The US spends ~$15,474 per person on healthcare. Japan spends ~$5,790 and has the highest life expectancy in the OECD. That gap is roughly $3.24 trillion per year.
This project finds it, one issue at a time. Each issue identifies one fixable problem, quantifies the waste from primary federal data, and recommends a specific policy fix. All code is open-source. Anyone can reproduce the analysis.
Subscribe on Substack | MIT License | Contributing
Latest Issue (#9): The Employer Trap — Plans paying their broker through one carrier-paid commission system, plans paying their administrator through another, and the worker paying for both in wages they never see. We pulled DOL Form 5500 Schedule A and Schedule C disclosures for plan year 2023, the first full year of post-CAA 2021 fiduciary transparency, and built peer-group fee benchmarks at plan level. $6.6B/year in conservative savings identified from broker commissions above the 3% DOL benchmark and admin fees above peer-group medians. Read it on Substack →
| # | Issue | Savings | Key Finding | Data Source |
|---|---|---|---|---|
| 1 | OTC Drug Overspending | $0.6B/yr | Medicare pays Rx prices for drugs you can buy off the shelf | CMS Part D 2023 |
| 2 | The Same Pill, A Different Price | $25.0B/yr | US pays 7–581x more than peer nations for the same drugs | CMS Part D, NHS Tariff, RAND |
| 3 | The 254% Problem | $73.0B/yr | Commercial insurers pay 254% of Medicare for identical hospital procedures | CMS HCRIS, RAND 5.1 |
| 4 | The Middlemen | $30.0B/yr | Three PBMs process 80% of US prescriptions and extract ~$30B/yr through spread pricing, rebate opacity, and formulary manipulation | FTC Interim Reports, Ohio Auditor, JAMA |
| 5 | The Paper Chase | $200.0B/yr | US spends $4,983/person on healthcare admin vs. $884 in peer nations; original HCRIS analysis of 4,518 hospitals reveals 6.2× variance in overhead costs | CMS HCRIS, CMS NHE, OECD, AMA |
| 6 | The Supply Closet | $28.0B/yr | Original HCRIS analysis of 5,480 hospitals reveals massive variance in per-discharge supply costs; CMI-adjusted P75/P25 ratios of 2.5–3.4× within same-size peer groups | CMS HCRIS FY2023 |
| 7 | The GLP-1 Gold Rush | $40.0B/yr | US GLP-1 spending grew 1,200-fold in 5 years ($57M→$71.7B); US pays 3–5× international prices; original 10-year BALANCE budget projection for Medicare GLP-1 coverage | CMS Part D, OECD, KFF, CBO |
| 8 | The Denial Machine | $32.0B/yr | Original CMS-0057-F extraction of 93 MA contracts (UHC denial rate 13.5%, appeal overturn 58–65%, ~3M denied entitled care annually); care suppression, vertical integration arbitrage, and AI-driven denial escalation | CMS-0057-F, UNH/HUM 10-K, Health Affairs, Stanford npj |
| 9 | The Employer Trap | $6.6B/yr | First plan-level analysis of post-CAA 2021 broker and admin-fee disclosures (8,180 health welfare plans, 23.8M participants); per-plan broker commissions above 3% DOL benchmark and admin fees above peer-group medians | DOL Form 5500 Schedule A and Schedule C 2023, KFF EHBS 2024, JAMA Network Open |
| Running Total | $435.2B/yr | 13.4% of the $3.24T gap |
Issue #9 is the first issue at the revised $3.24T denominator (CMS NHE 2024 final, released April 18, 2026). Issues #1–#8 published using the prior $3T denominator and are not retrofitted.
The same operations. Exposed to the same clinical evidence. Wildly different prices.
Source: iFHP International Health Cost Comparison 2024–2025. Prices are median insurer-paid amounts.
