Medicare's Hyperbaric Rulebook and the Six Indications It Leaves Out

An aged typewritten policy page on a steel desk with three of its rows struck through in red pencil, a small brass pressure gauge resting at the corner

Ask what Medicare will pay for and the answer is a list. Fifteen conditions, national in scope, currently carried as version 4 of NCD 20.29 — a document whose effective date is 3 April 2017. Ask a hyperbaric physician what the therapy is for and the answer is a different list: the Undersea and Hyperbaric Medical Society’s fifteen indications, the field’s own statement of scope since the manual’s first edition.

Lay the two side by side and the overlap is nine. Six indications in the specialty’s own manual have no national Medicare coverage at all, and two of those six are not merely missing — the rulebook names them in a section headed “Noncovered Conditions,” which states plainly that no program payment may be made for any condition not listed in the covered section.

This is not a rumour passed around wound-care conferences. Both documents are public, both are current, and neither has been reconciled with the other.

What the rulebook actually says

NCD 20.29 begins with a sentence that is easy to read past: “Program reimbursement for HBO therapy will be limited to that which is administered in a chamber (including the one man unit) and is limited to the following conditions.” What follows is a flat enumeration — carbon monoxide intoxication, decompression illness, gas embolism, gas gangrene, acute traumatic peripheral ischemia, crush injuries and the suturing of severed limbs, progressive necrotising infections, acute peripheral arterial insufficiency, preparation and preservation of compromised skin grafts, chronic refractory osteomyelitis, osteoradionecrosis, soft tissue radionecrosis, cyanide poisoning, actinomycosis, and diabetic wounds of the lower extremities meeting three stated criteria.

Some of that list reads like an artefact of its era. Actinomycosis, a bacterial infection that modern antibiotics handle, survives as the fourteenth item, qualified as an adjunct “when the disease process is refractory to antibiotics and surgical treatment.” Both were reasonable inclusions when the coverage language was drafted; the revision history shows the structural rule was settled in July 1997, when CMS “clarified that coverage is limited to conditions listed under §35-10.A.”

Then comes the exclusion. Skin burns (thermal). Cutaneous, decubitus and stasis ulcers. Sickle cell anaemia. Exceptional blood loss anaemia. Acute thermal and chemical pulmonary damage — glossed in the document as “smoke inhalation with pulmonary insufficiency.” Multiple sclerosis. Arthritic diseases. Acute cerebral oedema. Twenty-two entries in all.

The document contains no mention of the words abscess, retinal, sudden, hearing, avascular or osteonecrosis; the word flap does not appear, and neither does compartment.

The manual’s fifteen, and where they diverge

The UHMS manual, now in its fifteenth edition with seventeen numbered entries across fifteen indications, is the closest thing the field has to a statement of scope. It is also where the field records its own evolution: the sudden-hearing-loss entry carries the note “New! approved on October 8, 2011 by the UHMS Board of Directors.”

That entry, fifteen years old by the manual’s own dating, has no national Medicare home. Neither does central retinal artery occlusion, filed there as 06a. Neither does avascular necrosis of the femoral head, which the manual files under its own chapter as a loss of flow in “a terminal vascular bed” where “hypoxic conditions mediate the condition and can improve with a course of hyperbaric oxygen.”

Intracranial abscess, indication 08, is absent from the coverage rule. So is the acute thermal burn injury — indication 13 — which the manual introduces with epidemiology that makes the exclusion hard to read twice: more than two million burn injuries brought to medical attention in the United States each year, roughly 14,000 deaths, about 20,000 injuries severe enough for a specialised burn unit, and 75,000 hospitalisations, a quarter of them lasting more than two months.

The anaemia gap is a naming problem on top of a coverage problem. The manual’s indication 07 is “severe anemia,” defined around “marked loss of red blood cell mass by hemorrhage, hemolysis, or aplasia.” The phrase severe anaemia does not occur in NCD 20.29. What occurs, in the non-covered section, is “exceptional blood loss anemia” — the older label for the same clinical predicament, and one an administrator reading a claim would have to recognise as a synonym to reach any other conclusion.

Two of the mismatches are partial rather than total, and arguably more consequential because they sit inside covered territory. Compromised skin grafts are payable, “not for primary management of wounds” — but the manual’s indication 12 is titled “Compromised Grafts and Flaps,” and the coverage document never says flaps. And “problem wounds,” the manual’s deliberately broad 06b, is payable in Medicare only when the wound is diabetic, on a lower limb, Wagner grade III or higher, and has failed an adequate course of standard therapy. Decubitus ulcers and venous stasis ulcers, named in the exclusion list, are not payable at all.

Coverage is not evidence, in either direction

The tempting reading of all this is that CMS is behind the science. The documentation does not support that reading cleanly, and it is worth saying so before anyone quotes the six-item gap as a scandal.

Take burns first. Cochrane’s review of hyperbaric oxygen for thermal burns, last updated in 2004, found four randomised trials, judged two eligible, and concluded that the evidence was of poor methodological quality, that pooling would be inappropriate, and that the review “has not found sufficient evidence to support or refute the effectiveness of HBOT for the management of thermal burns.” One trial reported shorter mean healing times, 19.7 days against 43.8; the other found no difference in length of stay, mortality or number of operations once adjusted for injury severity.

