Nonunion PEMF Bone Growth Stimulator — desirability

The chair: the user. Here that is three people, and none of them is the same person: the adult with a long-bone fracture that has stopped healing, who has to wear the coil at home for months; the orthopaedic surgeon who has to write the prescription instead of booking a revision; and the DMEPOS (durable medical equipment, prosthetics, orthotics and supplies) supplier who has to put this device, not an incumbent's, in the patient's hands.

How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-018.md: the number is the claim's position in its ## Claims list and the tag is the one the candidate file carries today, copied and never adjusted here. A no-claim marker means nothing in the file speaks to the condition. See docs/deck-spec.md.

What changed since the last build. The Verifier moved tags on 2026-09-29. For this chair the most important move is the refutation of the volume claim: the file now records that the E0747 (a HCPCS code for durable medical equipment) line is smaller than the Generator guessed, which means fewer patients and prescribers sit in the user population this deck is about (Slide 1). The reimbursement machinery the patient's cost share runs through is now verified, which moves one link in the consent chain from assumption to fact (Slide 3) and leaves the behavioural links exactly where they were.

The rubric scores nothing on this deck directly. Factor 6 touches clinical risk and Factor 5 reads the adherence log as a moat; no factor asks whether any of these three people want the thing.

Slide 1 — The user, and the moment

Would have to be true: An adult with a long-bone fracture whose serial radiographs show healing has ceased for three or more months would have to reach a moment where their orthopaedic surgeon offers a home stimulator rather than going straight to revision fixation, and that moment would have to happen often enough to be a user population.

Where it stands: The clinical moment is precisely defined, and the definition is Medicare's own: NCD 150.2 (a Medicare national coverage determination number) says nonunion exists only once serial radiographs, two sets at least 90 days apart, confirm healing has ceased for three or more months claim 9verified, and the intended_use: restates it. The skull, vertebral and tumour-related nonunions the label leaves out are the ones NCD 150.2 excludes anyway claim 10verified. Whether surgeons offer a stimulator at that moment instead of revising is unchecked and cannot be checked by any connector claim 31unconfirmed. How often the moment happens is now smaller than the file first said. The Generator's guess that E0747 runs at low tens of thousands of Medicare services a year is refuted: the Verifier recorded 6,912 national Medicare fee-for-service services, about 4,900–5,400 beneficiaries, in data year 2024 claim 15refuted. The same figures show that some surgeons do prescribe stimulators for non-spinal indications claim 18verified. That is a Medicare FFS (fee-for-service) floor. It does not show the stimulator was chosen instead of revision. The wider nonunion population estimate is an unsourced Generator figure and stays unconfirmed claim 17unconfirmed.

What would settle it: Structured interviews with orthopaedic trauma surgeons about what they do at the three-month radiographic nonunion visit: stimulator, revision, both, or wait. No connector reaches surgeon decision-making. The volume leg is already settled for Medicare FFS. Adding non-Medicare volume needs a licensed or commercial-claims source that a human would have to attach.

If it's false: If surgeons at this moment go straight to revision, the device has no prescriber and every other slide on this deck is moot. The refuted volume claim already narrows this slide: the user population is the thousands of Medicare beneficiaries, not the tens of thousands the Generator guessed. The viability deck's TAM (total addressable market) slide takes the damage.

Slide 2 — What they'd have to do differently

Would have to be true: The patient would have to wear the garment-integrated coil for the prescribed hours every day across a multi-month course, and the low-profile form factor would have to make them measurably more likely to do so than the incumbents' devices. Before that, the surgeon would have to see non-adherence as a reason stimulator treatment fails, or the form factor answers a question nobody is asking.

Where it stands: Both halves of the patient condition are an explicit Generator assumption, unverifiable by design claim 29unconfirmed. The file does hold one real adherence figure, and it is context only, not evidence here. In the TRUST trial, 73% of patients administered at least half the recommended treatments with a daily self-administered LIPUS device claim 24verified. That is a different device, a different indication and a trial setting, and it suggests adherence is not free. Whether surgeons think non-adherence explains their failures is a condition this deck put on record; nothing checks it yet claim 42unverified. The prescribed hours per day and the course length appear in no claim, so the size of the behaviour change is a blank.

What would settle it: Surgeon interviews (the same round as Slide 1) that ask directly whether and how often non-adherence explains a stimulator failure. For patients, a wear-time study comparing the garment coil against an incumbent form factor. That is primary research with human participants, and a human decides whether to run it. Where adherence data exists in the incumbents' published trials, the literature scan can point at it; a pointer is not a tag.

