Is American Health Care Better Off? The State of US Medicine, AI and Biotech

· @Rich

Americans are living longer than ever and have better drugs on the way, yet the US spends about 2.5 times the peer-country average per person and ranks last of ten wealthy nations on a leading cross-country scorecard. Compared with ten years ago, outcomes and treatments are better, while coverage and affordability are mixed and, for many people, worse in 2026.

How the US compares with peer countries

The US is the world’s biggest health spender and gets less for it than most rich countries. The Commonwealth Fund’s Mirror, Mirror 2024 report, using 2021–2023 data and 70 measures, ranked the US 10th of 10 high-income countries. The US placed last on access, equity and health outcomes, 9th on administrative efficiency, and 2nd on care processes, its one strong category (Fortune on the report). I did not find a newer edition.

Measure United States Peer comparison Source
Health spending per person (2023) $13,818 OECD average $5,477 America’s Health Rankings
Health spending, share of GDP (2022) 16.5% 9.8% to 12.6% in the other nine countries Fortune on Commonwealth Fund
Life expectancy (2023) 78.4 years OECD average 81.1 years America’s Health Rankings
Infant deaths per 1,000 births (2023) 5.6 OECD average 4.2 America’s Health Rankings
People without insurance About 8% in 2025 The other nine countries have universal coverage Fortune on CDC data

Total US spending reached about $5.7 trillion in 2025, up 7.3% from 2024, and is projected to approach $9 trillion by 2034 (CMS figures via ACDIS). The two best-ranked systems in the Commonwealth Fund report, Australia and the Netherlands, were also among the lowest spenders.

Longevity and disease

US life expectancy hit a record 79 years in 2024, passing the pre-pandemic peak of 2019. Officials credit falling COVID-19 and drug overdose deaths (NPR). Preliminary CDC data show the 2025 death rate near 689 per 100,000, the lowest in more than a century of records, and down 22% since 2021 (CDC data via KRDO). That report does not state a 2025 life expectancy figure, so treat any headline number as unconfirmed.

The gap with peers is still wide. In 2023 the US ranked 30th of 38 OECD countries, 2.7 years behind the OECD average. That gap was 0.2 years in 2003 and peaked at 3.7 years in 2021 (America’s Health Rankings).

What kills Americans, per preliminary 2025 data:

  • Heart disease: nearly 695,000 deaths, the top cause.
  • Cancer: nearly 623,000 deaths.
  • Unintentional injuries, including overdoses, in third place. Overdose deaths were about 70,000 in 2025, down from nearly 80,000 in 2024.

The averages hide big differences. Death rates for Black and American Indian people run more than twice the rate for Asian people, and men die at a rate of 811 per 100,000 versus 583 for women (same CDC report).

Prescription drug prices

This is where the US gap is largest. A RAND study of 2022 data across the US and 33 other high-income countries found US prices were 278% of other countries’ prices across all drugs. For brand-name drugs, US list prices were 422% of comparison-country prices, and after manufacturer rebates the net gap was still more than threefold (RAND).

Generics are the exception. US generic prices were lower than in most comparison countries, and generics made up 90% of US prescription volume but only 8% of spending at list prices (same RAND report).

Two changes are narrowing the gap, slowly:

  • Medicare negotiation. Negotiated prices for the first 10 drugs, including Eliquis, Jardiance and Xarelto, took effect in 2026, at least 38% below 2023 list prices, with about $6 billion a year in Medicare savings. Fifteen more drugs, including Ozempic, follow in 2027 (Medicare Rights Center). The 2025 reconciliation law limits which drugs can be negotiated, which KFF estimates will raise Medicare spending by at least $5 billion (same source).
  • Cash-pay channels. Novo Nordisk announced Wegovy pills at $149 or $299 a month for cash payers depending on dose, and Wegovy injections at $349 (NBC News). The administration’s TrumpRx site is expected to link patients to such discounts.

Elderly care

Medicare does not pay for most long-term care, so families pay out of pocket until savings run low and Medicaid steps in (AP report on Genworth survey, 2016). The 2025 CareScout cost survey shows how large the bills are:

Care option Median cost Per year
Nursing home, private room $355 a day $129,575
Nursing home, semi-private room $315 a day $114,975
Non-medical home caregiver $35 an hour, 44 hours a week $80,080
Assisted living $6,200 a month $74,400
Adult day health care $95 a day, five days a week $24,700

Source: CareScout 2025 Cost of Care Survey. Assisted living rose 5% in a year, while adult day care fell 5%.

The practical options are to stay home with paid help, use adult day programs to ease the load on family caregivers, or move to assisted living or a nursing home. Each step up in support costs far more. Newer tools may help at the margins: Alzheimer’s blood tests and polypill-style 3D-printed medicines, covered below, aim at diagnosis and pill burden for older patients.