The employer-sponsored insurance system covers 136 million participants and converts system-level healthcare price excess into a hidden tax on wages. Premiums for employer-sponsored insurance climbed from 7.9 percent of total compensation in 1988 to 17.7 percent in 2019; the difference came out of wages that did not rise. The Consolidated Appropriations Act of 2021 changed this structurally: ERISA Section 408(b)(2)(B) now requires brokers and consultants to disclose all direct and indirect compensation above $1,000 to plan fiduciaries. Plan year 2023 is the first full post-CAA 2021 health-plan disclosure year. We pulled every Schedule A (broker compensation) and Schedule C (service-provider compensation) filed by 4A health welfare plans from DOL's "Latest" research file and built peer-group fee benchmarks at plan level — the public reference point that the Lewandowski v. J&J and Navarro v. Wells Fargo dismissals said was missing. Three booked components total $6.6 billion per year (range $6.6B to $12.2B): broker commissions above the 3 percent DOL benchmark, broker rate extension to self-insured plans, and admin-fee variance above peer-group medians at conservative 30 percent reducibility.
Source: DOL Form 5500 Schedule A and Schedule C, 2023 Latest file. n=425 plans filing both disclosures.
Read the full analysis → issue_09/newsletter_issue_09.md
Reproducing the analysis
cd issue_09
# Build the Schedule A and Schedule C analysis datasets
python 01_build_data.py
python 02_build_data_schedule_c.py
# Generate all four analysis charts (peer variance, broker-vs-admin boxplot, savings decomposition, running tracker)
python generate_all_charts.pyKey outputs:
results/savings_estimate_v2.json— Booked components and range with all assumptionsresults/schedule_c_admin_variance.csv— Per-peer-group admin fee P10/P25/P50/P75/P90results/schedule_a_c_linkage.csv— 425 plans filing both Schedule A and Schedule C, with broker rate and admin fee per participantresults/overlap_matrix.md— Component-level overlap accounting against Issues #3, #4, #5, #8results/meps_ic_verification.md— Cross-validation against MEPS-IC public tablesfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| DOL Form 5500 Schedule A 2023 (Latest research file) | Broker and consultant commissions disclosed by fully insured plans (7,036 plans) |
| DOL Form 5500 Schedule C 2023 (Latest research file) | Service-provider compensation disclosed by health welfare plans with trust funding (8,180 plans, 23.8M participants, $12.47B in disclosed fees) |
| KFF Employer Health Benefits Survey 2024 | Self-insured share (65%), per-worker premium, plan-design distribution |
| MEPS-IC 2024 (AHRQ) | State and national employer benefit verification tables |
| BLS Employer Costs for Employee Compensation | Quarterly health benefit share of total compensation, 2014–2025 |
| CMS National Health Expenditure 2024 final | Total private insurance spending; per-capita US figure for $3.24T denominator |
| Hager K, Emanuel EJ, Mozaffarian D, JAMA Network Open (Jan 2024) | Premium-share-of-compensation trajectory 1988–2019 by income decile and race |
| Baicker K, Chandra A, Journal of Labor Economics (2006) | Wage offset from premium growth, ~dollar-for-dollar over time |
| RAND Corporation Round 5.1 (2023) | Commercial hospital prices = 254% of Medicare (referenced in The Fix section) |
| Lewandowski v. Johnson and Johnson (D.N.J. Nov. 26, 2025) | Standing dismissal in ERISA fiduciary case for lack of public benchmark |
| Navarro v. Wells Fargo (D. Minn. Mar. 24, 2025) | Companion dismissal on the same logic |
| Marsh McLennan, Willis Towers Watson, Aon plc 10-K and DEF 14A filings | Broker-consulting firm financials and compensation structures |
| OpenSecrets.org federal lobbying disclosure (2020–2024) | Industry lobbying expenditures |
Key methodology notes
- Three booked components, each computed from federal filings:
- Component A ($2.18B): per-plan broker commission above the 3% DOL benchmark, 7,036 fully insured plans
- Component B ($2.77B): Schedule A broker rate ($82.68 per life) extended at conservative 0.5 ratio to ~115M employer-sponsored self-insured lives
- Component C ($1.68B): above-peer-median admin fees, 8,180 plan Schedule C sample at 30% reducibility
- 30% reducibility floor is conservative: post-Tibble v. Edison 401(k) reform compressed retirement plan administrative fees by 20–30% over five years; Ohio Medicaid PBM reform compressed by 12–18% in year one
