a rebate. The test? Zero kickback.

PBM rebate structures are well-documented. Prescription drugs generate manufacturer rebates paid to PBMs and plan sponsors. Diagnostic tests and lab work do not generate comparable rebates.
FTC Interim Report on PBMs (2024): <a href="https://www.ftc.gov/reports/pharmacy-benefit-managers-report">https://www.ftc.gov/reports/pharmacy-benefit-managers-report</a>

FTC PBM report, congressional testimony, industry analyses.
SEGMENT 2: NOON -- "THE HIDDEN MATH" [[5:50-11:10]]
#7 [[6:25]] [P] Primary

Blue Cross has no legal obligation to act in Joey's or his company's best interest.

Under ERISA, third-party administrators of self-insured plans have limited fiduciary obligations compared to fully insured plans. The TPA's contractual duty is to the plan document, not to individual members. This structural gap is widely documented in benefits law.

ERISA (Employee Retirement Income Security Act of 1974). Benefits law literature.

Widely documented in benefits law and legal analyses.
#8 [[6:40]] [P] Primary

Primary source: Joey. $75/week x 52 = $3,900/year. National context from KFF 2025 Employer Health Benefits Survey: Average worker contribution is $1,440/year for single coverage and $6,850/year for family coverage.

Primary source: Joey. $75/week x 52 = $3,900/year. National context from KFF 2025 Employer Health Benefits Survey: Average worker contribution is $1,440/year for single coverage and $6,850/year for family coverage.

Joey's figure falls between single and family national averages.
#9 [[6:45]] [P] Primary

His employer pays Blue Cross another $27,000 a year to manage the plan.

Primary source: Joey's W-2 and employer benefits documentation. As spoken in the episode: $27,000/year. 2025 update: Joey's W-2 shows employer contribution of $34,069.35, a 26% increase. National context from KFF 2025: Average total premium for family coverage is $26,993. Employers pay an average of $20,143.

Joey's employer now pays $34,000+ -- nearly $14,000 above the national average employer contribution ($20,143).
#10 [[7:00]] [J] Quality Journalism

The company that makes Humira pays Blue Cross thirty to forty percent of that $72,000 list price back as a kickback. Call it $25,000.

PBM rebate ranges on brand-name biologics widely reported at 30-40% of list price per FTC findings and congressional testimony. $72,000 x 30-40% = $21,600-$28,800. "$25,000" is a reasonable midpoint estimate. Specific rebate percentages vary by plan and PBM contract.

Rebate percentages are industry-standard ranges, not exact figures for Joey's specific plan.
#11 [[7:50]] [S] Secondary

Nobody can see the rebate. Not Joey. Not his employer. Not the pharmacist.

PBM rebate opacity documented by FTC and congressional investigations. Rebate agreements between manufacturers and PBMs are typically confidential. Employers in self-insured plans often do not see exact rebate amounts. Patients and pharmacists have no visibility into rebate flows.

Structural claim about PBM opacity. FTC report + multiple congressional hearings on PBM transparency.
#12 [[8:10]] [J] Quality Journalism

Cigna doctors denied 300,000 claims over two months. The average review time per claim was 1.2 seconds.

ProPublica investigation (March 2023) based on internal Cigna data and audit logs. Cigna's PxDx system was used to deny approximately 300,000 claims over two months. Audit logs showed average review time of 1.2 seconds per claim. Doctors approved flagged denials in bulk without reading patient files.

Anchor stat for the episode. 1.2 seconds is the thesis number.
#13 [ [8:30]] [J] Quality Journalism

One doctor denied 121,000 claims in sixty days.

ProPublica (March 2023). Internal audit logs showed one medical director processing approximately 121,000 claim denials in a two-month period through the PxDx system.

ProPublica PxDx investigation.
#14 [[8:45]] [J] Quality Journalism

The internal system is called PXDX. Procedure by diagnosis. Page 47 of their internal documents.

ProPublica (March 2023). PxDx (Procedure-to-Diagnosis) is Cigna's internal rules engine that cross-references procedure codes against diagnosis codes. If the combination is not in the approved pairs, the claim is flagged for denial. Referenced in Kisting-Leung v. Cigna class action court filings. Page 47 reference from internal documents cited in litigation.

