US Healthcare Costs: Why Americans Are Delaying or Skipping Care

OFFICIAL DATA. The Centers for Medicare & Medicaid Services’ National Health Expenditure Accounts — the government’s own official count, and the primary source this article uses for national totals — put total US health spending at $5.3 trillion in 2024, the most recent full year of actual (not projected) data, up 7.2% from 2023. That works out to $15,474 per person and 18.0% of GDP, up from 17.7% the year before. CMS’s own NHE Projections series estimates spending could reach roughly $8.6 trillion (about $24,200 per person) and 20.3% of GDP by 2033 — a projection under current trends, not a fact about the future.

These are three genuinely different numbers, and conflating them is one of the most common errors in healthcare-cost reporting: national health expenditure is the sum of everything spent on healthcare in the US by patients, employers, insurers and government combined. Per-capita spending divides that total by population — it is not what any individual actually pays. Out-of-pocket spending is the much smaller slice a patient personally hands over: premiums, deductibles, copays and coinsurance. A hospital’s list price is yet another, separate figure — rarely what an insured patient’s plan actually pays after negotiation. This article keeps those categories distinct throughout.

No single factor explains US healthcare spending. Health economists generally separate the question into three parts: price (what each service costs), utilization (how much care people use) and mix (which services and settings people use). Research comparing the US with peer countries consistently finds Americans do not use dramatically more healthcare by volume — fewer physician visits per person than many peer countries, in fact. What differs is price: the same MRI, the same birth, the same appliance costs several times more in the US than in comparable wealthy countries, largely because prices are negotiated privately between providers and insurers rather than set or capped by a national system.

Layered on top of price: administrative complexity from a multi-payer insurance system with thousands of distinct plans and billing rules; hospital and physician-group consolidation, which research links to higher negotiated prices in concentrated markets; facility fees charged by hospital-owned outpatient clinics for services a stand-alone office would bill less for; specialist-heavy, technology-intensive care patterns; and prescription drug prices set by manufacturers with far less government negotiating power than in most peer countries, historically. AiTimeline analysis: the honest summary is that US healthcare is expensive mainly because of what each unit of care costs, not primarily because Americans use much more of it — a distinction that matters because “use less care” and “pay less for the same care” are different policy problems.

This is the fact this article is built around: a person can be insured and still face a real financial barrier to care. Understanding why requires knowing what each insurance term actually means.

Premium — the amount paid, usually monthly, just to have coverage, regardless of whether care is used. Deductible — the amount a patient pays out of pocket before the plan starts paying its share for most services. Copay — a fixed dollar amount paid per visit or prescription (e.g., $30 for a primary-care visit). Coinsurance — a percentage of the cost the patient pays after the deductible is met (e.g., 20%) until an out-of-pocket maximum is reached. Out-of-pocket maximum — the most a patient pays in a plan year before the insurer covers 100% of covered costs. Network — the group of providers a plan has negotiated rates with; out-of-network care is typically billed at higher, less-protected rates.

Illustrative example — not an average US deductible: a worker enrolls in an employer plan with a $3,000 individual deductible. In February they need an MRI. Before insurance pays a substantial share, they owe the negotiated rate for that MRI, up to $3,000, out of pocket — even though their premium is deducted from every paycheck and their coverage is technically active. That gap between “covered” and “affordable in the moment” is where the delay decision happens.

For scale, not as a universal figure: the average deductible for a single employer-sponsored plan reached $1,886 in 2025, up 17% since 2020’s $1,617, and 34% of covered workers now carry a deductible of $2,000 or more (KFF 2025 Employer Health Benefits Survey). None of these numbers describe every plan — some employer plans and many ACA marketplace “gold” or “platinum” tiers carry far lower deductibles in exchange for higher premiums.