- Bias caveat: Schedule C is filed primarily by large, trust-funded plans. One cell (jumbo plans with mixed funding, 260 plans) accounts for 65% of in-sample lives and half the disclosed fees. Excluding the cell entirely drops Component C from $1.68B to $0.87B and the headline from $6.6B to $5.8B
- Broker-admin correlation: 425 plans filed both Schedule A and Schedule C. Pearson r = 0.030 (linear), Spearman rho = 0.199 (rank). Admin medians do climb 3.2× across broker quartiles, but within-quartile spread is much larger than the between-quartile shift, so broker rate is a weak predictor of admin fee — Components A and C measure different mechanisms and add without double-counting
- Range ceiling ($12.2B) reflects 50% reducibility on Component C plus a bounded extrapolation to the ~63 million ERISA-welfare participants in plans that file Form 5500 but skip Schedule C; the bounded extrapolation is described in the newsletter's CTA box but is not booked, pending matched plan-level claims data
- No overlap with prior issues: Schedule C dollars flow to TPAs, consultants, brokers, and lawyers — not to hospitals (Issue #3), PBMs (Issue #4), insurer underwriting (Issue #8), or general administrative complexity (Issue #5). See
results/overlap_matrix.mdfor the full accounting
Insurance companies use claim denials, prior authorization, and vertical integration as profit tools. We extracted per-contract prior authorization data from 93 Medicare Advantage contracts (61 UnitedHealthcare, 32 Humana) using the new CMS-0057-F transparency rule, covering 18.4 million prior authorization requests. UnitedHealthcare's volume-weighted denial rate: 13.5% (contradicting its headline "95.4% approved"). Per-contract variance: 0.7% to 25.2% (a 36× spread). Appeal overturn rates: 57.9% (UHC) and 64.7% (Humana). National extrapolation: approximately 3 million MA patients are denied entitled care every year and never appeal. Eliminating care suppression, vertical integration arbitrage, and AI-driven denial escalation would save approximately $32 billion per year.
Read the full analysis → issue_08/newsletter_issue_08.md
Reproducing the analysis
cd issue_08
# Build dataset from CMS-0057-F disclosures and SEC filings
python 01_build_data.py
# Generate analysis charts
python generate_all_charts.pyKey outputs:
results/— Per-contract denial rates, appeal analysis, savings modelfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS-0057-F Prior Authorization Transparency Rule Disclosures (April 2026) | Per-contract PA decision data for MA plans |
| UnitedHealth Group 10-K FY2024 | Revenue, operating margins, Optum segment |
| CVS Health, Elevance, Cigna, Humana 10-K filings | Insurer financials and MA enrollment |
| Health Affairs Nov 2025 | Optum vertical integration premium: 17% (61% in concentrated markets) |
| Stanford npj Digital Medicine Jan 2026 | AI increases denial rates 5–8 percentage points |
| AMA Physician Survey on Prior Authorization 2024 | 93% report PA delays care; 8% report PA contributed to death/disability |
| KFF CY2024 Part C PA Reporting Data | MA plan prior authorization volumes |
Key methodology notes
- Original CMS-0057-F per-contract extraction: 93 MA contracts, 18.4M PA requests
- Volume-weighted denial rates computed from per-contract data (not headline averages)
- Savings components: Care Suppression ($13.7B mid), Vertical Integration Arbitrage ($10.3B mid), AI Denial Escalation ($5.7B mid), Risk Adjustment ($0.3B)
- MLR Gaming ($11.8–20.7B) documented but excluded from booked savings
- Booked: $32B (upper-conservative within component ranges A+B+C+E); full range $22–37B
- Component D (deductible-delay extraction) described in the newsletter's MRI vignette but excluded from booked total pending matched patient-level claims + deductible-exposure data
- No overlap with Issue #5 (admin waste counts processing cost of PA; this counts the denied-care cost)
US GLP-1 spending grew from $57 million in 2018 to $71.7 billion in 2023, a 1,200-fold increase in five years. The US pays 3–5× more per dose than every other country buying the same drug. We built the first published 10-year budget projection of the CMS BALANCE Model, estimating $124 billion in cumulative Medicare costs to cover 4.6 million beneficiaries at negotiated prices ($245–350/month vs. $1,000+ retail). The $40 billion annual savings opportunity comes from aligning US GLP-1 prices with international levels through reference pricing and generic entry (semaglutide patent: December 2031).
Source: CMS BALANCE Model documentation, Peterson-KFF international drug pricing, Novo Nordisk and Eli Lilly SEC filings.
