All CMS-classified health research plus a deliberately high billed-charge envelope for the current solid-organ transplant volume. It omits most advanced delivery categories.
Healthcare spending · empirical decomposition
How much are we paying for the best medicine in the world?
America really does have extraordinary doctors, hospitals, research, and medical technology. The question is whether those capabilities explain a $5.279 trillion healthcare system.
The answer
A real strength. Not a $5 trillion explanation.
National accounts do not report an “exceptional medicine” category. High-end treatment is mixed into hospital, physician, and drug spending. So the honest answer is a transparent attribution range—not a fake point estimate.
$4.19
High-end capability
$95.81
The surrounding healthcare system
“Ordinary” does not mean wasteful or unimportant. It means the spending is not uniquely or unusually high-end. The central scenario is an allocation exercise, not a CMS category or a savings forecast. [1]
Three scenarios
What has to be assumed?
Adds 25% of all cancer care and 50% of domestic pharmaceutical R&D. Those fractions are explicit sensitivity assumptions—not observed allocations.
Credits all cancer care and all domestic pharmaceutical R&D, even though that includes routine oncology and research financed by foreign as well as U.S. revenue.
Actual capacity
How many people can the exceptional system treat?
The United States is a genuine leader in several categories. But even its clearest high-complexity volumes are small beside 340.1 million people and $5.279 trillion in annual spending.
Solid-organ transplants
Performed in 2024—0.01375% of the population. The United States led reporting countries at 143 transplants per million. [2][3]
Diagnoses at NCI centers
Seventy-four designated centers diagnose about 400,000 patients annually; the larger number treated is not reported. Diagnosis at a center does not mean every case is rare or uniquely complex. [4]
First CAR-T infusions
Reported to CIBMTR in 2023—0.00139% of the population. Actual paid national spending and foreign-patient counts are not published. [5]
TAVR procedures
Recorded at about 850 U.S. sites in 2024. This is very large capacity, although TAVR is also standard in wealthy peer systems. [6]
NICU admissions
Derived from 9.8% of 2023 live births. NICU care is advanced and lifesaving, but not uniquely American. [7]
Complex neuro and rare disease
No complete national series separates expert-center throughput from routine neurosurgery, spine care, chronic rare-disease management, or ordinary hospital care. NULL means unknown—not zero.
| Sector | Annual volume | Population share | What it proves | What it does not prove |
|---|---|---|---|---|
| Solid-organ transplant | 46,750 | 0.01375% | World-leading national throughput | National paid spending |
| NCI-designated cancer centers | ~400,000 diagnoses | 0.11761% | Large research-centered network | Complex-case volume |
| CAR-T | 4,721 | 0.00139% | Real advanced cell-therapy delivery | Every U.S. infusion or net price |
| TAVR | 106,000+ | 0.03117% | Large structural-heart capacity | Uniqueness to the United States |
| Congenital cardiac surgery | ~36,230/year | 0.01065% | Large North American registry | U.S.-only volume |
| Active interventional trials | 30,875 studies | Not applicable | Very large trial footprint | Unique patients treated |
Foreign patients
People do come. The national footprint is tiny.
~186,096
2024 air visitors who named health treatment as their main trip purpose
$1.140B
health-related travel exports—about 0.0216% of NHE
These federal series confirm inbound medical demand. They do not identify wealth, hospital, procedure, routine versus complex care, or trial participation. The visitor estimate excludes land arrivals; the BEA expenditure measure includes travel consumption, not just medical bills. [8][9]
Where the money goes
The accounting table reconciles exactly to the CMS total.
These rows are mutually exclusive. High-complexity care, provider administration, low-value care, and excess prices are embedded inside them. Adding those overlays to this table would double-count.
Hospital care
30.97%
Physician and clinical
21.02%
Long-term, residential, and home care
13.45%
Other personal healthcare
11.16%
Retail prescription drugs
8.85%
Explicit administration and insurance overhead
7.05%
Structures and equipment
3.19%
Government public health
2.99%
Research
1.33%
CMS 2024 historical NHE: $5,278.588 billion, or $15,474 per person. Commercial manufacturer R&D is excluded from the NHE research row. [1]
Download NHE decomposition · CSVInternational test
Advanced medicine exists in countries that spend far less.
The ten comparison countries all maintain advanced hospitals, specialists, research, clinical trials, and transplantation. Their 2024 unweighted average was $7,738 per person and 10.97% of GDP. The United States spent $14,885 and 17.16%. [10][11]
United States$14,885
Switzerland$9,963
Germany$9,365
Netherlands$8,436
Sweden$7,871
Australia$7,469
France$7,367
Canada$7,301
Denmark$7,071
United Kingdom$6,747
Japan$5,790
OECD’s standardized 2023 measures also reject a one-cause story: the U.S. health-sector price index was 55.7% above the ten-peer average, while price-deflated volume and mix were 34.9% higher. Both prices and the amount/mix of care matter. [12]
Download peer comparison · CSVInnovation
The strongest case deserves a real answer.
