Evidence · idea-018

claim 1verifiedUSAsk ↗

Federal Register document 2026-07366, published 2026-04-16 (91 FR, Vol. 91 No. 73), is a final order reclassifying non-invasive bone growth stimulators, product codes LOF and LPQ, from class III into class II subject to premarket notification, with an effective date of 2026-05-18

jurisdiction: US — confidence: verified — source: found via web search of federalregister.gov and govinfo.gov listings for "Physical Medicine Devices; Reclassification of Non-Invasive Bone Growth Stimulators"; the document itself was not opened and no sentence was quoted from it. Not primary-source confirmed.…

the full check

[Verifier 2026-09-18] policy 2026-07366 --expect-type rule --jurisdiction US → verified, Federal Register document_number=2026-07366 (91 FR 20352). Every element of the claim is confirmed against the document itself: 91 FR 20352, "Physical Medicine Devices; Reclassification of Non-Invasive Bone Growth Stimulators", Rule, published 2026-04-16, FDA (US Food and Drug Administration)/HHS, effective 2026-05-18. Text read verbatim via --expect: “The Food and Drug Administration (FDA) is issuing a final order to reclassify non-invasive bone growth stimulators (product codes LOF and LPQ), postamendments class III devices, into class II, subject to premarket notification.” (--expect "class II", "premarket notification", "LOF", "LPQ" all → verified) and “DATES: This order is effective May 18, 2026.” (--expect "effective" → verified). The --expect-type rule check did not refute, so the instrument really is a final rule/order and not a notice. One correction of detail, not of substance: the page citation is 91 FR 20352; the claim's "Vol. 91 No. 73" was not the form the Federal Register returned and is not relied on here. Replay: policy 2026-07366 --expect-type rule --jurisdiction US.

claim 2verifiedUSAsk ↗

The reclassified generic type is codified at 21 CFR (Title 21 of the Code of Federal Regulations) 890.5870, "Non-invasive bone growth stimulator", described as a prescription device that provides stimulation through electrical, magnetic, or ultrasonic fields and is intended to be used externally to promote osteogenesis as an adjunct to primary treatments for fracture fixation and spinal fusion, or as a treatment for established nonunions or failed fusions

jurisdiction: US — confidence: verified — source: found via web search of the eCFR entry for 21 CFR 890.5870; the section text was not read in full. Not primary-source confirmed. [Verifier 2026-09-18] `regulation 890.5870 --title 21 --expect "prescription device that provides stimulation through electrical,…

the full check

magnetic, or ultrasonic fields" --jurisdiction US → verified, eCFR versioner title-21, issue 2026-09-16, section=890.5870

. Identification text read **verbatim**: “§ 890.5870 Non-invasive bone growth stimulator. (a) Identification. A non-invasive bone growth stimulator is a prescription device that provides stimulation through electrical, magnetic, or ultrasonic fields. The device is intended to be used externally to promote osteogenesis as an adjunct to primary treatments for fracture fixation and spinal fusion or as a treatment for established nonunions or failed fusions. (b) Classification. Class II (special controls).” The claim restates this correctly in every element, including the three-part indication language the whole candidate's narrowing argument rests on (

--expect "as a treatment for established nonunions or failed fusions" → verified). Checked against the frontmatter intended_use:

 — established nonunion of a long bone in skeletally mature adults — which sits inside the third limb of that language; a change of indication to fresh-fracture acceleration would invalidate this finding (see claim 7unconfirmed). Replay:

regulation 890.5870 --title 21 --expect "as a treatment for established nonunions or failed fusions" --jurisdiction US`.

claim 3verifiedUSAsk ↗

21 CFR 890.5870(b)(2) sets out special controls with which a class II non-invasive bone growth stimulator must comply as a condition of the class II designation; the Generator did not retrieve their text and asserts nothing about what they require

jurisdiction: US — confidence: verified — note: the special controls' text is no longer unretrieved — it was read in full this pass, which closes the largest named evidence-burden unknown in the file (cited as a discount in Factors 1 and 7); and the citation in the claim is narrower than the claim's own…

the full check

description — source: web-search summary of the eCFR section structure only. This claim is deliberately written as "special controls exist at this citation", not as a statement of their content, because the content is the single largest unknown in the evidence burden for this candidate. [Verifier 2026-09-18] regulation 890.5870 --title 21 --expect "special controls" --jurisdiction US → verified, eCFR versioner title-21, issue 2026-09-16, section=890.5870. Special controls do exist at this citation and are a condition of the class II designation: “(b) Classification. Class II (FDA device class II) (special controls). The special controls for this device are: …”. Precision point: the special controls are the list at (b)(1)–(7), not (b)(2) alone; (b)(2) is specifically the non-clinical performance testing control. The full list, read verbatim, is: (1) clinical data demonstrating the device performs as intended, with imaging data demonstrating fusion at the treatment site; (2) non-clinical performance testing verifying and validating that (i) intended design outputs are delivered to the patient, (ii) thermal safety and thermal reliability, (iii) signal characteristics are within safe physiologic limits, (iv) device reliability consistent with expected use-life; (3) biocompatibility of patient-contacting components; (4) electrical safety and electromagnetic compatibility; (5) software verification, validation and hazard analysis; (6) labeling comprehension testing demonstrating the patient can correctly use the device based solely on reading the instructions for use; (7) labeling including warnings against use on compromised skin or known skin sensitivities, warnings for patients with implanted medical devices, and a detailed summary of the supporting clinical data and clinical outcomes. Two of these bear directly on this candidate and are flagged for the Scorer rather than scored here: (b)(1) requires clinical data, so the 510(k) (premarket notification) is not a bench-only submission; and (b)(1)'s imaging requirement is written in fusion terms, which a nonunion indication will have to address. Replay: regulation 890.5870 --title 21 --expect "Labeling comprehension testing" --jurisdiction US.

claim 4verifiedUSAsk ↗

The final order in claim 1verified was preceded by a proposed order published in the Federal Register on 2020-08-17 (85 FR), Federal Register document 2020-17543, also covering product codes LOF and LPQ

jurisdiction: US — confidence: verified — source: found via web search (public-inspection.federalregister.gov and govinfo listings). Not primary-source confirmed. Carried because a six-year gap between proposed and final order is a fact the Verifier can check cheaply and it bears on how settled the new route is.…

the full check

[Verifier 2026-09-18] policy 2020-17543 --expect-type proposed-rule --jurisdiction US → verified, Federal Register document_number=2020-17543 (85 FR 49986). Confirmed: 85 FR 49986, same title, Proposed Rule, published 2020-08-17, FDA/HHS, comment period closed 2020-10-16. Product codes confirmed in the document's own text (--expect "LOF", --expect "LPQ" both → verified): “FDA is proposing to reclassify non-invasive bone growth stimulators, postamendments class III devices (product codes LOF and LPQ), into class II (special controls), subject to premarket notification.” The six-year proposed-to-final gap the claim flags is therefore real: 2020-08-17 → 2026-04-16. Replay: policy 2020-17543 --expect-type proposed-rule --expect "LOF" --jurisdiction US.

claim 5verifiedUSAsk ↗

Before the effective date in claim 1verified, non-invasive bone growth stimulators were postamendments class III devices for which a PMA (premarket approval) under 21 CFR Part 814 was required; the 510(k) route this candidate proposes to launch on therefore did not exist in the United States before 2026-05-18

jurisdiction: US — confidence: verified — note: verified from the final order's own text plus a corroborating openFDA absence, not from a pre-2026 classification snapshot — openFDA serves current class only, and both codes now read class 2 — source: Generator's reading of the reclassification framing found in the…

the full check

web-search results for claim 1verified ("postamendments class III devices, into class II, subject to premarket notification"). Not primary-source confirmed; the pre-2026 class III status in particular has not been checked against an FDA classification record. [Verifier 2026-09-18] policy 2026-07366 --expect "postamendments class III" --jurisdiction US → verified, Federal Register document_number=2026-07366 (91 FR 20352); and policy 2026-07366 --expect "premarket approval" --jurisdiction US → verified; and policy 2026-07366 --expect "part 814" --jurisdiction US → verified. The order states in its own words that these were “postamendments class III devices”, recites the statutory scheme “class I (general controls), class II (special controls), and class III (premarket approval)”, and cites “21 CFR part 814, subparts A through E (Premarket Approval of Medical Devices)”. Corroboration for the second half of the claim: clearances --product-code LOF --since 1996-01-01 --until 2026-09-18 --jurisdiction US → none (0 records) and the same for LPQ — openFDA holds no 510(k) clearance ever under either code, which is what a class III generic type looks like and is consistent with the 510(k) route not existing before 2026-05-18. Caveat recorded rather than glossed: openFDA's classification endpoint serves the current class (both codes now return class 2), so the pre-2026 class III status is evidenced by the Federal Register text and the clearance absence, not by a historical classification record. Replay: policy 2026-07366 --expect "postamendments class III" --jurisdiction US.