Read the full analysis → issue_07/newsletter_issue_07.md
Reproducing the analysis
cd issue_07
# Run the BALANCE Model projection (no downloads needed — all data hardcoded from published sources)
python 01_build_data.py
# Generate analysis charts
python generate_all_charts.py # Charts 1–4 (market growth, price comparison, cost projection, pricing structure)
python generate_chart5.py # Chart 5 (savings tracker)Key outputs:
results/balance_projection_all_scenarios.csv— 10-year enrollment and cost projections (LOW/MID/HIGH)results/sensitivity_analysis.csv— Model uncertainty driversresults/health_benefit_roi.csv— Cost-benefit analysisresults/international_prices.csv— US vs. international GLP-1 price comparisonresults/market_growth.csv— Historical GLP-1 market trajectory (2018–2025)results/key_metrics.json— Summary of all headline numbers and assumptionsfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS BALANCE Model Documentation (2026) | Negotiated prices, eligibility criteria, behavioral engagement requirements |
| CBO "How Would Authorizing Medicare to Cover Anti-Obesity Medications Affect the Federal Budget?" (Oct 2024) | Budget scoring framework |
| CDC NHANES 2023-2024 | Medicare-specific obesity prevalence (35.5%) |
| JAMA Network Open 2024 | US GLP-1 spending analysis ($71.7B in 2023) |
| Peterson-KFF Health System Tracker | International drug price comparisons |
| Novo Nordisk Annual Report 2024 | GLP-1 revenue ($26.0B), Wegovy/Ozempic financials |
| Eli Lilly SEC filings Q3 2025 | Tirzepatide revenue ($35-38B estimated), patent timelines |
| SELECT trial (NEJM 2023) | 20% reduction in major adverse cardiovascular events |
| STEP/SURMOUNT trials (NEJM 2023) | 16–22% mean weight loss |
| White House Section 232 Proclamation (April 2, 2026) | Pharmaceutical tariff structure and MFN exemption |
Key methodology notes
- Eligible population: 67.5M Medicare beneficiaries × 35.5% obesity × 55% behavioral engagement = 13.2M
- Enrollment ramp: 5% (2027) → 12% (2028) → 40% plateau (2031+), 88% annual retention
- Three pricing scenarios: LOW ($2,940/yr), MID ($3,600/yr), HIGH ($4,200/yr)
- 10-year cost: $86.2B (LOW) to $170.1B (HIGH); MID = $124.8B
- Health benefit ROI: 33% conservative (prevented CV events and T2DM progression)
- Savings: $40B/yr from 50–70% US GLP-1 price reduction through reference pricing and generic competition
- No overlap with Issue #2 (which covers 9 specific top-spend Medicare drugs) or Issue #4 (PBM extraction)
Original HCRIS FY2023 analysis of 5,480 hospitals (142M total discharges) reveals massive unexplained variance in per-discharge supply costs. National total: $170.9B across medical supplies ($40.4B), implantable devices ($48.7B), and drugs charged to patients ($81.9B). CMI-adjusted, bed-size-stratified P75/P25 ratios range from 2.5× to 3.4×: hospitals in the same size class and acuity tier spend wildly different amounts on supplies for equivalent patients. Bringing the highest-cost quartile down to the 75th percentile within peer groups would save approximately $28 billion per year.
Source: CMS HCRIS FY2023, 5,480 hospitals with ≥50 discharges and nonzero supply costs.