SURVEY. In its Health Tracking Poll, KFF asked a nationally representative sample of US adults: “In the past 12 months, have you skipped or postponed getting the health care you needed because of the cost?” 36% answered yes — up from 25% when KFF asked the identical question in August 2023. The poll was fielded May 5–26, 2025, among 2,539 US adults (2,444 in English, 95 in Spanish), with a margin of error of ±3 percentage points for the full sample. The finding appears in KFF’s brief “Americans’ Challenges With Health Care Costs,” last updated April 30, 2026.

What this figure does and doesn’t show: it measures self-reported avoidance of needed care over cost, across any type of care (KFF’s broader release also breaks this down by gender, with women reporting higher rates than men). It does not mean 36% of Americans never see a doctor, and it is not a measure of the uninsured population specifically — KFF’s data shows insured adults report this too, just at a lower rate than the uninsured. It should never be shortened to “36% of Americans don’t see doctors” — that materially misstates what was asked.

OFFICIAL DATA / SURVEY. USAFacts reports that 26% of American adults skipped some form of medical treatment in 2025 because they could not afford it — down slightly from 28% in 2024. This figure traces back to the Federal Reserve’s Report on the Economic Well-Being of U.S. Households in 2025 (the “SHED” survey), fielded by Ipsos October 17–28, 2025 among 12,934 adults, published May 13, 2026. USAFacts’ explainer, last updated May 18, 2026, breaks down what was skipped: dental care (18%), doctor or specialist visits (15%), follow-up care (10%), mental healthcare (10%) and prescription medicine (9%).

This is a genuinely different measure from KFF’s 36%: the Fed/SHED question asks about “skipping some medical treatment” specifically framed around affordability in the past year (2025 calendar year), while KFF asks about “skipping or postponing needed care” over the prior 12 months from when the poll was fielded (mid-2024 to mid-2025). Different wording, different field periods, different sponsoring organizations, similar territory. Neither number is wrong, and they are not contradictory — they are two independent instruments converging on the same underlying reality: roughly a quarter to more than a third of US adults report going without some care they needed because of what it would cost.

The gap is large and real — insurance is doing genuine protective work. But 24% of insured adults still skipping needed treatment over cost is not a rounding error; it is tens of millions of people for whom a card in their wallet did not fully solve the affordability problem. AiTimeline analysis: the affordability problem and the uninsured problem are related but not identical — closing the coverage gap alone would not close the cost-avoidance gap.

A deductible is not a design flaw; it is a deliberate mechanism to keep premiums lower by shifting first-dollar cost to the patient. The tradeoff has intensified: the average single-coverage deductible for employer plans rose from $1,617 in 2020 to $1,886 in 2025, a 17% increase, and roughly a third of covered workers now face a deductible of $2,000 or more (KFF 2025 Employer Health Benefits Survey). High-deductible health plans, often paired with tax-advantaged Health Savings Accounts (HSAs), have become one of the most common employer plan designs precisely because they lower the visible premium — while raising the amount owed at the point of care.

ACA marketplace plans add a second layer of variation: bronze-tier plans typically carry the lowest premiums and highest deductibles, while gold and platinum tiers reverse that tradeoff. Not every deductible is unaffordable — for a household with savings or an HSA balance, a $1,886 deductible is a manageable, budgeted cost. For a household living paycheck to paycheck, the same number can be the entire reason a visit gets pushed to “next month.”

Prescription drugs made up roughly $467 billion of US health spending in 2024 — about 9% of total national health expenditure, up 7.9% from the prior year (CMS). Cost exposure varies enormously by drug type: generics are typically inexpensive; brand-name drugs, especially those without generic competition, can carry copays, coinsurance percentages or full pre-deductible list prices that patients on maintenance medications feel every single month, not just once.