Sudden hearing loss is closer to a coin-flip than a verdict. The 2012 Cochrane review pooled seven trials and 392 participants and found no significant gain in the chance of a 50 per cent improvement in pure-tone average, but a significant gain in the chance of a 25 per cent improvement — risk ratio 1.39, 95 per cent confidence interval 1.05 to 1.84 — plus an absolute threshold gain of 15.6 dB, and a number needed to treat of five. The authors’ own conclusion was that the application of hyperbaric oxygen “significantly improved hearing, but the clinical significance remains unclear,” with no evidence of benefit in chronic presentations.

Meanwhile the covered list includes actinomycosis, for which nobody has run a modern randomised trial, and the intracranial abscess chapter that Medicare declines to pay for rests on mortality series assembled between 1981 and 1993 — six case series summing to a 22 per cent death rate in 636 patients, then fifteen studies from 1987 to 1993 summing to 18 per cent in 634. That is a thin evidence base, and it is fair to say so.

The honest description is that the two lists answer different questions on different clocks. The manual asks what a trained hyperbaric physician could justifiably offer a patient. The coverage rule asks what a national payer will fund by default, and it was written to be conservative — cheap to administer, hard to expand, and structurally hostile to adjunctive therapies whose benefit is measured over months.

The thirty-day clock, and what the codes pay

The utilisation controls sit inside the covered list rather than beside it. Hyperbaric oxygen is covered as adjunctive therapy “only after there are no measurable signs of healing for at least 30 days of treatment with standard wound therapy.” Wounds must then be reassessed at least every 30 days, and continued treatment is not covered “if measurable signs of healing have not been demonstrated within any 30-day period.”

That is a workable test for a diabetic foot ulcer with a measurable area, and an awkward one for an adjunct to burn excision or for a femoral head that will not declare itself radiographically for months. The shape of the rule selects which covered indications get used in practice.

Payment is split across two codes, and UHMS’s own regulatory-affairs FAQ documents how they came about. 99183, physician supervision of hyperbaric oxygen therapy, “was given a value by CMS in 1994, and has remained unchanged since that time”; when the code came up for review in the 2015 physician fee schedule cycle its work value was cut from 2.34 to 2.11 relative value units, roughly a 9 per cent reduction in the professional payment, put by UHMS at about $112 per treatment in facility settings against $123.59 before. The facility code changed at the same time: C1300, billed in 30-minute intervals, was deleted and replaced by G0277, with the same descriptor but without the outpatient-hospital restriction.

Read those two facts together — a per-session professional payment set three decades ago and shaved since, and an authorisation rule that asks for measurable healing every thirty days — and the incentive geometry becomes visible: it rewards high-volume wound care with demonstrable area reduction, and does not reward a burn centre running adjunctive therapy through a long admission or a hearing clinic treating within seven days of onset.

What the regulator is actually policing

The FDA occupies a third position in this arrangement, and it is not the position of either list. Hyperbaric chambers are Class II devices under 21 CFR 868.5470, cleared through the 510(k) pathway under product code CBF. Clearance attaches to a manufacturer’s labelled indications and to the device’s safe operation — not to whether a payer will reimburse the treatment.

That second obligation has become the FDA’s louder message. The agency’s letter to health care providers of 25 August 2025 opens with “reports of fires that occurred with HBOT devices that resulted in serious injuries and deaths,” records that the root cause of those fires was not determined, and asks facilities to follow manufacturer instructions for use, train staff, monitor patients for the full duration of treatment, observe cleaning and maintenance intervals, exclude prohibited electrical and static-producing devices, and dress patients in cotton rather than wool or synthetics.

The society’s own warning runs along the same line but aims at a different target. UHMS describes treatment at pressures below 1.5 ATA as unproven, notes that such sessions are “widely available in so-called ‘wellness centers’ and health spas outside the setting of medical facilities,” that they are generally not physician-prescribed or supervised, and that the commercial growth of these services “has led to the use of unsafe and unapproved chamber vessels outside medical facilities,” putting those exposed “at risk for serious injury and even death.” Nothing in the coverage debate changes that: hyperbaric oxygen sold for longevity, anti-ageing or general wellness is not an FDA-cleared use, and the coverage lists discussed here do not touch it.

Where the line could move

CMS’s coverage database currently records two national analyses touching hyperbaric oxygen — CAG-00060N, on hypoxic and diabetic wounds of the lower extremities, and CAG-00060R, on the topical-oxygen section that was removed from NCD 20.29 in 2017. Both are closed. No reconsideration of the indication list is open, and the revisions since 2017 have been coding maintenance rather than clinical.

What could force one is what has been missing for two decades: a trial big enough to settle a question. SuDHO, published as a protocol on 4 November, randomises adults with at least 50 dB hearing loss across pure-tone average who can start treatment within seven days of onset, is assessor-blind, and targets 151 participants to detect a 10 dB difference at 14 weeks with 90 per cent power. If it reports a clean result, the coverage question for sudden hearing loss stops being a judgment call about weak pooled data.

The international picture is already less tidy than the American one. The Brazilian otolaryngology taskforce, publishing its guideline the same week, recommended systemic and intratympanic corticosteroids as initial treatment and put hyperbaric oxygen in a different register: therapy that “may be considered in selected patients.”

What a patient should take from all of this is narrower than either camp would like. Being on the Medicare list is not a certificate of proven benefit — actinomycosis has held a place on it for nearly thirty years. Being off the list is not a finding of ineffectiveness — a burn, a hearing loss or an osteonecrotic femoral head can be treated in a hyperbaric chamber tomorrow and still generate a bill nobody pays. The two lists measure different things, and the distance between them is where American hyperbaric medicine currently does its paperwork.

Sources

← Back to blog