Read the full analysis → issue_06/newsletter_issue_06.md
Reproducing the analysis
cd issue_06
# Build dataset from raw HCRIS FY2023/FY2024 flat files
# Downloads ~200MB per year, extracts supply cost centers, computes CMI-adjusted variance
python 01_build_data.py
# Generate analysis charts
python generate_chart1_supply_variance.py # Supply cost variance by bed size (CMI-adjusted)
python generate_chart2_surplus_nonprofits.py # Medical surplus redistribution
python generate_chart3.py # Supply cost decomposition ($170.9B)
python generate_chart4.py # Ownership breakdown (for-profit vs nonprofit vs govt)
python generate_chart5.py # Implant price variance
python generate_chart7_state_ranking.py # 50-state supply waste rankingKey outputs:
results/expanded_analysis_results.json— State rankings, teaching analysis, FY2023 vs FY2024 comparisonfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS HCRIS HOSP10-REPORTS FY2023 | Cost reports for 5,480 hospitals; Worksheet A (supply cost centers), Worksheet S-2 (CMI), Worksheet S-3 (discharges/beds) |
| UCSF Health Supply Chain Optimization Study | 6.5% universal savings benchmark from supply chain standardization |
| Bernstein et al. 2024 | 6.8× cost variance in surgeon preference items for lumbar fusion |
| MATTER / Afya Foundation | Nonprofit medical surplus redistribution data |
Key methodology notes
- 5,480 hospitals analyzed (≥50 discharges, nonzero supply costs) from raw HCRIS FY2023 federal cost reports
- Three supply cost categories: medical supplies (Worksheet A, line 53), implantable devices (line 55), drugs charged to patients (line 56)
- CMI adjustment: per-discharge costs divided by hospital Case Mix Index to normalize for patient acuity
- Bed-size stratification: Small (<100), Medium (100–299), Large (300–499), Major (500+)
- Savings scenarios: Q4→P75 within peers = $28.5B (conservative); above-median→P50 = $58.9B; UCSF 6.5% universal = $11.1B
- Ownership: For-Profit 1,576 hospitals ($18.8B total), Nonprofit 2,993 hospitals ($128.9B total, 75.5% of national spend), Government 911 hospitals ($23.0B total)
- No overlap with Issue #3 (hospital pricing): that covers commercial-to-Medicare price ratios; this covers within-hospital purchasing efficiency
The US spends $4,983 per person just to administer healthcare — 5.6× the $884 average across ten peer nations. An original analysis of 4,518 hospital cost reports (CMS HCRIS FY2023) reveals a 6.2× variance in administrative overhead per discharge nationally. Even within same-size, same-acuity peer groups, the gap is 2.0–3.1×. Prior authorization alone costs the system $21–93 billion per year. Standardized billing, automated prior auth, and all-payer rate setting would save approximately $200 billion per year.
Source: CMS HCRIS FY2023, 4,518 hospitals with ≥100 discharges.
Read the full analysis → issue_05/newsletter_issue_05.md
Reproducing the analysis
cd issue_05
# Build dataset from raw HCRIS FY2023 flat files
# Downloads ~200MB, extracts admin/overhead cost centers from Worksheet A
python 01_build_data.py
# Generate all analysis charts from the hospital dataset
python generate_all_charts.pyKey outputs:
results/hospital_admin_costs_fy2023.csv— 4,518 hospitals, 22 columns (admin costs, overhead breakdown, payer mix)figures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS HCRIS HOSP10-REPORTS FY2023 | Cost reports for 4,518 hospitals; administrative overhead, A&G costs, total expenses |
| CMS National Health Expenditure Accounts 2023 | Total US healthcare spending $4.867T; admin share benchmarks |
| OECD Health Statistics 2023 | Per-capita admin spending across 10 peer nations |
| Woolhandler/Himmelstein 2020, Annals of Internal Medicine | US healthcare admin costs: $812B (2017), updated to $1.13–1.66T (2023) |
| AMA Prior Authorization Survey 2024 | 93% of physicians report PA delays care; 7% report PA contributed to patient death |
| Health Affairs Nov 2025 | Full-system PA cost: $93.3B/year (payers $6B, manufacturers $24.8B, physicians $26.7B, patients $35.8B) |
| Gaffney, Himmelstein, Woolhandler & Kahn 2023 | International admin cost comparison methodology |
Key methodology notes
- Four original analyses: (1) per-capita international comparison (US $4,983 vs 10-peer avg $884), (2) Woolhandler update to 2023 ($1.13–1.66T), (3) prior auth national cost ($21–93B), (4) HCRIS hospital admin variance
- HCRIS analysis: 4,518 hospitals (≥100 discharges, nonzero A&G), $141.2B total A&G, $268.5B total overhead (32.2% of total costs)
- P75/P25 ratio: 6.2× nationally, 2.0–2.7× within bed-size peers (CMI-adjusted: 2.1–3.1×)
- Ownership: For-profit median $458/DC, nonprofit $1,980/DC, government $2,524/DC
- Savings: Q4→P75 = $18.0B, above-median→P50 = $39.8B (hospital admin only; $200B includes full system)
- Published dataset:
results/hospital_admin_costs_fy2023.csv(4,518 hospitals, 22 columns) - No overlap with Issues #1–4: those cover drug prices, hospital procedure prices, and PBM extraction; this covers administrative overhead
Three companies — CVS Caremark, Express Scripts, and OptumRx — process 80% of the 6.6 billion prescriptions Americans fill each year. The Federal Trade Commission spent two years investigating their practices and documented billions in extraction through six distinct mechanisms: spread pricing, rebate opacity, specialty drug markup, formulary manipulation, self-preferencing, and independent pharmacy destruction. The Ohio state auditor found $224.8 million in spread pricing from a single state's Medicaid program in a single year. Eliminating these extraction mechanisms — through rebate pass-through mandates, fiduciary standards, and formulary transparency — would save approximately $30 billion per year.