POLICY. Two federal changes are actively lowering some of this exposure in 2026. Under the Inflation Reduction Act, Medicare negotiated prices for the first 10 selected drugs took effect January 1, 2026, at discounts of roughly 38% to 79% off 2023 list prices. Medicare’s Part D annual out-of-pocket cap sits at $2,100 in 2026, and insulin for Medicare beneficiaries remains capped at $35 per month. These changes apply to Medicare, not to every insured American — employer and marketplace plan formularies, pharmacy networks and coinsurance structures still vary widely outside Medicare, and this article does not offer guidance on any individual medication.

SURVEY / OFFICIAL DATA — note the range. Medical debt is one of the hardest healthcare-cost figures to pin to a single number, because different sources measure different things. A widely cited KFF/NPR analysis found roughly 41% of US adults — more than 100 million people — carry some form of medical or dental debt in the broadest sense (including informal debt like credit-card balances or family loans used to cover a bill); that analysis is now several years old and should be read as a general order of magnitude, not a current-year figure. A narrower, more recent measure — medical debt actually reported on credit files — covers a smaller share, roughly 9% of adults (about 23 million people), totaling upward of $220 billion in collections by some counts, with broader household-liability estimates running considerably higher.

POLICY / COURT RULING. A regulatory attempt to shrink that credit-report exposure did not survive intact: the Consumer Financial Protection Bureau finalized a rule in January 2025 to remove medical debt from credit reports entirely, but a federal court in the Eastern District of Texas vacated that rule on July 11, 2025. As of this update, medical debt can still appear on a credit report in most states without that federal protection, though some states have passed their own versions. Medical debt matters beyond the balance itself: it can affect credit access, housing applications and, per KFF’s own research, future willingness to seek care — a feedback loop, not a one-time event.

A commonly repeated claim is that delaying care always costs more later. The honest version is narrower: delaying care can sometimes lead to more advanced disease, a costlier intervention or an emergency-department visit — but this depends heavily on the condition, and is not universal. A postponed cholesterol screening carries a different risk profile than a postponed evaluation of chest pain; a delayed refill of a maintenance medication is different again. Research on delayed care and outcomes generally supports the pathway existing for specific conditions — particularly ones that progress silently, like some cancers and cardiovascular disease — without supporting it as an across-the-board rule for every visit skipped.

Emergency departments exist to treat conditions that cannot wait, and the large majority of ER visits reflect exactly that. But research and clinician reporting also describe a secondary pattern: some patients who delayed primary or specialist care eventually arrive at an emergency department once a condition has progressed — not because ER care is inappropriate for them at that point, but because it became the point where care could no longer be postponed. Emergency care is typically billed at a higher rate than an equivalent primary-care visit, and for patients whose insurance network or coverage status complicates the bill, an ER visit can compound the exact affordability problem that caused the delay in the first place. None of this implies ER use is generally inappropriate — only that it is one visible endpoint of a cost-driven delay pattern for a subset of patients.

A “surprise medical bill” happens when a patient receives care from an out-of-network provider — often unknowingly, such as an anesthesiologist at an in-network hospital — and gets billed the difference between what their insurer paid and what that provider charged. The federal No Surprises Act, effective January 1, 2022, bans this practice for emergency care and for most out-of-network clinicians working at in-network facilities, capping what a patient can be charged at the in-network rate. It also requires “good faith estimates” for uninsured and self-pay patients. These patient-facing protections are still in effect.

COURT RULING. What changed behind the scenes: on August 12, 2026, the full, en banc 5th US Circuit Court of Appeals (17 judges, New Orleans) ruled in Texas Medical Association v. HHS that federal regulators acted unlawfully in how they defined the “qualifying payment amount” (QPA) — the benchmark rate insurers and independent arbitrators use to settle payment disputes with out-of-network providers under the Act. The court found regulators improperly let insurers include so-called “ghost rates” (contracted rates for services that were never actually performed) in the QPA calculation while excluding certain bonus and incentive payments — a combination the court said systematically understated real contracted rates.