claim 6verifiedUSAsk ↗

FDA product code LOF designates a non-invasive bone growth stimulator using electrical or magnetic fields, and product code LPQ designates the ultrasonic variant; both codes are named in the final order at claim 1verified

jurisdiction: US — confidence: verified — note: LOF matches the claim exactly; LPQ's registered device name is not the wording the claim uses — the code is confirmed under the same regulation, but a reader should not carry "LPQ = non-invasive bone growth stimulator, ultrasonic" forward as FDA's own string —…

the full check

source: found via web search (fda.report product-code listing for LOF, and the two codes named together in the Federal Register summaries). No openFDA classification record was retrieved. Not primary-source confirmed. [Verifier 2026-09-18] classification --product-code LOF --jurisdiction US → verified, openFDA device/classification product_code=LOF: “Device class 2: Stimulator, Bone Growth, Non-Invasive (regulation 890.5870, review panel OR)”. classification --product-code LPQ --jurisdiction US → verified, openFDA device/classification product_code=LPQ: “Device class 2: Stimulator, Ultrasound And Muscle, For Use Other Than Applying Therapeutic Deep (regulation 890.5870, review panel OR)”. Both codes are class 2 under 21 CFR 890.5870, which independently corroborates claims 1, 2 and 5. The second half of the claim — that both codes are named in the final order — is confirmed verbatim: policy 2026-07366 --expect "LOF" and --expect "LPQ" both → verified on “(product codes LOF and LPQ)”. The wording gap on LPQ is recorded above and is worth an FDA pre-submission question if the ultrasonic fork in device_class is ever taken seriously. Replay: classification --product-code LOF --jurisdiction US.

claim 7unconfirmedUSAsk ↗

What the narrowing buys, stated as the regulatory claim it is: an intended use confined to the indications named in 21 CFR 890.5870 (fracture-fixation adjunct, spinal-fusion adjunct, established nonunion or failed fusion) falls inside the reclassified generic type and is reachable by 510(k), whereas an intended use claiming acceleration of healing of fresh or acute fractures, or general promotion of bone formation, falls outside it and would require a De Novo (De Novo classification request) or a PMA

jurisdiction: US — confidence: unconfirmed — note: this is the load-bearing claim of the candidate and it remains unevidenced. Half of it now rests on verified text; the other half is an FDA substantial-equivalence and scope judgement that no connector in this repo can make — source: Generator's inference from…

the full check

the indication language in claim 2verified, not a statement FDA has made about this device. The second half in particular (what the broad indication would require) names no regulation and no product code because the Generator identified none, and is not asserting one. This is the load-bearing claim of the whole candidate and it is the one most worth an FDA pre-submission rather than a connector lookup. [Verifier 2026-09-18] No connector answers this claim, and the two that come closest were run and say so. regulation 890.5870 --title 21 --expect "as a treatment for established nonunions or failed fusions" --jurisdiction US → verified (eCFR versioner title-21, issue 2026-09-16, section=890.5870) establishes only the first premise: the regulation's indication language is exactly the three things the claim says it is, and this candidate's intended_use: sits inside the third. Whether an intended use inside that language is therefore reachable by 510(k) — and, more importantly, what a wider indication would actually require — is a scope and substantial-equivalence determination FDA makes, not a fact in any database. The second half was probed directly against the order's own text and found no support in either direction: policy 2026-07366 --expect "fresh fracture" --jurisdiction US → unconfirmed, --expect "acceleration of healing" → unconfirmed, --expect "De Novo" → unconfirmed (document retrieved in full, phrases absent). Absent phrases are not refutation, and per this repo's scope limit a regulation or an order does not settle how FDA interprets it. The check remains an FDA pre-submission, as the claim itself says. Replay: regulation 890.5870 --title 21 --expect "as a treatment for established nonunions or failed fusions" --jurisdiction US.

claim 8verifiedUSAsk ↗

Medicare National Coverage Determination 150.2 (Osteogenic Stimulators) covers non-invasive electrical osteogenic stimulators for nonunion of long bone fractures, failed fusion where a minimum of nine months has elapsed since the last surgery, congenital pseudarthroses, and as an adjunct to spinal fusion surgery for patients at high risk of pseudarthrosis due to a previously failed fusion at the same site or undergoing multiple-level fusion

jurisdiction: US — confidence: verified — source: found via web search of the CMS (Centers for Medicare & Medicaid Services) Medicare Coverage Database NCD (national coverage determination) entry (ncdid=65); the NCD text itself was not retrieved through a connector and no version/effective date was captured. Not primary-source confirmed. [Verifier 2026-09-18] ncd 150.2 --jurisdiction US…

the full check

→ verified, CMS Pub. 100-3 Medicare NCD Manual, chapter 1 part 2 (Rev. 12299, issued 10-12-23), section=150.2. The policy's own text, read verbatim: “B. Nationally Covered Indications — 1. Noninvasive Stimulator. The noninvasive stimulator device is covered only for the following indications: • Nonunion of long bone fractures; • Failed fusion, where a minimum of 9 months has elapsed since the last surgery; • Congenital pseudarthroses; • Effective July 1, 1996, as an adjunct to spinal fusion surgery for patients at high risk of pseudarthrosis due to previously failed spinal fusion at the same site or for those undergoing multiple level fusion. A multiple level fusion involves 3 or more vertebrae (e.g., L3-L5, L4-S1, etc).” All four indications in the claim are confirmed, in the claim's own order. Two things the NCD text adds that the claim does not carry and a reader should: the word “only” (“covered only for the following indications”) — this is a closed list, not a floor — and the NCD's actual title is “Osteogenic Stimulator” (singular). Scope limit stated by the connector and passed through: this is the eligibility language CMS published, not a determination that any particular device or claim is payable. Replay: ncd 150.2 --jurisdiction US.

claim 9verifiedUSAsk ↗

NCD 150.2 (a Medicare national coverage determination number) states that nonunion of a long bone fracture is considered to exist only when serial radiographs have confirmed that fracture healing has ceased for three or more months prior to starting treatment with the electrical osteogenic stimulator, and that serial radiographs must include a minimum of two sets of radiographs, each including multiple views of the fracture site, separated by a minimum of 90 days

jurisdiction: US — confidence: verified — source: found via web search of the CMS Medicare Coverage Database NCD 150.2 entry. Not primary-source confirmed. This candidate's intended_use is written to restate this criterion deliberately, so if the criterion text differs from what is claimed here, the intended use is…

the full check

the thing that needs editing, not just the claim. [Verifier 2026-09-18] ncd 150.2 --jurisdiction US → verified, CMS Pub. 100-3 Medicare NCD Manual, chapter 1 part 2 (Rev. 12299, issued 10-12-23), section=150.2. Read verbatim: “• Effective April 1, 2000, nonunion of long bone fractures is considered to exist only when serial radiographs have confirmed that fracture healing has ceased for 3 or more months prior to starting treatment with the electrical osteogenic stimulator. Serial radiographs must include a minimum of 2 sets of radiographs, each including multiple views of the fracture site, separated by a minimum of 90 days.” The claim restates this criterion accurately in both halves (three-month cessation; two sets of radiographs, multiple views, 90 days apart). The candidate's intended_use: is therefore aligned with the coverage criterion as the claim intends, and no edit to the intended use is required on this finding. One adjacent line in the same policy the Scorer should see: an earlier bullet, “Effective September 15, 1980, nonunion of long bone fractures is considered to exist only after 6 or more months have elapsed without healing of the fracture”, sits in the text alongside the 2000 criterion; whether a contractor reads the two cumulatively is a policy-reading question, not a database one. Replay: ncd 150.2 --jurisdiction US.

claim 10verifiedUSAsk ↗

NCD 150.2 excludes nonunion fractures of the skull and vertebrae, and tumor-related nonunions, from coverage

jurisdiction: US — confidence: verified — source: found via web search of the CMS Medicare Coverage Database NCD 150.2 entry. Not primary-source confirmed. The intended_use above excludes the same three categories, so the launch indication claims nothing Medicare would not pay for. [Verifier 2026-09-18] `ncd…

the full check

150.2 --expect "Nonunion fractures of the skull, vertebrae and those that are tumor-related are excluded from coverage" --jurisdiction US → verified, CMS Pub. 100-3 Medicare NCD Manual, chapter 1 part 2 (Rev. 12299, issued 10-12-23), section=150.2

. The full sentence matched **verbatim**, not by keyword: “Nonunion fractures of the skull, vertebrae and those that are tumor-related are excluded from coverage.” The candidate's

intended_use:

 excludes the same three categories, so the claim's point stands on primary text: the launch indication claims nothing NCD 150.2 excludes. Replay:

ncd 150.2 --expect "Nonunion fractures of the skull, vertebrae and those that are tumor-related are excluded from coverage" --jurisdiction US`.

claim 11verifiedUSAsk ↗

NCD 150.2 provides that, effective 2001-01-01, ultrasonic osteogenic stimulators are covered as medically reasonable and necessary for the treatment of nonunion fractures

jurisdiction: US — confidence: verified — source: found via web search of the CMS Medicare Coverage Database NCD 150.2 entry. Not primary-source confirmed. Carried because it is what makes the LPQ/E0760 engineering fork in device_class a real commercial alternative rather than a hypothetical one. **[Verifier…

the full check

2026-09-18]**

ncd 150.2 --expect "Effective January 1, 2001, ultrasonic osteogenic stimulators are covered as medically reasonable and necessary for the treatment of nonunion fractures" --jurisdiction US