Higher expected global revenue does increase the incentive to fund risky drug development; major revenue cuts would reduce some future launches. But sunk R&D does not mechanically set a launched drug’s price. NIH-funded science matters causally, foreign revenue matters, and other countries contribute researchers, trials, public funding, manufacturing, and purchases. The evidence supports “some innovation at the margin,” not “the world’s innovation would collapse.” [16]
Do not double-count
Administration, prices, low-value care, and fraud overlap.
| Evidence overlay | Estimate | How to read it |
|---|---|---|
| Explicit CMS administration and insurance overhead | $372.132B (2024) | Mutually exclusive NHE row; provider administration is elsewhere |
| Broad insurer + provider administration | $812B (2017) | Embedded across hospital, physician, long-term care, and insurance rows [17] |
| Low-value/overtreatment | $75.7B–$101.2B (2019 review) | Embedded in clinical rows; not an additive category [18] |
| Medicare FFS/C/D + Medicaid improper payments | $85.45B (FY2024) | Includes documentation and payment-rule errors; not equivalent to fraud [19] |
| Private-plan hospital prices | 254% of Medicare (2022) | Same-service price comparison, not a national savings estimate [20] |
| U.S. gross brand-drug prices | 422% of 33-country comparison (2022) | Before confidential rebates; generics differ [21] |
The $72 trillion question
$72.135 trillion is real—but it is a projection.
It is the sum of CMS’s ten projected nominal years from 2025 through 2034. The central scenario attributes $3.019 trillion of that decade to unusually high-end capability. The conservative scenario attributes $1.444 trillion; the high scenario, $6.021 trillion.
Method
What this estimate is—and is not.
- Start with the exact 2024 CMS National Health Expenditure Accounts.
- Count actual patients and procedures where national registries exist.
- Keep missing national figures NULL rather than turning anecdotes into capacity.
- Treat administrative cost, excess-price, low-value-care, and improper-payment studies as overlapping evidence—not additive buckets.
- Construct three inspectable attribution scenarios and disclose every assumption.
- Compare the result with ten wealthy countries using one OECD vintage and consistent measures.
The underlying calculation is reproducible, and every machine-readable figure carries a source URL, year, status, confidence level, and limitation. [22]
Sources and citations
Primary data first.
- CMS, National Health Expenditures. Historical 1960–2024 and projections 2025–2034. Download official data.
- HRSA/SRTR, 2024 Annual Data Report. U.S. organ transplantation totals, organ chapters, donors, and waitlists. Open report.
- WHO–ONT Global Observatory on Donation and Transplantation, 2024. International transplant activity. Open peer-reviewed report.
- National Cancer Institute. Seventy-four designated centers and approximately 400,000 annual diagnoses at those centers. Open NCI page.
- CIBMTR. 2023 U.S. hematopoietic cell transplantation and cellular therapy activity, including CAR-T. Open report.
- ACC/STS TVT Registry. More than 106,000 TAVR procedures in 2024. Open registry presentation.
- CDC/NCHS Data Brief 525. National NICU admission rate, 2016–2023. Open report.
- National Travel and Tourism Office, SIAT. 2024 inbound air-visitor totals and health-treatment purpose shares. Open data page.
- Bureau of Economic Analysis. Revised 2024 international transactions, including health-related travel exports. Open release.
- OECD, Health at a Glance 2025. 2024 health expenditure per person in PPP dollars. Open data.
- OECD, Health at a Glance 2025. 2024 health expenditure relative to GDP. Open data.
- OECD, Health at a Glance 2025. 2023 comparative health-sector prices and volume per person. Open data.
- NIH Data Book. FY2024 budget authority. Open data.
- NSF NCSES, 2023 BERD. Business R&D performed in the United States, including pharmaceuticals and medicines. Open data.
- FDA CDER. Fifty novel drug approvals in 2024. Open approvals.
- Congressional Budget Office. Pharmaceutical R&D, expected revenue, and drug development. Open report.
- Himmelstein, Campbell, and Woolhandler. U.S.–Canada administrative-cost comparison using 2017 data. Open abstract.
- Shrank, Rogstad, and Parekh, JAMA 2019. Review of six overlapping healthcare-waste domains. Open study.
- CMS FY2024 Improper Payments Fact Sheet. Medicare and Medicaid estimates and explicit warning that improper payments are not all fraud. Open fact sheet.
- RAND Hospital Price Transparency Study, Round 5. Private-plan payments relative to Medicare in 2022. Open report.
- RAND International Prescription Drug Price Comparison. U.S. gross drug prices relative to 33 OECD countries in 2022. Open report.
- AFWH reproducible research package. Machine-readable inputs, outputs, source URLs, confidence labels, and scenario methods. Download evidence file.
Bottom line
The best estimate is about $4 of every $100.
A wide, honest range is $2–$8. The remaining $92–$98 buys the broader healthcare system around America’s exceptional medical nodes. Frontier capability is real. It is not a plausible dominant explanation for frontier spending.