What this ruling does: it invalidates the current QPA calculation methodology and sends federal agencies back to rewrite it. What it does not do: it does not remove a patient’s protection from surprise bills, and the court itself said the existing formula stays in interim use while a replacement is written, to avoid “all-out chaos.” The practical stakes are between insurers and providers, not directly billed to patients: provider wins in No Surprises Act arbitration disputes already surged from $4.1 billion in awarded amounts in 2024 to $14.9 billion in 2025, with providers prevailing in roughly 85% of disputes — a trend this ruling is likely to reinforce, which could indirectly influence negotiated rates and premiums over time, though that effect is not yet measured.

Federal rules now require hospitals to publish machine-readable files of their negotiated rates and a consumer-facing shoppable-services tool, alongside the No Surprises Act’s good-faith-estimate requirement for the uninsured. POLICY. CMS’s price-transparency enforcement was tightened for 2026: a new Calendar Year 2026 rule took effect January 1, 2026, with stricter enforcement — including a new standardized data format, a named-executive attestation requirement and provider ID tagging — beginning April 1, 2026.

AiTimeline analysis: price transparency has neither failed completely nor solved the affordability problem. Published rate files have made systematic price research possible for researchers, employers and journalists, and have documented enormous, hard-to-justify price variation for identical procedures within the same city. But they help an individual patient only if that patient has realistic alternative providers to choose between, time to compare a spreadsheet of codes before an urgent or emergency need, and a plan structure where the comparison actually changes what they owe. For a patient with one nearby hospital and one in-network cardiologist, transparency data is informative but not actionable in the moment a decision has to be made.

Healthcare costs do not sit in a separate household budget line from everything else. Survey evidence on financial hardship consistently finds medical bills named alongside rent, groceries, utilities and transportation as competing claims on the same limited monthly income — and unlike most of those expenses, a medical bill can arrive as a single large, unpredictable shock rather than a steady, budgetable cost. This is one reason cost-driven care avoidance concentrates among lower- and middle-income households: the deductible is not competing against discretionary spending, it is competing against the next month’s essentials.

Preventive services — many routine screenings, checkups and vaccinations — are treated differently under federal insurance rules than most other care: ACA-compliant plans are generally required to cover a defined list of preventive services with no cost-sharing, meaning no deductible or copay applies. This matters directly for the cost-avoidance story, because it means some of the care most useful for catching problems early is, by policy design, not supposed to carry the same financial barrier as a diagnostic or treatment visit — though plan compliance, coding and network issues can still create out-of-pocket surprises in practice. This article does not recommend specific screenings; discuss what’s appropriate for you with a healthcare professional.

Money moves through the US healthcare system in a longer chain than a single bill suggests. A patient (or their employer, on their behalf) pays a premium into an insurance plan. The insurer negotiates rates with and pays a healthcare provider for services rendered, while the patient separately covers their deductible, copay and coinsurance directly to the provider or plan. In parallel, government programs — Medicare, Medicaid, and ACA marketplace subsidies funded by federal and state taxes — pay providers directly for enrolled populations, or subsidize private premiums. Providers, in turn, pay pharmaceutical and medical-device suppliers for the drugs and equipment used in care. Every one of premiums, taxes, deductibles, copays and coinsurance is a distinct payment stream, often flowing simultaneously for the same episode of care — part of why a single medical bill can be so difficult for a patient to parse.

OFFICIAL DATA. The Commonwealth Fund’s 2024 Mirror, Mirror comparison — its most recent edition, published September 19, 2024 — ranked the US last of 10 high-income countries (Australia, Canada, France, Germany, the Netherlands, New Zealand, Sweden, Switzerland, the UK and the US) on overall health-system performance, including last on equity, health outcomes and access to care, despite spending the most of any country studied — more than 16% of GDP in 2022 versus roughly 8–12% for its peers. OECD data separately puts US per-capita health spending at roughly $14,885 (PPP-adjusted), the highest in the OECD. US life expectancy stood at about 79.0 years in 2024, versus a peer-country average of roughly 82.7 years.