→ verified, CMS Pub. 100-3 Medicare NCD Manual, chapter 1 part 2 (Rev. 12299, issued 10-12-23), section=150.2. The whole sentence the claim asserts matched verbatim, including the 2001-01-01 effective date. The LPQ/E0760 engineering fork in device_class therefore has a national coverage instrument behind it as well, which is what this claim was carried for. Carry forward the same scope limit as claims 8-10: coverage eligibility language, not a payment determination, and the E0760 (a HCPCS code for durable medical equipment) payment and utilization legs remain unchecked (claims 14, 15, 16). Replay:

ncd 150.2 --expect "Effective January 1, 2001, ultrasonic osteogenic stimulators are covered as medically reasonable and necessary for the treatment of nonunion fractures" --jurisdiction US

claim 12unconfirmedUSAsk ↗

A DME MAC (durable medical equipment Medicare administrative contractor) Local Coverage Article titled "Osteogenesis Stimulators — Policy Article", article ID A52513, exists in the CMS Medicare Coverage Database and sets documentation and coverage conditions applicable to these HCPCS (Healthcare Common Procedure Coding System) codes

jurisdiction: US — confidence: unconfirmed — note: [2026-09-29] tag unchanged; the missing-file reason is gone and a different one replaces it. The coverage index is now cached and the A52513 row is in it, but the connector is code-keyed and the article's codes column is blank, so no connector can verify the…

the full check

article or what it requires. Earlier note: missing-connector gap, not a silent source — nothing was looked up in either direction, and this must not be read as evidence the article does not exist — source: found via web search of cms.gov medicare-coverage-database article listings; the article was not opened and nothing is asserted here about what documentation it requires. Not primary-source confirmed. Flagged as the highest-value single document for anyone assessing whether the onboard adherence log in mechanism has commercial value. [Verifier 2026-09-18] coverage E0747 --jurisdiction US → unconfirmed: “No local cache at data/medicare_coverage_policies.csv — Medicare coverage policy for E0747 cannot be checked until it exists. See data/README.md for the exact CMS source and expected columns. Do not substitute a web search for this file.” Two separate reasons this stays open, and both should be on record. First, the coverage connector has no data file in this environment, so the lookup could not run at all. Second, even with the file, coverage is code-keyed: it answers whether any policy names a HCPCS code, and cannot be pointed at an article ID. A DME MAC Local Coverage Article (A52513) is not an NCD, so ncd is the wrong instrument too — ncd reads Pub. 100-3 by section number and has no article route. Closing this needs either the cached coverage index plus a human reading the article at the source_ref it yields, or a human attaching the article directly. The claim's own assessment stands unchallenged: this is the highest-value single document for anyone testing whether the adherence log in mechanism has commercial value (claim 30unconfirmed). Replay: coverage E0747 --jurisdiction US. [Verifier 2026-09-29] coverage E0747 --jurisdiction US → refuted, local cache of CMS Medicare Coverage Database (no policy names this code), code=E0747: “No NCD or LCD (local coverage determination) in the cached coverage file names E0747.” This is not recorded as a refutation of this claim, and it is not a finding about Medicare coverage. The same cached MCD article export (data/medicare_coverage_policies.csv, from downloads.cms.gov/medicare-coverage-database/downloads/exports/current_article.zip, retrieved 2026-09-29) has a row

A52513, Article, "Osteogenesis Stimulators - Policy Article", codes=<blank>, version 50, effective 2026-05-18, related LCD L33796

, plus A59402 "Response to Comments: Osteogenesis Stimulators - DL33796". So the refuted comes from our linkage: the policy's code column is empty in the export (this is the trap recorded for idea-001 claim 9verified in references/reimbursement.md). NCD 150.2 covers the category (claims 8-11). The article's existence and title can be seen in the cache file, but no connector returns a check on an article ID, and the claim's substantive half, that the article sets documentation and coverage conditions for these codes, needs someone to read the article. That has not been done. The tag stays unconfirmed. Next step: a human opens A52513 (and LCD L33796 (a Medicare local coverage determination number), which the effective date of 2026-05-18 suggests was revised at reclassification; that is an inference, not checked).

claim 13verifiedUSAsk ↗

HCPCS Level II code E0747 exists with the descriptor "Osteogenesis stimulator, electrical, non-invasive, other than spinal applications"

jurisdiction: US — confidence: verified — note: [2026-09-29] moved unconfirmed → verified: the HCPCS Level II file is now cached and the descriptor matches exactly. The earlier missing-file note is kept below as history. Earlier note: missing-connector gap, not a silent source. The code's existence is…

the full check

unchecked in either direction; note that NCD 150.2 (claims 8-11) evidences national coverage of the device category but says nothing about this code string — source: found via web search of CMS coverage-article and payer-policy pages listing the code; not checked against the HCPCS quarterly release file. Not primary-source confirmed. [Verifier 2026-09-18] hcpcs E0747 --jurisdiction US → unconfirmed: “No local cache at data/hcpcs_level_ii.csv — HCPCS Level II code E0747 cannot be checked until it exists. See data/README.md for the exact CMS source and expected columns. Do not substitute a web search for this file.” No source_ref returned. The HCPCS quarterly release is the file that settles both the code's existence and its descriptor; it is not cached in this environment and a web search is not an acceptable substitute for it. Replay:

hcpcs E0747 --expect "Osteogenesis stimulator, electrical, non-invasive, other than spinal applications" --jurisdiction US

. [Verifier 2026-09-29]

hcpcs E0747 --expect "Osteogenesis stimulator, electrical, non-invasive, other than spinal applications" --jurisdiction US

→ verified, local cache of CMS HCPCS Level II file, code=E0747: “HCPCS E0747: Osteogenesis stimulator, electrical, non-invasive, other than spinal applications”, which is the claimed descriptor verbatim. Scope: this shows the code exists. It does not show payment (claims 16, 23) or coverage (claims 8-12).

claim 14verifiedUSAsk ↗

HCPCS Level II codes E0748 ("Osteogenesis stimulator, electrical, non-invasive, spinal applications") and E0760 ("Osteogenesis stimulator, low intensity ultrasound, non-invasive") exist alongside E0747

jurisdiction: US — confidence: verified — note: [2026-09-29] moved unconfirmed → verified: both codes are in the now-cached HCPCS Level II file with the claimed descriptors verbatim. Earlier note: missing-connector gap, not a silent source — same file, same reason as claim 13verified — source: found via web search of…

the full check

the same CMS coverage-article and payer-policy pages. Not primary-source confirmed. Carried separately from claim 13verified because E0748 is the line this candidate's narrowing deliberately gives up (claim 18verified) and E0760 is the line the LPQ fork would use instead. [Verifier 2026-09-18] hcpcs E0748 --jurisdiction US → unconfirmed and hcpcs E0760 --jurisdiction US → unconfirmed, both: “No local cache at data/hcpcs_level_ii.csv — HCPCS Level II code … cannot be checked until it exists.” No source_ref returned for either. Neither code's existence nor its descriptor is established or denied by this pass. Replay: hcpcs E0748 --jurisdiction US; hcpcs E0760 --jurisdiction US. [Verifier 2026-09-29] hcpcs E0748 --expect "Osteogenesis stimulator, electrical, non-invasive, spinal applications" --jurisdiction US → verified, local cache of CMS HCPCS Level II file, code=E0748: “HCPCS E0748: Osteogenesis stimulator, electrical, non-invasive, spinal applications”. hcpcs E0760 --expect "Osteogenesis stimulator, low intensity ultrasound, non-invasive" --jurisdiction US → verified, local cache of CMS HCPCS Level II file, code=E0760: “HCPCS E0760: Osteogenesis stimulator, low intensity ultrasound, non-invasive”. hcpcs E0747 --jurisdiction US → verified (claim 13verified), so all three codes do exist alongside each other. This shows the codes exist. It says nothing about payment or coverage.

claim 15refutedUSAsk ↗

TAM (total addressable market) leg 1 (volume, the checkable leg): the annual number of Medicare fee-for-service allowed services for HCPCS E0747 nationally is on the order of low tens of thousands per year

jurisdiction: US — confidence: refuted — note: [2026-09-29] moved unconfirmed → refuted. The DMEPOS (durable medical equipment, prosthetics, orthotics and supplies) utilization file is now cached. E0747 is 6,912 national Medicare FFS (fee-for-service) services, below even the bottom of "low tens of thousands". Read as patients, it is lower still: about 4,900–5,450 beneficiaries. The TAM…

the full check

volume leg is smaller than claimed. This does not mean the code has no volume. Superseded earlier note: missing-file gap, and the connector says so explicitly: “Nothing was looked up in either direction.” This is the file's own nominated first check and it could not be run — the TAM's one checkable leg remains an unsourced guess — source: Generator's order-of-magnitude guess with no data behind it. No CMS DMEPOS utilization file, Medicare Provider Utilization and Payment file, or procedures connector output was consulted this pass. This is the single most checkable number in the file and the one that should be pulled first: a real E0747 allowed-services count from CMS utilization data either supports or destroys the TAM without touching any other claim. [Verifier 2026-09-18] procedures E0747 --jurisdiction US → unconfirmed, no source_ref: “E0747 is a supplier-billed DMEPOS code, and the cached release (medicare_procedure_volumes.csv) covers physician-billed services — it carries essentially no E-prefixed codes at all. Nothing was looked up in either direction. This is a missing-file gap, not evidence that the code has no volume, and it must not be read as one. The release that can answer it is CMS ‘Medicare Durable Medical Equipment, Devices & Supplies — by Geography and Service’ (data.cms.gov); it is not cached, and this connector cannot yet read its columns — see docs/backlog.md 1.13.” The Generator's “low tens of thousands per year” is therefore neither supported nor contradicted, and the Scorer must not read this line as a null result. Two caveats to carry whenever the file does land: Medicare FFS excludes Medicare Advantage (Medicare Advantage), commercial, Medicaid, VA and cash-pay, so any figure is a floor on US volume, never a market size; and a DMEPOS allowed-services count is a supplier-billing count, not a patient count. Replay: procedures E0747 --jurisdiction US. [Verifier 2026-09-29] procedures E0747 --expect-volume 10000 --jurisdiction US → refuted,

local cache of CMS Medicare DMEPOS utilization file (referring provider geography and service), code=E0747, geography=National