Source: Drug Channels Institute 2024; Bernard & Sloan 2025.
Read the full analysis → issue_04/newsletter_issue_04.md
Reproducing the analysis
cd issue_04
# Generate analysis charts from cited federal data and academic literature
python chart1_pbm_market.py # PBM market concentration (Drug Channels Institute 2024)
python chart2_harm_spread.py # Spread pricing extraction mechanisms (FTC, Ohio Auditor)
python chart4_biosimilar_v4.py # Biosimilar adoption by state PBM law (CMS Part D, JAMA)
python chart5_insulin_prices.py # Insulin price trajectory (IQVIA, CMS)Key outputs:
results/biosimilar_analysis_2023.csv— CMS Part D biosimilar adoption dataresults/key_metrics.csv— Core PBM extraction metricsfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| FTC Interim Report #1 (July 2024) | $7.3B in PBM-owned specialty pharmacy markups, 2017–2022; $334B annual rebate flow |
| FTC Interim Report #2 (January 2025) | Vertical integration details and self-preferencing evidence |
| Ohio State Auditor (2018) | $224.8M spread pricing extracted from Ohio Medicaid in one year |
| Mattingly, Hyman & Bai 2023, JAMA Health Forum | Comprehensive review of PBM economics and agency conflicts |
| Drug Channels Institute 2024 | PBM market share: CVS 34%, ESI 24%, OptumRx 22% |
| Bernard & Sloan 2025, J Gen Internal Med | Total US prescription drug spending $722.5B (2023) |
| Kwon, Sarpatwari & Dusetzina 2025, JAMA Health Forum | Biosimilar adoption rates by state PBM law stringency |
| Chea, Sydor & Popovian 2023 | 57.4% of ESI formulary exclusions with questionable patient benefit |
| Knox, Gagneja & Kraschel 2021, JAMA Health Forum | 16.1% of rural independent pharmacies closed 2003–2018 |
| IQVIA National Prescription Audit | Manufacturer rebates: $334B annually paid to PBMs/plans |
Key methodology notes
- Savings model is conservative, built from six distinct non-overlapping mechanisms
- Mechanism 1 (spread pricing): $3.0B — Ohio audit extrapolated to national Medicaid managed care
- Mechanism 2 (rebate pass-through): $10.0B — PBMs retain estimated 5–10% above disclosed admin fees on $334B rebate pool
- Mechanism 3 (specialty markup): $1.5B — FTC-documented $7.3B over 5 years, annualized
- Mechanism 4 (formulary reform/biosimilar preference): $10.0B — biosimilar adoption gap vs. states with strong PBM laws
- Mechanism 5 (admin transparency savings): $5.5B — waste from opaque PBM reporting requirements
- Total booked: $30.0B/year (range $30–50B)
- No overlap with Issues #1–3: Issue #2 addresses manufacturer-level drug prices; Issue #4 addresses the intermediary extraction layer on top of those prices
- CAA 2026 (enacted Feb 3, 2026) includes rebate pass-through effective 2029 for commercial plans; FTC settled with Express Scripts Feb 4, 2026 ($700M/yr projected savings from one PBM)
Commercial insurers pay 254% of Medicare rates for identical hospital procedures. A hip replacement costs $29,000 in the US and under $11,000 in most peer nations. Capping commercial hospital payments at 200% of Medicare — the mechanism already used by Montana Medicaid and thousands of self-insured employers — would save approximately $73 billion per year.
