Ground ambulances were carved out of the No Surprises Act — the committee that replaced the rule filed its report without naming a payment multiple

What is happening?

The No Surprises Act built a federal balance-billing regime for hospitals, clinicians and air ambulances. Ground ambulances sit outside it. On the freshest national estimate available, 85 percent of emergency ground ambulance transports were delivered out of network, measured on commercial claims from 2014 through 2017, and 28 percent of all such transports produced a potential surprise bill over the same period, the smaller share sitting inside the larger one rather than beside it. Both figures were still being cited as current in 2026 because no newer national estimate has been published.

The carve-out is affirmative statutory text, not a silence. Section 117 of the Act, enacted as part of the Consolidated Appropriations Act, 2021 and signed 2020-12-27, handled ground ambulances by ordering an advisory committee within 90 days and a report within 180 days of its first meeting. It contains no prohibition on balance billing, no payment standard, no rulemaking authority for the Secretaries, no deadline for Congress to act on what the committee recommended, and no sunset. The whole operative demand of the section is that a document be produced.

The document was produced. The report bears a cover date of 2024-03-29, and CMS states it was issued to the Secretaries 2024-08-28. In fiscal year 2026 the committee held zero meetings, issued zero reports, recorded $0.00 across every cost line and 0.00 federal staff FTE, and the agency recommendation for the next fiscal year is Terminate. As of 2026-08-08 the CMS page reads Committee Status - Currently Inactive, and no federal ground ambulance balance-billing statute is in force.

Whose problem is this?

The body convened to settle the question was an advisory committee, and an advisory committee discharges its duty by delivering a report. Everyone who could act on it afterward was left free to do nothing, and doing nothing violates no provision.

Where does this problem end?

The boundary matters here because the federal regime already exists and only one service was written out of it. What is missing is not a diagnosis and not a draft but a number that nobody has been willing to name.

What is the state now, and what should it be?

Now

Needs a new measurementthe target state: Section 117 attached no target figure, no deadline for Congress and no success metric to the outcome, and no source opened here publishes what a solved state would look like numerically. The committee did describe the shape of a fix — a balance-billing prohibition coupled with a guaranteed minimum payment, patient cost sharing capped at the lesser of $100 or 10 percent of the required payment and applied before the deductible, a ban on billing the patient before the claim reaches the insurer, and payment within 30 days of a clean claim — but it declined to name the payment multiple that would make that shape operative.

How big is it?

Roughly 840,000 to 2,550,000 privately insured people a year. The base is about 3 million privately insured patients transported to emergency rooms by ground ambulance each year, from a page published 2021-06-24 that does not state its data year at all. Applying the two shares from the 2014-2017 study gives 840,000 at the 28 percent potential-surprise-bill share and 2,550,000 at the 85 percent out-of-network share.

Carry the vintage wherever the range travels. Neither input supports an as-of 2026 statement, so the honest form is on the order of one to two and a half million privately insured people a year, computed from a base of unknown vintage and shares from 2014 through 2017.

The width of the range is not statistical uncertainty. It is the distance between two different questions — how many were billed above the allowed amount, and how many had no negotiated rate standing between them and the billed charge. A separate 2021 series covering all ground trips under employer plans puts the out-of-network share at 35 percent, which would place a midpoint near 1.05 million, but it is not substitutable because the denominator and the service mix are different.

The 22 to 24 states with laws are deliberately not netted out. Netting would need per-state scope and per-state fully insured population, and neither was retrievable. Because no state law reaches self-funded plans, a correct netting would remove far less than the state count suggests, and reporting a netted figure would be false precision.

What this number cannot count. Non-emergency ground transports, which are excluded from the base and are the exact category on which the committee finalized nothing. Responses where no transport occurs, which are roughly 30 percent of emergency calls on a 2023 federal presentation of national emergency medical services data, and where there is generally no reimbursement at all. The uninsured, who fall outside a privately insured denominator. Medicare Advantage enrollees, whose ground ambulance out-of-pocket obligation the committee flagged separately as unaddressed. Medicaid enrollees. People who did not call for an ambulance out of fear of an unknown bill, which is the harm the committee named and the one no denominator can hold. And repeat transports, because every figure counts trips rather than people and no source states the mapping.