(release per data/provenance.json: mup_dme_ry26_p05_v10_dy24_geor.csv, data year 2024, retrieved 2026-09-29): “E0747 (national, supplier-billed (DMEPOS): purchase 4,881 services / 4,881 beneficiaries, rental 2,031 services / 566 beneficiaries): claimed 10,000 overstates the cached figure of 6,912 services by 45%, outside the 25% tolerance.” The 10,000 floor is the most generous reading of the claim's "low tens of thousands", and it still fails. The bare replay procedures E0747 --jurisdiction US → verified returns the same figures and tests no number. How to read them: 2,031 of the 6,912 services are rental months across 566 beneficiaries, so the device/patient count is about 4,881 purchases plus up to 566 renters (not summed: one person can rent and then buy), i.e. roughly 4.9k–5.4k beneficiaries a year. The previous record's caveats still apply: Medicare FFS only (no MA (Medicare Advantage), commercial, Medicaid, VA or cash-pay), so this is a floor on US volume, never a market size. Because it is a floor, a non-Medicare-inclusive restatement of the claim could legitimately be higher, but the claim as written counts Medicare FFS allowed services and it is refuted on those.

claim 16verifiedUSAsk ↗

TAM leg 2 (price): the Medicare DMEPOS fee schedule purchase allowance for one E0747 device is in the low thousands of US dollars

jurisdiction: US — confidence: verified — note: [2026-09-29] moved unconfirmed → verified, with two qualifications that matter more than the tag. (1) The purchase allowance is $4,743.80–$5,580.94 (NU/KF, non-rural, contiguous states/DC, July 2026). That clears the $1,000 floor the claim's replay tested, but it…

the full check

is mid-single thousands, not "low". Read in the plain sense, the claim understates the price, so the Scorer should use the figures, not the words. (2) Every E0747 line on the schedule carries a second modifier, KF (NU/KF, RR/KF, UE/KF), and there is no non-KF line. The HCPCS meaning of KF (FDA class III device) was not checked by a connector here. Whether a device cleared as Class II under 890.5870 bills on this line at this fee is not answered by the file. The connector's floor test cannot touch that question, and it bears directly on this candidate's price leg. Superseded earlier note: missing-connector gap, not a silent source. The reimbursement factor's price leg is still an unsourced guess, so the existence of a code has not been shown to be a demonstration that anything pays above cost — source: Generator's unsourced guess, deliberately stated as a range of magnitude rather than a figure because the Generator has no DMEPOS fee schedule file and did not retrieve one. The check is the CMS DMEPOS fee schedule public-use file for the current year; note that fee schedule amounts vary by state and by purchase-versus-rental modifier, so "one number" may not exist. [Verifier 2026-09-18] payment E0747 --jurisdiction US → unconfirmed, no source_ref: “No local cache at data/hcpcs_payment_rates.csv — payment rate for E0747 cannot be checked until it exists. See data/README.md for the exact CMS source and expected columns. Do not substitute a web search for this file.” The claim's own nominated check — the CMS DMEPOS fee schedule public-use file — is the right one and is not available here. The claim's caveat that fee schedule amounts vary by state and by purchase-versus-rental modifier is carried forward unchallenged: “one number” may not exist even once the file lands. Replay: payment E0747 --min-expected 1000 --jurisdiction US. [Verifier 2026-09-29] payment E0747 --min-expected 1000 --modifier NU/KF --jurisdiction US → verified, local cache of CMS DMEPOS fee schedule, code=E0747, modifier=NU/KF, schedule=DMEPOS, release=July 2026: “NU/KF (new purchase), category IN — inexpensive and other routinely purchased items: non-rural $4,743.80 (CA) to $5,580.94 (AR) across 49 contiguous states/DC. Clears the claimed floor of $1,000.00 in every contiguous state/DC.” The fee is Medicare's allowed amount, not its payment: beneficiary coinsurance comes out of it, and it varies by state and by rural ZIP. The recorded replay payment E0747 --min-expected 1000 --jurisdiction US needs a modifier (there are three lines). --modifier NU → unconfirmed (“has no 'NU' line. Its lines are: NU/KF, RR/KF, UE/KF”). Other lines, for reference: RR/KF rental $471.38–$554.57 per month; UE/KF used $3,524.55–$4,146.53.

claim 17unconfirmedUSAsk ↗

TAM leg 3 (addressable share, and the narrowed denominator): the launch indication's denominator is not US fractures but the subset that progresses to an established, radiographically confirmed nonunion of a long bone, excluding skull, vertebral and tumor-associated nonunions; the Generator estimates this at roughly 5-10% of an annual US fracture volume on the order of six million, i.e. a target population one to two orders of magnitude smaller than the fresh-fracture population the broad indication would have reached

jurisdiction: US — confidence: unconfirmed — note: permanently unconfirmed absent a human-attached licensed source — unverifiable by design, not a miss, and it may not be upgraded by any agent. Note that the three legs decompose and the other two are tracked separately (claims 15, 16), so this line is now the…

the full check

residual share/epidemiology leg — source: Generator's unsourced estimate on both numbers. The 5-10% nonunion rate and the ~6 million annual US fracture figure are widely repeated background figures that the Generator did not retrieve from any primary epidemiological source this pass, and neither is specific to long bones or to the radiographic definition in claim 9verified. Written this way so the narrowed denominator is visible: the narrowing shrinks the TAM on purpose, and this claim is where that shrinkage is recorded rather than hidden. [Verifier 2026-09-18] market "nonunion of long bone, 5-10% of ~6 million annual US fractures, addressable share" → unverifiable, no source_ref: “Market-size claim … has no free, authoritative, machine-queryable source. Stays unconfirmed permanently unless a human attaches a licensed report and edits the claim by hand.” Neither the 5-10% nonunion rate nor the ~6 million annual US fracture figure was retrieved from any primary epidemiological source, and neither is specific to long bones or to the radiographic definition verified at claim 9verified — the Generator's own caveat, which this pass confirms rather than closes. The checkable leg of the TAM is claim 15refuted's utilization count and it is blocked on a missing file, not on judgement. Replay: market "nonunion of long bone, 5-10% of ~6 million annual US fractures, addressable share".

claim 18verifiedUSAsk ↗

What the narrowing costs, commercially: confining the launch indication to non-spinal established nonunion gives up the spinal-fusion adjunct indication billed under HCPCS E0748, which NCD 150.2 separately covers for high-risk pseudarthrosis patients (claim 8verified) and which the Generator believes is a larger annual Medicare DME (durable medical equipment) line than E0747

jurisdiction: US — confidence: verified — note: [2026-09-29] moved unconfirmed → verified: the relative-size half now resolves. Against the cached DMEPOS utilization file, E0748 is about four times the E0747 line by services and about five times by beneficiaries, at nearly the same purchase fee. The coverage half…

the full check

was already verified (claim 8verified). Earlier note: the coverage half is now verified through claim 8verified; the relative-size half is not, for a tooling reason — so the commercial cost of the narrowing is still asserted, not measured — source: the coverage half is claim 8verified's source; the relative-size half is the Generator's unsourced belief and is exactly as checkable as claim 15refuted, against the same CMS utilization data, for both codes at once. [Verifier 2026-09-18] Two halves, checked separately. Coverage half: ncd 150.2 --jurisdiction US → verified (CMS Pub. 100-3 Medicare NCD Manual, chapter 1 part 2 (Rev. 12299, issued 10-12-23), section=150.2) — NCD 150.2 does separately cover the noninvasive stimulator “as an adjunct to spinal fusion surgery for patients at high risk of pseudarthrosis due to previously failed spinal fusion at the same site or for those undergoing multiple level fusion”, verbatim, so the spinal line the narrowing gives up is a really-covered line and not a hypothetical one. Relative-size half: procedures E0748 --jurisdiction US → unconfirmed and procedures E0747 --jurisdiction US → unconfirmed, both “supplier-billed DMEPOS code … Nothing was looked up in either direction … missing-file gap, not evidence that the code has no volume” (docs/backlog.md 1.13). The belief that E0748 is the larger Medicare line is therefore untested, and one query against the CMS DMEPOS utilization release would settle claim 15refuted and this line at once. Replay: procedures E0748 --jurisdiction US. [Verifier 2026-09-29] Relative-size half: procedures E0748 --jurisdiction US → verified,

local cache of CMS Medicare DMEPOS utilization file (referring provider geography and service), code=E0748, geography=National