This is not simply “other countries have free healthcare.” Peer countries finance care through some mix of general taxation, mandatory social-insurance contributions and regulated private insurance — the cost is paid collectively through taxes rather than concentrated at the point of care through deductibles and coinsurance, and most peer systems negotiate or regulate prices nationally in ways the fragmented US system generally does not. The tradeoff is real: broader financing pools reduce individual point-of-care exposure, but come with their own tax burdens, wait-time dynamics and coverage-design choices that vary significantly by country.

POLICY. ACA marketplace: enhanced premium tax credits, in place since 2021, expired at the end of 2025. KFF reports 2026 marketplace enrollment fell 4.9% to 23.1 million, down from a record 24.3 million in 2025, as average net premium payments for enrollees who stayed covered rose roughly 58%, from $113 to $178 a month. Separately, the Congressional Budget Office had projected the subsidy expiration could push some enrollees’ costs up around 114% on average and result in 7.3 million fewer marketplace enrollees, 4.8 million of whom become uninsured — a forecast, not the final observed number, and one KFF’s actual reported enrollment figures only partially track.

POLICY. Medicaid: the One Big Beautiful Bill Act, signed July 4, 2025, introduces 80-hours-per-month work or “community engagement” requirements for many non-disabled adult Medicaid enrollees, with states required to implement by January 1, 2027. The Congressional Budget Office estimates up to 7.8 million people could lose Medicaid coverage from this requirement combined with more frequent eligibility checks and reduced federal matching funds.

REPORTED. Hospital finances: hospital-industry advocacy groups have begun projecting that coverage losses from the changes above will raise uncompensated-care costs — one cited estimate from the Greater New York Hospital Association projects roughly $1.3 billion a year in added uncompensated care for New York hospitals alone. This is a single state’s advocacy estimate, not an audited national figure, and should be read as “reported” and “projected,” not as settled fact.

COURT RULING. Surprise billing: the 5th Circuit’s August 12, 2026 ruling on the No Surprises Act’s QPA formula (detailed above) sends part of the billing-dispute rulebook back to federal agencies for revision.

POLICY. Price transparency: CMS’s tightened hospital price-transparency enforcement began April 1, 2026.

AiTimeline analysis. Total national spending, per-capita cost and even the uninsured rate are all important, but none of them is the single most revealing signal in this data. The more revealing question is not “what does healthcare cost?” but “what do people stop doing because it costs too much?” A system can spend more than any other country in the world — which the US does — and still generate a rising share of adults who skip a needed appointment, delay a prescription refill or avoid a follow-up visit. That behavioral shift, documented independently by KFF and the Federal Reserve in 2025 data, is a more direct measure of lived affordability than any aggregate spending total, because it captures the point where cost actually changes what a patient does.

Artificial intelligence is already being deployed against several cost drivers named earlier in this article, mostly on the administrative and access side rather than the clinical-price side. Documented and emerging use cases include: automated appointment scheduling and reminders (reducing missed-visit and rebooking costs); AI-assisted clinical documentation (reducing physician administrative time, one contributor to overhead built into billed rates); expanded telehealth, which can lower the cost and logistical barrier of a primary-care touchpoint; AI-assisted diagnostic support tools used alongside clinician judgment; patient-navigation tools helping people find in-network, lower-cost providers; and billing-automation and price-transparency tools that make published rate-file data more usable for patients and employers.

The risks are real and specific, not hypothetical: AI clinical tools can produce incorrect or biased recommendations, particularly for underrepresented patient populations in training data; automated systems raise genuine patient-privacy questions; over-reliance on an AI tool without clinician oversight risks misdiagnosis; and unequal access to AI-enabled telehealth or navigation tools could widen, not narrow, existing affordability gaps for patients without reliable internet access or digital literacy. AI tools should not be treated as a substitute for professional medical evaluation — nothing in this section is a recommendation to use an AI tool in place of seeing a qualified healthcare professional.