Under what conditions does it arise?

1. The exemption is written text, and the party that obtained it says so on the record. Air ambulances went into the Act and ground ambulances did not. The national trade association for ambulance services stated on the record in 2024 that the committee was formed following the successful advocacy of the association to carve ground ambulance services out of the No Surprises Act. The grounds stated on the record were the inability to choose patients, the safety-net role in rural and medically underserved urban areas, and the standing cost of readiness.

2. The substitute institution has no forcing function. A report, once delivered, discharges the statutory duty, and the federal advisory-committee record says exactly that. There is no deadline for Congress, no default rule that switches on if Congress does nothing, no sunset that returns the question to the calendar, and no reporting obligation attached to the outcome. Inaction here is not a violation of anything.

3. The machinery of the No Surprises Act does not transplant onto this service. That regime sets out-of-network payment from the median in-network rate in a region, resolves disputes through arbitration, and relies on notice and consent at the point of service. With roughly 85 percent of emergency claims out of network there is no meaningful median to compute. With approximately 75 percent of ground ambulance services billing fewer than three transports a day, and 47 percent classified rural or super-rural, both figures as recorded in the 2024 committee report, arbitration fees can exceed the disputed charge, and the committee declined to recommend arbitration as a fallback. And a patient in an emergency cannot read a waiver and decline.

4. A federal fix therefore means legislating a payment rate, and nobody wants to own the number. A large share of ground ambulance providers are governmental — fire departments and municipal services funded partly by local taxes and set through public processes. A federal rate either overrides local rate-setting, which the emergency services side argues would underfund readiness in a mostly rural sector, or defers to it, which the insurance side argues hands local governments an uncapped claim on premium dollars.

5. State law cannot close the gap even in principle. State insurance law reaches fully insured plans and does not reach self-funded employer plans. In 2025, 67 percent of covered workers were enrolled in self-funded plans, ranging from 27 percent at firms with 10 to 199 workers to 80 percent at firms with 200 or more, and employer-sponsored insurance covered 154 million people under age 65. One state that built a rate-reporting regime in 2026 reaches self-funded plans only if those plans elect to participate, which concedes the ceiling rather than passing through it.

What has been tried?

Two things were tried at once — carve the service out of the general rule, and convene experts to design a rule for it alone. The first succeeded. The second produced a report whose decisive parameter was left blank, and the body that produced it is now recommended for termination.

What was found?

Why is it still unsolved?

Institutional exemption — the federal rule exists, one service was written out of it by name, and what replaced the rule for that service was a committee whose duty ended when it filed.

Start with what the exemption is. Balance billing by hospitals, clinicians and air ambulances is prohibited by federal law. For ground ambulances the same statute instead created a committee, and the party that sought that outcome describes it in public as a won objective rather than as an accident of drafting. This matters for the diagnosis, because an absent regime and an exempted service fail differently. An absent regime can be built when attention arrives. An exemption has a constituency that already spent something to obtain it, and the reasons it gave are real ones — a service that cannot select its patients, that must stand ready in places where volume never repays readiness, and that in much of the country is a fire department rather than a firm.

The second movement is that the substitute was designed without a return path. Section 117 set two clocks, one for establishing the committee and one for the report, and attached no clock to anything that would follow. The committee ran nearly three years late against both clocks and that lateness cost nobody anything, because the statute made delivery the whole of the obligation. Once delivered, the federal record marks the statutory purpose complete and the committee inactive. Compare this with the parts of the same Act that Congress actually meant to bind, where the statute set the payment mechanism itself rather than asking a committee what it should be. Here there is no default rule that switches on, no sunset, no report on the report. An agency or a Congress that has stopped looks identical to one that is still working.