(data year 2024): “purchase 28,270 services / 28,162 beneficiaries”, no rental line. procedures E0747 --jurisdiction US → verified, same source, code=E0747: “purchase 4,881 services / 4,881 beneficiaries, rental 2,031 services / 566 beneficiaries: 6,912 services”. E0748 is larger by services (28,270 vs 6,912) and by beneficiaries (≈28.2k vs ≈4.9k–5.4k). Dollars point the same way: payment E0748 --jurisdiction US → verified, local cache of CMS DMEPOS fee schedule, code=E0748, modifier=NU/KF+RR/KF+UE/KF, schedule=DMEPOS, release=July 2026, NU/KF non-rural $5,544.80, against E0747 NU/KF $4,743.80–$5,580.94 (claim 16verified). The claim's belief that E0748 is the larger Medicare DME line is supported under all three readings. Medicare FFS only, so these are floors. The commercial cost of the narrowing is therefore real and large relative to the line this candidate keeps.

claim 19unconfirmedUSAsk ↗

Non-invasive bone growth stimulators are currently marketed in the US by Orthofix (product names "Physio-Stim" and "Spinal-Stim"), Zimmer Biomet (product name "OrthoPak"), and Bioventus (ultrasonic, product name "EXOGEN")

jurisdiction: US — confidence: unconfirmed — note: the refuted results quoted below answer a different question from the one the claim asks and are deliberately not carried into this tag. A PMA-marketed device does not appear in openFDA's 510(k) endpoint, and this repo has no PMA, GUDID (Global Unique Device Identification Database) or…

the full check

registration-and-listing connector at all — source: Generator's background knowledge plus incidental confirmation of the string "Biomet OrthoPak Non-invasive Bone Growth Stimulator System" in a GUDID device listing surfaced by web search. No PMA number, K-number, or GUDID identifier is asserted for any of these products — none was retrieved, and the Generator is not naming one it did not find. [Verifier 2026-09-18] Four lookups run, and the reasoning about what they can and cannot show matters more than the statuses. 510k "OrthoPak" --applicant "Biomet" --jurisdiction US → refuted (“No 510(k) record found”), and likewise 510k "EXOGEN" --applicant "Bioventus", 510k "Physio-Stim" --applicant "Orthofix", 510k "Spinal-Stim" --applicant "Orthofix" — all four returned refuted with no source_ref. Without the --applicant narrowing the same names return unconfirmed (510k "OrthoPak", 510k "EXOGEN", 510k "Physio-Stim" → “No 510(k) record found for device_name=…”). These do not refute the claim. The claim asserts current US marketing and explicitly asserts no K-number; claims 1 and 5 establish that this entire generic type was class III — PMA, not 510(k) — until 2026-05-18, so an absence from the 510(k) database is exactly what a legally PMA-marketed device looks like and is consistent with the claim rather than against it. Corroborating the same point from the other side: clearances --product-code LOF --since 1996-01-01 --until 2026-09-18 --jurisdiction US → none (0), same for LPQ, and clearances --device-name "osteogenesis stimulator" --since 1996-01-01 --until 2026-09-18 → none (0). Settling this needs a PMA/P-number or GUDID lookup, or FDA's registration-and-listing data; none exists in this repo. Tagging this refuted off a 510(k) miss would be answering the wrong question with a real-looking status. Replay: 510k "OrthoPak" --applicant "Biomet" --jurisdiction US.

claim 20verifiedUSAsk ↗

No 510(k) cleared under 21 CFR 890.5870 was identified by the Generator as of 2026-09-18, and whether a device legally marketed under a PMA before the 2026-05-18 reclassification is available as a predicate for a subsequent 510(k) is unresolved

jurisdiction: US — confidence: verified — note: both halves hold as written, and the first is now a real query rather than the Generator's silence. An empty result is a finding: openFDA was asked and returned zero. Read the note below before treating this as good news — it is the claim's own warning, now evidenced…

the full check

— source: Generator inference, no search performed. No openFDA 510(k) query against product codes LOF or LPQ was run this pass. This matters more than it looks: if no usable predicate exists, the "cheap pathway" the whole candidate rests on is a 510(k) with no predicate, which is not a cheap pathway at all. First check is an openFDA clearances query on LOF and LPQ with a decision-date window starting 2026-05-18. [Verifier 2026-09-18] clearances --product-code LOF --since 2026-05-18 --until 2026-09-18 --limit 50 --jurisdiction US → none, query product_code:"LOF" AND decision_date:[2026-05-18 TO 2026-09-18], count 0: “No 510(k) clearances for product_code:‘LOF’ in 2026-05-18 to 2026-09-18.” Same query on LPQ → none, count 0. Widened to the whole openFDA record rather than the post-reclassification window: clearances --product-code LOF --since 1996-01-01 --until 2026-09-18 → none (0) and LPQ → none (0) — no 510(k) has ever been cleared under either product code. Cross-checked by device name, not just code: clearances --device-name "bone growth stimulator" --since 2026-05-18 --until 2026-09-18 → none (0) and clearances --device-name "osteogenesis stimulator" --since 1996-01-01 --until 2026-09-18 → none (0). This is an empty query, not a query not run, and it is not white space in the market either — claims 1, 5 and 19 say the incumbents are there under PMAs. The consequence is the one the claim names and it points down, not up: four months after the route opened, there is no device cleared under 890.5870 to serve as a predicate, so the ‘cheap pathway’ this candidate rests on is currently a 510(k) with no predicate under its own new regulation. The second half of the claim — whether a pre-reclassification PMA device is available as a predicate afterwards — remains genuinely unresolved: that is an FDA substantial-equivalence and transition-policy judgement, no connector in this repo answers it, and the final order was probed for it directly (policy 2026-07366 --expect "predicate" --jurisdiction US → verified, but the matched text is the order's generic recitation of section 513(i) substantial-equivalence procedure, not a statement about these devices). Replay: clearances --product-code LOF --since 2026-05-18 --until 2026-09-18 --jurisdiction US.

claim 21unconfirmedUSAsk ↗

The 510(k) route described in claims 1, 2 and 5 became available on the same date to every incumbent named in claim 19unconfirmed and to every other entrant, so the pathway itself confers no exclusivity on this candidate

jurisdiction: US — confidence: unconfirmed — note: the factual premise is solid and the inference from it is not checkable here — the claim's force as the strongest argument against Moat is undiminished, but it is reasoning, not a sourced finding — source: Generator's reading of the final order's general applicability…

the full check

to the generic type (claim 1verified). Recorded as a claim rather than left in prose because it is the strongest argument against the Moat factor and should be visible to the Scorer, not only to a reader of the rationale note. [Verifier 2026-09-18] policy 2026-07366 --expect "LOF" --jurisdiction US → verified, Federal Register document_number=2026-07366 (91 FR 20352), on the text “reclassify non-invasive bone growth stimulators (product codes LOF and LPQ), postamendments class III devices, into class II, subject to premarket notification”, effective 2026-05-18 — so the order is written against product codes, i.e. the device type, and names no firm. What that implies for exclusivity is a reading of the regulatory scheme, and per this repo's scope limit a Federal Register document does not settle what it means in practice. Recorded for precision: policy 2026-07366 --expect "generic type" --jurisdiction US → unconfirmed (document retrieved in full, that exact phrase absent) — the phrase the claim's reasoning leans on is not one the order uses, though the product-code framing carries the same sense. Corroboration from the other direction: claim 20verified's clearance queries show no entrant, incumbent or otherwise, has used the route yet. Replay: policy 2026-07366 --expect "LOF" --jurisdiction US.

claim 22unconfirmedUSAsk ↗

Federal Register document 2026-16368 (91 FR 51710) limits eligibility for the RAPID coverage pathway to class II and class III FDA Breakthrough-designated devices, requires entry at the IDE presubmission stage with devices already market-authorized or already under an IDE excluded, and requires the device to be separately payable — so a device launching by 510(k) on an existing predicate, without a Breakthrough designation, is outside that eligibility

jurisdiction: US — confidence: unconfirmed — note: three of the four eligibility criteria are now confirmed verbatim and the fourth is not; the conclusion drawn from them is the Generator's inference. And a fact the claim does not carry: this is a notice open for comment until 2026-10-13, so RAPID is a…

the full check

published intention, not an operating pathway — source: the Federal Register document number and page citation are carried directly so the Verifier can check the document cold; the final clause (that this candidate's route falls outside the eligibility) is the Generator's inference from those criteria, not language in the document about this device. Carried because forgoing a coverage-at-authorization lane is part of what the narrowing costs, and it should be visible rather than assumed away. [Verifier 2026-09-18] policy 2026-16368 --expect-type notice --jurisdiction US → verified, Federal Register document_number=2026-16368 (91 FR 51710): “Medicare Program; Regulatory Alignment for Predictable and Immediate Device (RAPID) Coverage Pathway”, Notice, published 2026-08-11, CMS/HHS, comment period closes 2026-10-13 — the connector's own caution applies: “a document published for comment states what an agency intends, not a settled requirement.” The --expect-type notice check did not refute, so the instrument is correctly described. Criteria read verbatim from the document: “The RAPID coverage pathway will provide accelerated Medicare beneficiary access to certain eligible Class II and Class III (FDA device class III) U.S. Food and Drug Administration (FDA) Breakthrough-designated Devices” (--expect "Breakthrough", --expect "class II and class III" → verified); “The device must also be at the IDE presubmission stage and be the subject of an IDE study that enrolls Medicare beneficiaries…” and “FDA designated Breakthrough Devices planning to submit a PMA application…” (--expect "presubmission" → verified); “Separately payable devices that can, if approved, be billed to Medicare; and Not otherwise excluded from coverage through law or regulation”, preceded by “Not already the subject of a controlling Medicare NCD” (--expect "separately payable" → verified). Not confirmed: the claim's “already market-authorized or already under an IDE excluded” clause — --expect "market-authorized", --expect "already market authorized", --expect "already received market authorization", --expect "Devices that have already received", --expect "excluded from participation" all → unconfirmed against the full 63,621-character text. Unconfirmed, not refuted: long prose may say it in other words. The final clause — that this candidate's 510(k)-on-an-existing-predicate route falls outside eligibility — is the Generator's inference, as the claim itself says, and interpretation is outside what policy can settle. Two independent reasons it looks likely anyway, offered as reasoning and not as evidence: this candidate has no Breakthrough designation, and the criterion “not already the subject of a controlling Medicare NCD” sits directly against the NCD 150.2 coverage verified at claims 8-11. Replay: policy 2026-16368 --expect-type notice --expect "Breakthrough" --jurisdiction US.