Page 47 reference from internal documents. Court filings + investigative journalism.
#15 [[9:10]] [J] Quality Journalism

UnitedHealth built a tool called nH Predict. The algorithm looked at six million patient records and decided your recovery should average out to fourteen days.

STAT News "Denied by AI" investigative series (2023-2024). nH Predict developed by NaviHealth, a UnitedHealth subsidiary (now Home & Community Care under Optum). Tool uses six million historical patient records to generate expected length-of-stay predictions for post-acute care. Senate Permanent Subcommittee on Investigations "Refusal of Recovery" report (2024) documented cases where nH Predict cutoffs overrode treating physician recommendations. UnitedHealth's post-acute care denial rate more than doubled after deploying NaviHealth, climbing from 8.7% to 22.7% between 2019 and 2022.

STAT News + Senate investigation. Denial rate doubled after nH Predict deployment.
#16 [[9:50]] [P] Primary ALLEGATION

A federal lawsuit alleges UnitedHealth knew the algorithm was wrong ninety percent of the time when families actually fought back.

Estate of Lokken v. UnitedHealth Group -- Class Action Complaint (D. Minn. Nov 2023). Plaintiff allegation based on internal data showing approximately 90% overturn rate when families appealed nH Predict-based coverage terminations. Case still actively advancing through federal court as of March 2026 after judge ordered broad document discovery in March 2025.

Plaintiff allegation, not adjudicated finding. Marked as allegation. Three supporting legal sources.
SEGMENT 3: EVENING -- "THE VERDICT" [[11:15-16:00]]
#17 [[11:30]] [J] Quality Journalism

When people do appeal, ninety percent win.

The episode rounds to "ninety percent." Three underlying data points support the 80-90% range: (1) KFF January 2026 analysis of CMS data shows 80.7% of appealed Medicare Advantage prior authorization denials were partially or fully overturned in 2024. (2) The Lokken lawsuit alleges approximately 90% overturn rate for nH Predict appeals specifically. (3) A January 2026 Health Affairs study from Stanford found an 82% overturn rate across MA appeals.

The 80-90% range is well-supported across multiple sources. "Ninety percent" is the rounded figure used in the episode.
#18 [[11:45]] [P] Primary

0.2% appeal.

KFF analysis of ACA marketplace data. In 2024, consumers appealed fewer than 1% of approximately 85 million in-network denied claims (262,982 appeals filed = approximately 0.3%). The 0.2% figure cited in the episode comes from earlier KFF ACA marketplace analyses. Both figures support the thesis. Scope: ACA marketplace plans on HealthCare.gov only, not all insured populations. Medicare Advantage appeal rates are higher (11.5% in 2024) but still low relative to the overturn rate.

2024 data shows approximately 0.3%, slightly above the 0.2% cited in the episode. Both support the thesis.
#19 [[13:00]] [S] Secondary

I buy $47 of UNH stock. That's about one-sixth of a share. Fractional shareholders still get to attend the annual meeting.
Line was flubbed in recording. Said 47 Shares

UNH share price at time of recording: approximately $280-300. $47 / ~$282 = roughly 1/6 of a share. Fractional share ownership through most brokerages does technically entitle holders to shareholder meeting attendance, though policies vary by broker and company.

Comedy bit (The Letter I Didn't Send). Math checks at recording-date share price.
SOURCE SUMMARY
Category Count
Total claims documented 19
Primary sources [P] 5
Quality journalism [J] 9
Secondary sources [S] 5

Key Source Documents

Source Type Claims
ProPublica: Cigna PXDX (Mar 2023) [J] #12, #13, #14
ProPublica: McNaughton/UHC (Feb 2023) [J] #3
STAT News: "Denied by AI" (2023-2024) #15
Lokken v. UnitedHealth (D. Minn. 2023) [p] #16, #17
Senate PSI: "Refusal of Recovery" (2024) [p] #15
KFF: MA Prior Auth (Jan 2026) [p] #17
KFF: ACA Claims Denials (Mar 2026) [p] #18
KFF: Employer Health Benefits (2025) [s] #8, #9
FTC PBM Interim Report (2024) [S] #6, #10, #11
ERISA / Benefits Law [S] #7

Last verified: March 2, 2026. Found an error? corrections@theranter.com