The third movement is the load-bearing one, and it is not that Congress ignored the advice. The advice arrived with its decisive parameter blank. The committee agreed unanimously that a prohibition must be coupled with a guaranteed payment, then could not agree on whether to mandate a minimum payment, on how to cap it, or on how any of it would apply to self-funded plans, and it said so in its own summary. It declined to name a multiple of Medicare. It dropped its non-emergency recommendations entirely because the package they were bundled with failed. And the registry of state and local rates that its own payment hierarchy depends on had to be recommended into existence, because no such registry exists. Meanwhile the workaround that filled the vacuum runs into a ceiling it cannot pass — state insurance law reaches fully insured plans, and two thirds of covered workers are in plans it does not reach. So a patient can live in a state that is counted as protected, be transported by a service that is legally barred from balance billing that state, and still receive the bill because of how an employer funds a health plan. That is a fact no patient knows at the moment of calling for help.

What observation would mean it is solved?

Candidates — (a) Congress enacts a ground ambulance balance-billing prohibition with a stated minimum payment standard and an effective date (b) the share of emergency ground ambulance transports producing a balance bill falls on a national measurement (c) the count of states with ground ambulance protections rises toward the whole country.

(a) alone is weaker than it looks. The committee was unanimous that a prohibition without a guaranteed payment is not a fix, and it refused to name the payment number, so the enacted multiple is the entire question rather than a detail of it. A statute could also exempt self-funded plans, which would leave two thirds of covered workers where they are while the headline reads that the gap was closed.

(b) alone cannot currently be observed. No national estimate of balance-bill exposure has been published on data more recent than 2017, and every 2025 and 2026 publication recycles that study. A measure that has not been refreshed in roughly a decade cannot register movement in either direction, and building one would itself be part of the fix rather than evidence of it.

(c) alone counts states rather than people. No source opened here reports how many people the 22 to 24 state laws actually protect, and no source reports any enforcement activity under any of them — no penalties, no complaint volumes, no audits. New York has had a law since 2015 and no published evaluation says whether balance billing stopped there. The three have to be read together, and (b) has to be read against whether anyone is still measuring.

What is it connected to?

Fills with researchair ambulance balance billing under the same statute, the doctrine of federal preemption of state insurance regulation for self-funded plans, rural emergency medical services funding and service viability, the Medicare ambulance fee schedule, and medical debt collection. Relation type and evidence grade were not confirmed in this round.

What these sources do not say

• How many people actually received a ground ambulance balance bill in a year. Every published figure is an out-of-network rate or a potential exposure share. Nobody publishes bills sent, bills collected, or amounts finally paid by patients after appeals. The quantity the entire policy argument is about is measured by no one.
• Why no successor estimate exists. The freshest national exposure figure rests on 2014 through 2017 claims from three insurers, and the 2026 sources citing it do so in the present tense without flagging the vintage. No federal agency has produced a replacement since the Act took effect, and no source states why not.
• What multiple of Medicare a federal fix would use. The report says explicitly that it did not recommend one and gives lack of consensus as the reason, but it does not say what range was on the table or what any candidate number would cost or save.
• How much of the country the proposed payment hierarchy would actually cover. Its second tier depends on state and locally set rates, and the federal registry that would identify them does not exist. No source estimates how many jurisdictions set qualifying rates or how many would meet the guardrails, so the most-used tier may be the emptiest and nobody has checked.
• How many people the state laws protect. The state count is published to the unit and the protected population is published nowhere. The two inputs needed to compute it, per-state scope and per-state fully insured population, are not jointly available in any source opened here, and the state-by-state table behind the count could not be retrieved at all.
• Whether any state law works. No source opened here reports enforcement activity, penalty counts, complaint volumes or compliance audits under any of the 22 to 24 laws.
• What Congress did with the report. Section 117 names five committees that must receive it and records nothing about what they did. The legislative databases that would settle whether any bill followed refused automated retrieval in this round, so the silence is recorded as not found rather than as does not exist.
• What the premium effect of a fix would be, or the dollar cost of leaving it alone. The only figure in circulation is a congressional budget estimate of a 0.5 to 1 percent premium reduction for the No Surprises Act as a whole, cited by analogy and not an estimate of any ground ambulance provision. The mirror figure, the aggregate cost of inaction to patients, is absent entirely, and the committee recorded that no evidence was presented to it on the magnitude of the offsetting effects it discussed.