claim 23verifiedUSAsk ↗

A device dispensed to a Medicare beneficiary by a DMEPOS supplier and paid under HCPCS E0747 is billed and paid in its own right rather than absorbed into a facility or episode payment

jurisdiction: US — confidence: verified — note: [2026-09-29] moved unconfirmed → verified. Read strictly: the claim is conditional (a device dispensed by a DMEPOS supplier and paid under E0747), and the question is whether that payment is its own line. The DMEPOS fee schedule prices E0747 as its own item in…

the full check

payment category IN, and the utilization file shows suppliers actually being allowed for it. This goes beyond a code-existence check. Not settled: settings where DME is folded into a facility payment (e.g. a SNF Part A stay), whether a Class II device bills on the KF line (claim 16verified), and whether the amount covers cost (claim 41unconfirmed). Earlier note: missing-connector gap, not a silent source — and this is the claim that turns ‘a code exists’ into ‘the thing gets paid for separately’, so the Reimbursement factor should not be read as resting on it — source: Generator's reading of what a DMEPOS fee schedule line implies; no CMS document was consulted for this specific proposition and no claims data was examined. Checkable against the DMEPOS fee schedule and the payment rules for the code. [Verifier 2026-09-18] payment E0747 --jurisdiction US → unconfirmed, no source_ref: “No local cache at data/hcpcs_payment_rates.csv — payment rate for E0747 cannot be checked until it exists… Do not substitute a web search for this file.” hcpcs E0747 --jurisdiction US → unconfirmed (“No local cache at data/hcpcs_level_ii.csv”). apc would be the connector for the packaged-versus-separately-payable question in the outpatient setting and it has no data file in this environment either. So none of the three routes to this proposition could be run: the code's existence, its payment amount, and its payment status are all unchecked in either direction. Recorded so the distinction survives: NCD 150.2 (claims 8-11) establishes coverage eligibility and says nothing about whether payment is separate, bundled or adequate. Replay: payment E0747 --jurisdiction US. [Verifier 2026-09-29] payment E0747 --jurisdiction US → verified, local cache of CMS DMEPOS fee schedule, code=E0747, modifier=NU/KF+RR/KF+UE/KF, schedule=DMEPOS, release=July 2026: three separately priced lines, all “category IN — inexpensive and other routinely purchased items”, NU/KF $4,743.80–$5,580.94, RR/KF $471.38–$554.57, UE/KF $3,524.55–$4,146.53 (allowed amounts, non-rural). procedures E0747 --jurisdiction US → verified, local cache of CMS Medicare DMEPOS utilization file …, code=E0747, geography=National: 4,881 supplier-billed purchase services plus 2,031 rental services allowed nationally (data year 2024). So the code is paid to suppliers in its own right. hcpcs E0747 --jurisdiction US → verified, local cache of CMS HCPCS Level II file, code=E0747 (the file is now cached). apc was not run: the claim is about the DMEPOS supplier setting, not hospital outpatient packaging.

claim 24verifiedUSAsk ↗

A large multicentre randomised trial of low-intensity pulsed ultrasound in operatively managed tibial shaft fractures — the TRUST trial, reported in The BMJ in 2016 — found no benefit on radiographic healing or on functional recovery relative to a sham device

jurisdiction: US — confidence: verified — note: the trial is now a citation rather than a search handle — PMID (PubMed identifier) 27797787 — and every element the claim recalled is confirmed, including the negative result — source: Generator's recollection from background knowledge. The citation was not retrieved this pass and…

the full check

no PMID was captured, so the Verifier should treat the trial name, year and journal as a search handle rather than a citation. Carried deliberately because it is evidence against the broad fresh-fracture indication this candidate gives up, and because a Scorer should not read the narrowing as purely a regulatory convenience: the widest version of this label is also the version with the best-known negative trial behind it. [Verifier 2026-09-18]

literature "Re-evaluation of low intensity pulsed ultrasound in treatment of tibial fractures TRUST" --since 2015-01-01 --until 2017-12-31

→ found (1 record), then literature-detail 27797787 → found, PubMed PMID=27797787 (efetch), https://pubmed.ncbi.nlm.nih.gov/27797787/. “Re-evaluation of low intensity pulsed ultrasound in treatment of tibial fractures (TRUST): randomized clinical trial”, BMJ, 2016, type: Journal Article, Multicenter Study, Randomized Controlled Trial. Design, from the structured abstract: “A concealed, randomized, blinded, sham controlled clinical trial with a parallel group design of 501 patients… 43 North American academic trauma centers… patients with an open or closed tibial fracture amenable to intramedullary nail fixation… allocated centrally to self administer daily LIPUS (n=250) or use a sham device (n=251).” Result, verbatim: “Results showed no impact on SF-36 PCS scores between LIPUS and control groups (mean difference 0.55, 95% CI −0.75 to 1.84; P=0.41)… There was also no difference in time to radiographic healing (hazard ratio 1.07, 95% CI 0.86 to 1.34; P=0.55).” Conclusion, verbatim: “Postoperative use of LIPUS after tibial fracture fixation does not accelerate radiographic healing and fails to improve functional recovery.” Registration NCT00667849 (a ClinicalTrials.gov registration number). Two limits on what this verifies. The connector's own caveat: a structured abstract is “deeper than a title, still not appraised: no methods, no risk-of-bias, no funding disclosure.” And the claim's own scope — this is LIPUS in operatively managed fresh tibial fractures, i.e. the broad indication this candidate gives up; it is not evidence about PEMF, and not evidence about established nonunion, which is claim 25unconfirmed's hole. One detail of wording: the trial enrolled tibial fractures amenable to IM nailing rather than “tibial shaft” specifically. Replay: literature-detail 27797787.

claim 25unconfirmedUSAsk ↗

The published clinical evidence base for pulsed electromagnetic field stimulation specifically in established nonunion — as distinct from fresh fractures — was not retrieved by the Generator this pass, and no trial, systematic review or PMID is named here in either direction

jurisdiction: US — confidence: unconfirmed — note: the claim is a true statement about what the Generator did, which no connector can verify; but the substantive implication is now partly answered and the answer is that the field is not empty — read the leads below before treating Factor 6's hole as unclosable…

the full check

— source: no literature search performed. Written as an explicit statement of an unexamined field rather than omitted, so Factor 6 carries a named hole instead of silence. A PubMed sweep on PEMF/electrical stimulation in nonunion is the cheapest way to close it. [Verifier 2026-09-18] No connector verifies a statement about another agent's process, so the tag stays unconfirmed on the claim as written. What a real sweep found, recorded as leads, not evidence, and not appraised: literature "pulsed electromagnetic field nonunion fracture" --since 1990-01-01 --until 2026-09-18 --limit 15 → found, 80 total matches, including PMID 39387850 “Effects of pulsed electromagnetic fields on bone fractures: a systematic review update” (2024 Dec), PMID 42270187 “Bone Growth Stimulators in Orthopedic Trauma: History, Mechanisms, and Clinical Applications” (2026 Jul), PMID 42305571 “Pulsed Electromagnetic Field (PEMF) Stimulation for the Treatment of Fifth Metatarsal Fracture Nonunion” (2026), PMID 41244766 “Evaluation of the efficacy of physical agent modalities in patients with fractures: a systematic review and network meta-analysis” (2025), PMID 41885227 “Extracorporeal magnetotransduction therapy as an adjunctive and primary modality for bone healing: a scoping review” (2026). literature "electrical stimulation established nonunion bone" --since 1990-01-01 --until 2026-09-18 --limit 15 → found, 17 total matches. None of these has been read, appraised or checked for whether it addresses established nonunion as distinct from fresh fracture — a title is not a finding, and per this repo's rules a scan result carries no confidence tag. The claim's point that Factor 6 rests on an unexamined field therefore still stands; what has changed is that the field demonstrably exists and closing the hole is a literature-detail sweep away, not a research programme. Replay: literature "pulsed electromagnetic field nonunion fracture" --since 1990-01-01 --until 2026-09-18 --limit 15.