If it's false: If surgeons do not see non-adherence as a failure mode, the garment coil and the adherence log, the candidate's only differentiators (see mechanism:), address a problem this chair does not have. Slides 3 and 5 lose their premise, and the viability deck's moat slide loses the one mechanism that is not the shared pathway.

Slide 3 — Who else has to say yes

Would have to be true: Four parties would each have to agree. The surgeon would have to prescribe a new-entrant device over the incumbent brands. A DMEPOS supplier would have to stock and dispense it. The payer would have to cover it for this patient. The patient would have to accept it into their home for months.

Where it stands:

• Surgeon: the preference for a new entrant over brands with decades of prescriber familiarity is an explicit assumption claim 28unconfirmed. The incumbents themselves are named from background knowledge and not confirmed by any PMA (premarket approval) or GUDID (Global Unique Device Identification Database) lookup claim 19unconfirmed.
• DMEPOS supplier: whether suppliers will carry a new entrant alongside the incumbents they already distribute is unchecked claim 32unconfirmed.
• Payer: Medicare's side is the strongest part of the chain. NCD 150.2 covers the non-invasive stimulator for nonunion of long bone fractures claim 8verified, and E0747 is paid to DMEPOS suppliers as its own line claim 23verified. Commercial and Medicare Advantage coverage is unchecked; the viability deck carries it on its first slide claim 34unconfirmed. Whether a Class II (FDA device class II) device bills on the E0747 fee lines at all is also open, and the viability deck carries that as its load-bearing condition claim 45unverified.
• Patient: see Slide 4.

What would settle it: Surgeon interviews (Slide 1). DMEPOS supplier interviews on what it takes to add a stimulator brand to a line card. The Medicare payer link is already verified at the coverage level. The rest of the payer link sits on the viability deck.

If it's false: Any single refusal stops the device reaching the patient. The surgeon and supplier links are where this candidate has no advantage the file can point to: claim 21unconfirmed says the incumbents gained the same 510(k) (premarket notification) route on the same day.

Slide 4 — What it costs them

Would have to be true: The patient would have to accept three costs. The first is money: the Medicare cost share and any non-covered balance on an E0747 device. The second is time: hours a day under a coil for months. The third is scrutiny: a device that records how many minutes they actually wore it and sends that record to their surgeon and to the supplier.

Where it stands: The allowed amount the cost share comes out of is verified: an E0747 new purchase is $4,743.80–$5,580.94 non-rural across the contiguous states and DC, July 2026 claim 16verified. The coinsurance percentage was not checked by any connector, so the patient's actual out-of-pocket figure is a visible blank. Whether patients accept that share is unchecked claim 33unconfirmed. The time cost is the adherence assumption on Slide 2 claim 29unconfirmed. Whether patients accept being monitored is a condition this deck put on record claim 43unverified. The adherence log is sold to the supplier and the surgeon; the person it watches is the patient.

What would settle it: The Medicare DMEPOS coinsurance rule for the out-of-pocket blank is a regulatory lookup a human or a future connector could run. Patient acceptance of cost and of monitoring needs patient interviews or a preference study with nonunion patients. No connector can answer either one.

If it's false: If patients refuse or quietly defeat monitoring, the log records less than it claims to. That puts the feasibility deck's tamper-evidence condition (claim 48unverified) under load, and it weakens the commercial case for the log (claim 30unconfirmed). If the cost share alone deters patients, the Medicare line in claim 15refuted shrinks further in practice.

Slide 5 — What would make them believe it

Would have to be true: The surgeon would have to believe two things. First, that PEMF (pulsed electromagnetic field) helps an established nonunion, meaning a trial or systematic review in nonunion rather than in fresh fracture. Second, that the adherence report the device exports is worth reading when they decide whether to continue treatment or revise.

Where it stands: The evidence a surgeon would ask for has not been read. The Generator retrieved nothing on PEMF in established nonunion claim 25unconfirmed, and the condition that controlled evidence shows PEMF improves union rates there is unchecked claim 40unconfirmed. A Verifier sweep found leads, including a 2024 systematic review update on PEMF in fractures. They are unread and unappraised, and a title is not a finding. The best-known trial in the adjacent space points the other way. TRUST found no benefit of LIPUS on radiographic healing or function in operatively managed tibial fractures claim 24verified. That is fresh fracture and ultrasound, not nonunion and PEMF, but it is the trial a sceptical surgeon will raise. Whether surgeons would use an exported adherence report in their decision is a condition this deck put on record claim 44unverified.