Read the full analysis → issue_03/newsletter_issue_03.md
Reproducing the analysis
cd issue_03
# Build HCRIS cost report dataset and compute cost-to-charge ratios
python 01_build_data.py
# Generate analysis charts
python 02_visualize.pyKey outputs:
results/hospital_ccr_2023.csv— Hospital cost-to-charge ratios (3,193 hospitals)results/savings_calculation.json— Full savings model parametersresults/procedure_prices.json— International procedure price comparisonsfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS HCRIS HOSP10-REPORTS FY2023 | Cost reports for 3,193 hospitals; cost-to-charge ratios and operating costs |
| RAND Round 5.1 Hospital Pricing Study (2023) | Commercial insurer payments = 254% of Medicare for identical procedures |
| International Federation of Health Plans 2024-2025 | Procedure prices by country (hip replacement, bypass, etc.) |
| Peterson-KFF Health System Tracker | US vs. peer-nation procedure cost comparisons |
| CMS National Health Expenditure Accounts 2023 | Total US hospital spending $1.361T; private insurance share 38.8% |
| NASHP Montana Analysis (April 2021) | Independent evaluation of reference-based hospital pricing impact |
Key methodology notes
- Savings formula: $528B commercial hospital spend × 65% addressable × 21.3% price reduction (254%→200% of Medicare) = $73B
- 3,193 hospitals analyzed from raw HCRIS FY2023 federal cost reports
- For-profit hospitals: 4.11× median markup (highest); nonprofit: 2.46×; government: 2.22×. 37% of all hospitals charge 3× or more
- Correction (2026-03-17): Original release mislabeled CMS ownership codes, swapping nonprofit and for-profit categories. The $73B savings estimate was unaffected (derived from RAND/CMS NHE national data). See
issue_03/CTRL_TYPE_AUDIT.mdfor details. - Fix mechanism (Commercial Reference Pricing) is already implemented in Montana and by thousands of self-insured employers
- No overlap with Issues #1 or #2 (those cover drug prices only; this covers hospital/procedure prices)
Medicare pays 7–25× more than peer nations for the same brand-name drugs. International reference pricing — benchmarking Medicare negotiations against what Germany, France, Japan, UK, and Australia pay — would save approximately $25 billion per year.
Source: CMS Part D 2023 gross spend, Peterson-KFF 11-country OECD average prices. Savings = gross differential before rebate adjustment.
Read the full analysis → issue_02/newsletter_issue_02_FINAL.md
Reproducing the analysis
cd issue_02
# Build reference price dataset (NHS Drug Tariff + RAND international averages)
python 01_build_reference_data.py
# Generate analysis charts
python 02_visualize.pyKey outputs:
results/nhs_vs_medicare.csv— Medicare vs. NHS Drug Tariff price comparisonsresults/kff_drug_comparison.csv— 11-country OECD drug price benchmarksresults/rand_country_ratios.csv— RAND international price ratiosfigures/— All analysis charts
Data sources
| Source | Description |
|---|---|
| CMS Medicare Part D Spending by Drug (2023) | Gross drug spend and claim counts by drug name |
| NHS Drug Tariff Part VIIIA (March 2026) | UK generic reimbursement prices post-patent expiry |
| RAND RRA788-3 (Feb 2024) | International prescription drug price comparisons using 2022 data |
| Peterson-KFF Health System Tracker (Dec 2024) | 11-country OECD drug price benchmarks |
Key methodology notes
- Medicare figures are gross cost (pre-rebate) from CMS Part D Public Use File
- ~49% net rebate adjustment applied for top-spend brand drugs, triangulated from MedPAC and Feldman et al.
- NHS prices are post-patent generic reimbursement rates — representing the molecule's commodity price
- International average = Peterson-KFF 11-country OECD analysis
Medicare Part D pays prescription prices for drugs available cheaply over-the-counter. Step therapy reform — requiring OTC equivalents before prescription coverage activates — would redirect roughly $0.6 billion per year in unnecessary spending.