claim 26unconfirmedUSAsk ↗

FTO (freedom to operate): no patent search of any kind was performed for this candidate — not for the PEMF waveform or coil geometry, not for the garment-integrated coil form factor, and not for the tamper-evident adherence-logging and compliance-reporting function. The FTO field is unexamined, and no patent number or family is named here because none was found

jurisdiction: US — confidence: unconfirmed — note: and the claim's own reading of its consequence is correct and is endorsed here: unexamined is not clear, and the Stage 5 ‘FTO blocked’ kill condition cannot fire on this line. It must not be read the other way either — source: Generator's honest statement of what…

the full check

was not done. This is verifiable-as-unexamined: the Stage 5 "FTO blocked" kill condition cannot fire on it, and it must not be read as "searched and clear.". [Verifier 2026-09-18] No lookup was possible and none was faked. patent takes a patent identifier as its only argument and this claim names none — there is nothing to replay. (EPO_OPS_KEY/EPO_OPS_SECRET are present in this environment, so the blocker is the absent patent number, not a missing credential; a future pass with a real number can run.) This repo has no assignee-search or CPC-search connector, and substituting a generic patent-search-engine query for one is explicitly out of bounds. So the FTO field is recorded as unexamined, which is what the claim says: no art was found and none was searched, in either direction, for the PEMF waveform, the coil geometry, the garment-integrated form factor, or the tamper-evident adherence-logging function. Replay: none possible — patent <patent_id> requires an identifier this claim does not supply.

claim 27unconfirmedUSAsk ↗

FTO, second and narrower: whether Orthofix, Zimmer Biomet, Bioventus or Enovis hold live, enforceable US patents covering non-invasive bone growth stimulator waveforms, coil designs, or treatment-compliance monitoring applicable to a nonunion-indicated PEMF device is unknown to the Generator; the underlying PEMF bone-healing art dates to the 1970s and 1980s and the foundational patents from that era would be long expired on any ordinary term, but no specific patent, family, assignee record or expiry date was located or checked

jurisdiction: US — confidence: unconfirmed — note: the age-of-the-art reasoning is an inference about a class of patents and is not evidence about any patent; note that it says nothing at all about the two modern differentiators, where live art is most likely — source: Generator's background knowledge of the…

the full check

field's age, with no search. The expiry reasoning is an inference about a class of patents, not a finding about any patent. Named assignees are the starting point for a PatentsView or EPO (European Patent Office) OPS search, which is the check. [Verifier 2026-09-18] patent is available in this environment (EPO_OPS_KEY and EPO_OPS_SECRET are both set) but requires a patent identifier, and this claim names no patent number, family or publication, so no invocation exists to run and none was invented. There is no assignee-lookup connector, so “what do Orthofix, Zimmer Biomet, Bioventus and Enovis hold” cannot be asked here at all. Recorded explicitly because the reasoning is inviting: even a verified result from patent would be a database fact about filing and grant dates, carrying a naive-expiration caveat — never a legal FTO conclusion, and certainly not a conclusion about a whole era of art. The 1970s-80s foundational-patent expiry belief is untested, and it would not cover the garment-integrated coil geometry or the adherence-logging and compliance-reporting function even if it were true. The named assignees remain the starting point for a PatentsView or EPO OPS search a human or a future connector must run. Replay: none possible — patent <patent_id> requires an identifier this claim does not supply.

claim 28unconfirmedUSAsk ↗

Assumption, not a fact: the idea depends on orthopaedic trauma surgeons being willing to prescribe a new-entrant bone growth stimulator for an established nonunion in preference to the incumbent brands named in claim 19unconfirmed, which have decades of prescriber familiarity and established DMEPOS distribution. There is no connector for prescriber preference and no primary research behind this belief; it is stated explicitly as an assumption so it stays visible to the Scorer and Verifier rather than being phrased as a fact

jurisdiction: US — confidence: unconfirmed — note: unverifiable by design, not a lookup that missed. The claim is correctly written as an assumption and stays one; no agent may resolve it, and a model's own view of what surgeons would do would be a fabricated stakeholder opinion — source: Generator's assumption,…

the full check

no source. [Verifier 2026-09-18] adoption "orthopaedic trauma surgeons willing to prescribe a new-entrant bone growth stimulator" → unverifiable, no source_ref: “Adoption/preference claim … asserts what clinicians, payers, or patients would do. No free, authoritative, machine-queryable source answers that — it needs primary research (interviews, survey, published preference study)… Do NOT substitute a model's own estimate of stakeholder behaviour: a synthetic stakeholder opinion is a fabricated claim.” The connector's suggested rescoping was attempted and is itself blocked: the procedures proxy (how often the incumbents are actually dispensed) is claim 15refuted's missing DMEPOS release, and the coverage proxy is claim 12unconfirmed's missing policy file. Closing this is prescriber interviews — a paid primary-research line item, not a lookup. Replay: adoption "orthopaedic trauma surgeons willing to prescribe a new-entrant bone growth stimulator".

claim 29unconfirmedUSAsk ↗

Assumption, not a fact: the idea depends on patients with an established nonunion wearing the device for the prescribed number of hours per day over a multi-month course, and on a low-profile garment-integrated coil materially improving that adherence relative to existing form factors. No adherence data for existing bone growth stimulators was retrieved and none is cited; stated explicitly as an assumption

jurisdiction: US — confidence: unconfirmed — note: unverifiable by design, not a lookup that missed — correctly written as an assumption and it stays one — source: Generator's assumption, no source. [Verifier 2026-09-18] `adoption "patients with established nonunion wear the device the prescribed hours;…

the full check

low-profile garment coil improves adherence" → **unverifiable**, no source_ref

 — same reason as claim 28unconfirmed: a patient-behaviour claim needs primary research and no connector can answer it. Worth putting on record from the one primary source read this pass, as context and **not** as evidence for this claim: the TRUST trial at claim 24verified (PMID 27797787) reports “Patient compliance was moderate; 73% of patients administered ≥50% of all recommended treatments” for a daily self-administered LIPUS device — a different device, a different indication and a trial setting, so it neither supports nor contradicts the adherence assumption here, but it is the nearest real adherence figure in the file and suggests the assumption is not free. Replay:

adoption "patients with established nonunion wear the device the prescribed hours; low-profile garment coil improves adherence"`.

claim 30unconfirmedUSAsk ↗

Assumption, not a fact: the idea depends on DMEPOS suppliers and payers treating an onboard, tamper-evident device-generated adherence log as materially useful documentation — for continued-coverage support, for audit defence, or for denial reduction — rather than as a feature nobody pays for. Nothing in claim 12unconfirmed's policy article was read, and no supplier or payer was consulted; stated explicitly as an assumption because it is the entire commercial case for the one differentiating feature in mechanism

jurisdiction: US — confidence: unconfirmed — note: unverifiable by design. This is the entire commercial case for the one differentiating feature in mechanism, and it is unevidenced in either direction — the cheapest partial test is claim 12unconfirmed's policy article, which is blocked on a missing data file rather than…

the full check

on judgement — source: Generator's assumption, no source. [Verifier 2026-09-18]

adoption "DMEPOS suppliers and payers treat a tamper-evident device-generated adherence log as materially useful documentation"

→ unverifiable, no source_ref — a payer/supplier-behaviour claim needing primary research. The checkable proxy the connector recommends is exactly the document claim 12unconfirmed names: LCA A52513's documentation conditions. That route was attempted and failed for a tooling reason — coverage E0747 --jurisdiction US → unconfirmed, “No local cache at data/medicare_coverage_policies.csv” — and coverage is code-keyed in any case and cannot resolve an article ID. So nothing was read about what documentation DME MACs require, and no supplier or payer was consulted. Replay:

adoption "DMEPOS suppliers and payers treat a tamper-evident device-generated adherence log as materially useful documentation"

claim 31unconfirmedUSAsk ↗

Orthopaedic surgeons managing a radiographically confirmed established long-bone nonunion will offer a non-invasive bone growth stimulator rather than proceeding directly to revision surgical fixation

jurisdiction: US — confidence: unconfirmed — note: Surgeon behaviour: no connector can answer this, and no model opinion stands in for primary research. The checkable proxy is procedures E0747 (claim 15refuted: 4,881 purchases / 566 renters in Medicare FFS, data year 2024). It shows stimulators are prescribed for…

the full check

non-spinal indications. It does not show that they are chosen instead of revision fixation. — source: deck-surfaced assumption (desirability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] adoption "surgeons offer a bone growth stimulator for established long-bone nonunion rather than revision fixation" → unverifiable. The connector's answer is unverifiable, which maps to unconfirmed permanently unless a human attaches primary research or a licensed source.

claim 32unconfirmedUSAsk ↗

DMEPOS suppliers will stock and dispense a new-entrant nonunion PEMF bone growth stimulator alongside the incumbent brands they already distribute

jurisdiction: US — confidence: unconfirmed — note: Supplier behaviour: no connector can answer this, and no model opinion stands in for primary research. — source: deck-surfaced assumption (desirability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] `adoption "DMEPOS…

the full check

suppliers will stock a new-entrant nonunion PEMF stimulator" → unverifiable. The connector's answer is unverifiable`, which maps to unconfirmed permanently unless a human attaches primary research or a licensed source.