What would settle it: A literature-detail read of the leads recorded under claim 25unconfirmed, starting with the systematic review, to find out whether any of them addresses established long-bone nonunion specifically. Each finding becomes its own claim line for the Verifier. For the report, show a mock compliance report to surgeons during the Slide 1 interviews.

If it's false: If the nonunion evidence is thin or negative, the surgeon has no reason to prescribe any stimulator, and Slide 1 fails with it. If surgeons would not read the report, the log's value to the prescriber is gone, and whatever value it keeps rests only on the supplier and payer case (claim 30unconfirmed).

Slide 6 — The cheapest way to find out

Would have to be true: One round of interviews with orthopaedic trauma surgeons who prescribe stimulators today would have to show three things. They offer stimulators at the three-month nonunion visit. They see non-adherence as a reason treatment fails. They would read a device-generated wear-time report. If most say no to the second question, that observation falsifies the differentiator.

Where it stands: No interview has been run. The three questions are claims 31, 42 and 44, and none has anything verified behind it claim 31unconfirmed claim 42unverified claim 44unverified.

What would settle it: A human commissions and runs the interviews. It is a paid primary-research line item, not a lookup, and the adoption connector says as much for every behavioural claim in this file.

If it's false: Slides 1, 2 and 5 fail together, and the desirability case for this device reduces to "a stimulator like the ones that already exist".

Slide 7 — Where this deck outruns the file

Every condition above with nothing verified behind it. Read this slide first. After write-back every condition has a claim number, so the short no-claim list is not good news. Every behavioural condition on this deck rests on unconfirmed or unverified.

• Surgeons offer a stimulator at the nonunion visit rather than revising — claim 31unconfirmed — surgeon interviews; no connector.
• The nonunion user population outside Medicare FFS — claim 17unconfirmed — a licensed epidemiology or claims source a human attaches.
• Patients wear the device the prescribed hours; the garment coil improves that — claim 29unconfirmed — patient wear-time study.
• Surgeons see non-adherence as a cause of stimulator failure — claim 42unverified — surgeon interviews (written back by this deck).
• The prescribed hours per day and the course length — [no claim] — the incumbents' labelling or IFU (instructions for use), once a human retrieves one.
• Surgeons prescribe a new entrant over the incumbent brands — claim 28unconfirmed — surgeon interviews.
• The incumbents are who the file says they are — claim 19unconfirmed — a PMA or GUDID lookup no connector here performs.
• DMEPOS suppliers stock a new entrant — claim 32unconfirmed — supplier interviews.
• Commercial and Medicare Advantage (Medicare Advantage) coverage — claim 34unconfirmed — see the viability deck.
• A Class II device bills on the E0747 fee lines — claim 45unverified — see the viability deck.
• Patients accept the cost share — claim 33unconfirmed — patient research; the coinsurance percentage is itself unchecked.
• The patient's out-of-pocket dollar figure — [no claim] — the Medicare DMEPOS coinsurance rule, read by a human.
• Patients accept being monitored — claim 43unverified — patient interviews (written back by this deck).
• Controlled evidence that PEMF improves union in established nonunion — claim 40unconfirmed — literature-detail on the leads under claim 25unconfirmed.
• The nonunion evidence base as a whole — claim 25unconfirmed — same sweep.
• Surgeons use an exported adherence report — claim 44unverified — mock-report interviews (written back by this deck).

Sixteen conditions and none of them verified. The verified material on this deck (claims 8, 9, 10, 16, 18, 23, 24) describes the coverage rules, the price and the volume floor around these users, not what the users would do. That is the pattern docs/deck-spec.md predicts for a pipeline that scores supply-side questions only.

Slide 8 — The load-bearing condition

If only one thing from this chair could be checked: whether orthopaedic surgeons who prescribe stimulators for nonunion see patient non-adherence as a reason treatment fails claim 42unverified.

This one rather than the others: the category is already prescribed, and the verified E0747 volume shows it happens, so the open desirability question is not whether anyone wants a stimulator. It is whether anyone wants this one. Both differentiators, the garment coil and the adherence log, are answers to non-adherence. If surgeons do not see non-adherence as a problem, claims 29, 30, 43 and 44 lose their reason to be checked. It is also the cheapest condition on the deck to test, because it is one question in an interview round that Slides 1 and 5 need anyway.

Naming it is not a recommendation, a gate, or a kill. It is the answer to "check what first."