Read the full analysis → issue_01/newsletter_issue_01_FINAL.md
Reproducing the analysis
cd issue_01
# One-time environment setup
chmod +x 01_setup.sh && ./01_setup.sh
source .venv/bin/activate
# Download CMS Part D data (~200 MB)
python 02_download_data.py
# Build local DuckDB database
python 03_build_database.py
# Run analysis
python 04_analyze.py
# Generate analysis charts
python 05_visualize.pyKey outputs:
results/by_drug_2023.csv— Per-drug Medicare Part D spending and OTC price comparisonsresults/bene_overpayment_2023.csv— Beneficiary-level overpayment estimatesfigures/— All analysis charts
Data sources
| Source | URL |
|---|---|
| CMS Part D Spending by Drug (2023) | https://data.cms.gov/summary-statistics-on-use-and-payments/medicare-medicaid-spending-by-drug/medicare-part-d-spending-by-drug |
| JAMA — OTC Equivalents Study (Socal 2023) | https://pmc.ncbi.nlm.nih.gov/articles/PMC10722384/ |
| MedPAC Part D Report (2024) | https://www.medpac.gov/wp-content/uploads/2024/03/Mar24_Ch11_MedPAC_Report_To_Congress_SEC.pdf |
Key methodology notes
- OTC unit prices sourced from current retail at major US pharmacies (March 2026)
- 30-unit-per-claim approximation; see
issue_01/VALIDATION_REPORT.mdfor full methodology - Savings figures are conservative — do not account for PBM rebates or dispensing fees
Through 9 issues: ~$435.2 billion in identified savings (13.4% of the $3.24T gap)
We've identified $435.2 billion in fixable waste using free federal datasets. To go deeper, we need claim-level data that costs money to access: Medicare claims with diagnosis codes, all-payer state databases, hospital price transparency records, and legal research tools. Issue #8 made this concrete: the deductible-delay extraction mechanism described in the MRI vignette, where an insurer denial pushes a patient to cash and captures the deductible spread on the next claim, cannot be measured rigorously without paired patient-level claims plus deductible-exposure data. That is why Component D stays out of our booked total and why this fund exists.
Visit the AHC Data Access Fund → | Sponsor on GitHub →
Six datasets. Per-dataset crowdfunding via Stripe (no account required, any amount $5+). Your money is used only when a dataset is fully funded. Every contributor is listed publicly (or anonymously) on the fund page. Code is always open-source. Findings are always published. Holding licensed data (Truven/MarketScan, Optum Clinformatics, IQVIA Pharmetrics, Definitive Healthcare, Press Ganey, Sage Transparency)? Donate access — the fund page has a dedicated channel for proprietary dataset partnerships.
| Phase | Datasets | Cost | What It Unlocks |
|---|---|---|---|
| 1 | CMS Medicare Claims (5% sample) + Colorado All-Payer Claims | $3,500 | Patient-level denial outcomes, commercial vs. Medicare pricing, drug cost analysis |
| 2 | Hospital Discharge Data (CA+NY) + Price Transparency + Legal Research | $5,700 | Low-value care identification, real negotiated rates, antitrust case law |
| 3 | CMS Full Medicare (65M patients via VRDC) | $35,000 | The same data Harvard, Dartmouth, and RAND use. JAMA-publishable, congressionally-citable. |
Already have access to one of these datasets? We can collaborate directly. Your existing DUA + our published code = findings neither of us could produce alone. Get in touch →
Issue #10 examines the Procedure Mill: physicians say roughly one in five medical decisions in the US is unnecessary, and for procedures specifically the figure is closer to one in nine (Lyu, Wick, Cabrera, Makary, Overtreatment in the United States, PLOS ONE, 2017). The Lown Institute estimates $75 to $100 billion in low-value care annually. The Choosing Wisely initiative has named more than 600 procedures that evidence does not support for most patients. Medicare claims data lets us compute the volume, cost, and geographic distribution of overuse at the procedure level. Subscribe on Substack to get it when it drops.
Every analysis uses primary sources: CMS cost reports, Part D claims data, OECD health statistics, RAND pricing studies. Every number has a citation. Every script is reproducible from a clean clone. Caveats are named explicitly. The math is the argument.
No institutional affiliations. No university. No think tank. No funder who might find the analysis inconvenient. Funded by readers and data sponsors who want the numbers to be public.
Built by Andrew Rexroad. Questions, corrections, or data tips: vonrexroad@gmail.com