Are we better off than ten years ago?

It depends on what you measure. Health outcomes and treatments have improved. Costs and coverage have not, at least for people who buy insurance themselves or through an employer.

Better

  • Life expectancy is at a record 79 years, and the 2025 death rate is the lowest on record (NPR, CDC data via KRDO).
  • Medicare can now negotiate prices on some drugs, and 2026 prices on the first ten are at least 38% below 2023 list prices (Medicare Rights Center).
  • Patients have more options: cash-pay programs from drugmakers and retailers, new weight-loss pills, a first FDA-cleared Alzheimer’s blood test, and personalized gene editing (covered below).

Worse or uncertain

  • The average employer family premium reached $26,993 in 2025, up 26% in five years, with workers paying about $6,850 of it. The average single-coverage deductible rose 43% in ten years (KFF).
  • The enhanced ACA premium tax credits expired at the end of 2025. KFF projects marketplace enrollment falling from about 22 million to about 17 million in 2026 (KFF analysis via CT Public).
  • The CBO estimates the 2025 Medicaid changes could add 10 million uninsured people over a decade (Fortune). The uninsured rate was about 8% in 2025, before most of these changes bite.

People disagree about the causes and the fixes: some stress the cost to taxpayers and incentives, others stress lost coverage. I take no side here. On control, the clearest shift is toward cash-pay and direct-to-consumer options, which give some patients lower prices and more choice but sit outside insurance. I did not find consistent ten-year data on patient control, so I would not claim a trend.

AI and medicine: can it cut costs?

The best-evidenced savings are in paperwork, not in the exam room. A March 2025 task force report by the Peterson Health Technology Institute found that revenue-cycle work, meaning billing, coding, denials and prior authorization, costs an estimated 3% to 4% of net patient revenue. One study it cites projects that automation and AI documentation could cut financial transaction spending by 15% to 20%, or $15 billion to $20 billion a year nationally (PHTI). One health system automated 74% of its radiology coding and cut claim denials by 58.7%.

AI note-taking tools for doctors are spreading fast. The report found consistent signs of less burnout, including a 40% relative drop in one pilot, but mixed evidence of time savings and no proven cost savings. Risks include errors in AI-written notes and more thorough documentation that supports higher billing, which could raise spending (same report). Most evidence comes from early adopters, so this picture may have moved since 2025.

AI is also speeding up drug discovery. Insilico Medicine’s AI-designed pulmonary fibrosis drug, rentosertib, improved lung function by an average of 98.4 mL at the 60 mg dose over 12 weeks in a 71-patient phase IIa trial, while the placebo group declined by 20.3 mL. The company says early discovery took about 18 months and under $6 million, versus $100 million to $200 million traditionally. Those cost figures are company disclosures, the trial was small, and phase III trials still cost hundreds of millions (ai2.work summary).

My read: administrative AI could trim costs within a few years, drug discovery savings will take longer to reach prices, and patients will only see lower bills if insurers and hospitals pass the savings on.

Can medication be bioprinted?

It helps to separate two ideas. 3D-printed medication prints the pill itself. Bioprinting prints living cells into tissue. Printed pills are real; bioprinted tissue is mostly used to test drugs or, in the long run, to replace organs.

Printed pills. The FDA approved Spritam, the first 3D-printed drug, on August 3, 2015. It is levetiracetam for epilepsy, printed as a porous tablet that dissolves faster (NOVA). The bigger promise is personalization: custom doses and shapes, and multi-drug polypills that reduce pill burden for older patients. A University College London team printed warfarin tablets in different doses and shapes on smartphone-based printers, and they released the drug over 24 hours in a lab digestive-tract model. That was proof of concept only, with no human trials, and regulators still need a framework before anyone prints medicines at home (The Engineer). I found no evidence that personalized printing is routine in US pharmacies.

Bioprinted tissue. Printed tumor models that mimic a patient’s cancer, analyzed with AI, are an active research area for testing treatments (Artificial Intelligence in Health, 2026). The larger goal, bioprinted kidney and liver tissue, is funded by ARPA-H at roughly $25 million per team (Newswise, D Magazine).

Bottom line: yes for 3D-printed pills, one approved product in 2015 and personalization still in the lab; not yet for bioprinted medicine beyond research models.

Generics, biosimilars and cures on the horizon

Generics are the quiet success of US drug policy. The Association for Accessible Medicines, an industry group, reports that generics and biosimilars saved patients and the system about $467 billion in 2024. Generics were about 90% of prescriptions but only about 12% of drug spending. Biosimilars, the copies of biologic drugs, saved $20.2 billion in 2024 and $56.2 billion since 2015 (AAM).