See the evidence

The federal record was read directly and the central quantitative claim was not. The committee report was retrieved and read in full as text, which is where the verbatim Section 117 language, the payment hierarchy, the deadlock statement and the operational findings about small and rural providers come from. The committee landing page and the federal advisory-committee database entry were also read directly, and together they carry the entire current-status picture. Against that, the 85 percent figure comes second hand — the journal article is paywalled, so the number is cited through the federal report that quotes it twice with the citation attached, and only the companion 28 percent figure was read verbatim in an abstract. Where sources disagree the disagreement is left visible rather than resolved. The delivery date of the report is given four different ways by four federal or industry sources, from a 2024-03-29 cover date to a 2024-09-04 announcement, and none of them explains the roughly five-month spread. The state count is 22 in a March 2026 post and 24 in a July 2026 snapshot from the same research group, most likely reflecting further enactments rather than a methodological dispute, and both are reported with their dates rather than merged. The out-of-network share appears as 85 percent on 2014-2017 emergency transports and as 35 percent on 2021 trips of all kinds, which are different questions rather than rival answers, and a law-school commentary offers a third figure of nearly 80 percent with no denominator at all, which is why it is graded below the others. This is a Path A output, so observation_refs is empty and provenance_mode: press-derived.

People affected

Estimated range 840,000–2,550,000 As of 2014-2017 shares on a base of unstated data year

Sensitivity The width of the interval is not statistical uncertainty. It is the distance between two different questions asked of the same 2014-2017 study — how many patients were billed above the allowed amount, and how many had no network protection at all. The lower bound is harm that materialized on the narrowest published measure; the upper bound is the population the absence of a federal rule leaves unshielded. A separate series covering all ground trips under employer plans put the out-of-network share at 35 percent in 2021, which would place a midpoint near 1.05 million, but it is not substitutable because its denominator includes non-emergency trips and its data year is different. What the number fails to count runs in one direction only, upward: non-emergency ground transports are excluded from the base and are the exact category on which the advisory committee finalized no recommendation; responses where no transport occurs are roughly 30 percent of emergency calls on 2023 national emergency medical services data and generally produce no reimbursement at all; the uninsured, Medicare Advantage enrollees and Medicaid enrollees all fall outside a privately insured denominator; and people who did not call for an ambulance out of fear of an unknown bill are the harm the committee named and the one no denominator can hold. The limit in the opposite direction is that the 22 to 24 states with balance-billing protections are deliberately not netted out, so the range does not subtract anyone already covered by state law. Netting was not attempted because per-state scope and per-state fully insured population were not retrievable, and because no state law reaches self-funded employer plans, which covered 67 percent of covered workers in 2025, a correct netting would remove far less than the state count suggests. Finally, every input counts transports or trips rather than distinct people, and no source states how repeat transports map onto individuals, so the range may double-count anyone transported more than once in a year.

Regional breakdown No source opened here gives a state-by-state count of ground ambulance transports, of out-of-network transports, or of people protected by state law. The state-by-state table behind the count of 22 to 24 protected states lives in a client-rendered interactive that was not present in the archived snapshot and the live page refused retrieval, so neither the list of states nor their payment standards or market scope could be extracted. Splitting the national figure by state population would be proportional allocation and would be wrong in both directions at once, because the protection a person actually has depends on the state law where they were transported and on whether their employer plan is self-funded, and those two axes are distributed differently from population.