Three candidates for something close to a cure:

  • Type 1 diabetes. In June 2025, Vertex reported that 10 people given the full dose of its stem-cell-derived islet therapy, zimislecel, no longer needed daily insulin after a year, and two others made some insulin but still needed it. All still needed immunosuppressive drugs. Vertex moved its planned FDA submission up to 2026, with availability possible as early as 2027 (diaTribe). I did not confirm the current filing status.
  • Genetic diseases. People have already received gene-editing therapies for sickle cell disease and high cholesterol, and personalized editing has treated at least one infant (next section) (MIT Technology Review).
  • Alzheimer’s diagnosis. The FDA cleared the first Alzheimer’s blood test in May 2025, for adults 55 and older who already have memory problems (Ideastream). That speeds diagnosis; it is not a cure.

GLP-1 weight-loss drugs are not cures either, but price competition is pushing monthly cash prices down. TrumpRx is expected to list injections starting near $350 a month, falling toward $250 within two years (NBC News, linked above).

DNA mapping, CRISPR and AI

The clearest proof that gene editing can be personal came from a baby. Kyle “KJ” Muldoon Jr. was born with a rare disorder that prevents his body from clearing toxic ammonia. A University of Pennsylvania team led by Rebecca Ahrens-Nicklas and Kiran Musunuru built a base-editing treatment, a form of CRISPR that fixes single DNA letters, tested it in human cells, mice and monkeys, and gave KJ a first dose at seven months old. By the time of the report he was doing well, though doctors say it is too early to judge how effective the approach is (MIT Technology Review).

The economics matter as much as the science:

  • The treatment cost about as much as a liver transplant, roughly $1 million. Musunuru expects a drop to a few hundred thousand dollars within a few years.
  • The FDA agreed on a trial protocol that could enroll as few as five patients, and agency officials described a new approval pathway for personalized therapies in the New England Journal of Medicine. The team hopes for approval within 3 to 5 years.

AI supplies the map. Google DeepMind’s AlphaGenome predicts how changes in the 98% of DNA that does not code for proteins affect gene activity, reading up to a million DNA letters at once (DeepMind). It is not yet validated for predicting an individual’s genome (GenomeWeb). The plausible chain is AI to find which variants matter, base editing to fix them, and a regulatory path built for one-patient-at-a-time drugs.

Preventive health

Prevention is where the US gap is most fixable. Among OECD countries, more than three million people under 75 died in a recent year from causes that could have been prevented or better treated (OECD via DevelopmentAid). The Commonwealth Fund found the US has the highest rate of such deaths among ten peers (Fortune).

New screening tools are arriving, with caveats:

  • Multi-cancer blood test. The NHS-Galleri trial gave Grail’s test to more than 142,000 volunteers aged 50 to 77 in England. It missed its main goal of reducing stage 3 and 4 cancers combined. But yearly screening substantially reduced stage 4 diagnoses, and the test was reported to be accurate and safe. Researchers say analysis will continue before health services decide how to use it (King’s College London).
  • Alzheimer’s blood test. Cleared in 2025 for people already showing memory problems, not as a screen for healthy adults (Ideastream, linked above).
  • Overdose and COVID deaths. The biggest recent prevention win was a drop in both, which drove the life expectancy rebound.

The practical advice is unglamorous: keep up with age-appropriate screening and vaccinations, and ask your doctor which tests fit your age and risk. No new test replaces that yet.

What is next in biotech and AI

The next 12 to 36 months have a few concrete markers:

  1. Insurer rate filings for 2027. They will show whether the 2026 coverage losses and premium spikes were a one-year shock or the start of a trend (KFF analysis via CT Public).
  2. Medicare prices for 15 more drugs, including Ozempic, in 2027 (Medicare Rights Center, linked above).
  3. A decision on zimislecel, the stem-cell therapy for type 1 diabetes, if Vertex files as planned.
  4. A new FDA path for bespoke gene editing, with approval of a KJ-style therapy hoped for in 3 to 5 years.
  5. First safety data on epigenetic reprogramming. Life Biosciences’ ER-100 trial for optic-nerve disease is expected to report early results from late 2026 into 2027, with safety as the main goal (Fortune, Live in the Future).
  6. Bioprinted kidney and liver tissue from the ARPA-H PRINT teams, funded in stages tied to milestones.
  7. Independent evidence on health-care AI, especially whether billing automation actually lowers what patients and insurers pay.

The pattern across all seven: science is moving faster than the system that pays for it. The biggest open question is not whether these tools work, but who gets them and at what price.

Sources

Pages opened October 30, 2026. Several are news summaries of primary reports; figures are as those pages reported them.

This article is general information, not medical, legal or financial advice.