claim 33unconfirmedUSAsk ↗

Patients prescribed a non-invasive bone growth stimulator will accept the Medicare beneficiary cost share, and any non-covered balance, for an E0747 device

jurisdiction: US — confidence: unconfirmed — note: Patient behaviour: no connector can answer this, and no model opinion stands in for primary research. For context only: the allowed amount the cost share comes out of is $4,743.80–$5,580.94 (NU/KF, claim 16verified). The coinsurance percentage was not checked by any…

the full check

connector. — source: deck-surfaced assumption (desirability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] adoption "patients accept the Medicare cost share for an E0747 device" → unverifiable. The connector's answer is unverifiable, which maps to unconfirmed permanently unless a human attaches primary research or a licensed source.

claim 34unconfirmedUSAsk ↗

Commercial insurers and Medicare Advantage plans cover non-invasive bone growth stimulators for established long-bone nonunion on terms comparable to NCD 150.2

jurisdiction: US — confidence: unconfirmed — note: Compound claim. Commercial half: no connector reads commercial payer policy, so it is unchecked. MA half, by rule rather than by practice: regulation 422.101 --title 42 --expect "national coverage determinations" --jurisdiction US → verified, `eCFR versioner…

the full check

title-42, issue 2026-08-13, section=422.101

: each MA organization must “(b) Comply with— (1) CMS's national coverage determinations”. That is the regulatory text requiring MA plans to follow NCD 150.2. How plans apply it in practice (prior authorisation, utilisation management) is outside what the regulation text settles. Overall unconfirmed. — source: deck-surfaced assumption (viability deck, 2026-09-20) — never checked; recorded so the Verifier can see it **[Verifier 2026-09-29]**

adoption "commercial and Medicare Advantage payers cover bone growth stimulators for nonunion on NCD 150.2 terms" → unverifiable. The connector's answer is unverifiable`, which maps to unconfirmed permanently unless a human attaches primary research or a licensed source.

claim 35unconfirmedUSAsk ↗

A new-entrant nonunion PEMF stimulator will take a share of annual US E0747 dispensing volume from Orthofix, Zimmer Biomet and Bioventus

jurisdiction: US — confidence: unconfirmed — note: Market share is permanently unverifiable without a licensed source. The denominator is checkable and has now been checked, which shrinks it: claim 15refuted is refuted, E0747 Medicare FFS volume is 6,912 services (≈4.9k–5.4k beneficiaries), data year 2024. — source:…

the full check

deck-surfaced assumption (viability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] market "new-entrant share of US E0747 dispensing volume" → unverifiable. The connector's answer is unverifiable, which maps to unconfirmed permanently unless a human attaches primary research or a licensed source.

claim 36unconfirmedUSAsk ↗

Patent protection on the garment-integrated coil geometry or on the tamper-evident adherence log is obtainable and would deter a fast-follower

jurisdiction: US — confidence: unconfirmed — note: this pass found no connector that can check the claim. Patentability and deterrence are legal judgements, and no Stage 3 connector answers either. patent checks the dates of a named patent, and none is named. patent-scan produces a lead list and never a tag. No…

the full check

search was run this pass (see claims 26-27). An FTO or patentability opinion is Tier 3 legal work. — source: deck-surfaced assumption (viability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] No invocation: no typed connector checks this claim.

claim 37unconfirmedUSAsk ↗

No incumbent US bone growth stimulator manufacturer will ship a low-profile garment-integrated or adherence-logging device for nonunion before this candidate reaches market

jurisdiction: US — confidence: unconfirmed — note: this pass found no connector that can check the claim. This forecasts what competitors will do. No connector answers a forecast, and adoption covers clinicians, payers and patients, not manufacturers. A clearances scan on LOF/LPQ would be a pointer to what has…

the full check

been cleared so far, and never a tag. Not run. — source: deck-surfaced assumption (viability deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] No invocation: no typed connector checks this claim.

claim 38unconfirmedUSAsk ↗

A device legally marketed under a PMA before the 2026-05-18 reclassification is available to a new entrant as a predicate for a 510(k) under 21 CFR 890.5870

jurisdiction: US — confidence: unconfirmed — note: The phrase is in the final order, but the order does not settle the claim. policy → verified, Federal Register document_number=2026-07366 (91 FR 20352), read from GovInfo: “…sufficient evidence that the device performs equivalently to a legally marketed predicate…

the full check

device…”. From the same GovInfo text (read via WebFetch at the URL the connector returned): “persons who intend to market this device type must submit a 510(k) … and must obtain FDA clearance”, and a 510(k) applicant “may be able to rely on clinical data from a legally marketed non-invasive bone growth stimulator with the same intended use”. The order assumes legally marketed predicates exist. In the passages retrieved it does not say that a device PMA-approved before 2026-05-18 serves as a 510(k) predicate afterwards. That is an FDA interpretation question, which policy cannot settle (references/regulatory.md). classification --product-code LOF --jurisdiction US → verified, openFDA device/classification product_code=LOF: “Device class 2 … regulation 890.5870”. 510k "bone growth stimulator" --jurisdiction US → unconfirmed: “No 510(k) record found for device_name='bone growth stimulator'”, so no post-reclassification clearance citing a PMA predicate was found by that name search. That is not evidence none exists. Intended use checked against: the frontmatter intended_use: (established long-bone nonunion, adult, prescription home use). — source: deck-surfaced assumption (feasibility deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] policy 2026-07366 --expect "legally marketed predicate device" --jurisdiction US → verified.

claim 39unconfirmedUSAsk ↗

A garment-integrated coil thin enough to be worn under ordinary clothing delivers, at the fracture gap, the field magnitude the incumbent non-invasive bone growth stimulators deliver, within the thermal and signal limits required by 21 CFR 890.5870(b)(2)

jurisdiction: US — confidence: unconfirmed — note: The engineering performance assertion has no connector. It needs bench data. The regulation text was retrieved so that the limits the claim invokes are on record: eCFR versioner title-21, issue 2026-09-25, section=890.5870. (b)(2) requires non-clinical performance…

the full check

testing verifying “(i) Intended design outputs are delivered to the patient; (ii) Thermal safety and thermal reliability; (iii) Signal characteristics are within safe physiologic limits; and (iv) Device reliability is consistent with the expected use-life.” The regulation names categories of verification. It sets no numeric field or thermal limit, so there is no number to test a thin coil against. Material and outside this claim: (b)(1) also requires that “Clinical data must demonstrate that the device performs as intended under anticipated conditions of use. Imaging data must demonstrate fusion at the treatment site.” Every 510(k) under this regulation carries a clinical-data special control. That is relevant to claims 3, 7 and 20 and to the Capital factor, and is noted here for the Scorer. — source: deck-surfaced assumption (feasibility deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] regulation 890.5870 --jurisdiction US → verified.

claim 40unconfirmedUSAsk ↗

Published controlled evidence shows that pulsed electromagnetic field stimulation improves union rates in established long-bone nonunion relative to standard fracture management alone

jurisdiction: US — confidence: unconfirmed — note: this pass found no connector that can check the claim. No Stage 3 check connector answers a claim about the published evidence base. literature is a Stage 0 scan: a pointer, never a tag. Not run as evidence. The claim needs a named trial or systematic review, and…

the full check

then its own claim line. claim 25unconfirmed records the same hole. — source: deck-surfaced assumption (feasibility deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] No invocation: no typed connector checks this claim.

claim 41unconfirmedUSAsk ↗

The garment-integrated coil and onboard logging electronics can be manufactured at a unit cost below the Medicare DMEPOS purchase allowance for E0747

jurisdiction: US — confidence: unconfirmed — note: Compound claim, and only the price half resolves. The price half is the purchase allowance, NU/KF $4,743.80–$5,580.94 non-rural (July 2026, local cache of CMS DMEPOS fee schedule, code=E0747, modifier=NU/KF). That is Medicare's allowed amount, not what the…

the full check

manufacturer realises after the supplier's margin, and whether a Class II device bills on the KF line is open (claim 16verified). The unit-cost half has no connector and no source. — source: deck-surfaced assumption (feasibility deck, 2026-09-20) — never checked; recorded so the Verifier can see it [Verifier 2026-09-29] payment E0747 --min-expected 1000 --modifier NU/KF --jurisdiction US → verified.

claim 42unverifiedUSAsk ↗

Orthopaedic surgeons who prescribe non-invasive bone growth stimulators for established long-bone nonunion identify patient non-adherence to the prescribed daily treatment time as a cause of the treatment failures they see in practice

jurisdiction: US — confidence: unverified — source: deck-surfaced assumption (desirability deck, 2026-09-29) — never checked; recorded so the Verifier can see it — evidence class 3 (stakeholder future)

claim 43unverifiedUSAsk ↗

Patients prescribed a home bone growth stimulator for established long-bone nonunion will accept a device that records their actual field-on treatment minutes and exports that record to their prescriber and their DMEPOS supplier

claim 44unverifiedUSAsk ↗

Orthopaedic surgeons will use a device-exported adherence report when deciding whether to continue non-invasive stimulator treatment of a long-bone nonunion or proceed to revision fixation

claim 45unverifiedUSAsk ↗

A non-invasive bone growth stimulator cleared as Class II by 510(k) under 21 CFR 890.5870 is billable to Medicare under HCPCS E0747 on the KF-modified DMEPOS fee schedule lines (NU/KF, RR/KF, UE/KF), at the allowances recorded in claim 16verified

jurisdiction: US — confidence: unverified — source: deck-surfaced assumption (viability deck, 2026-09-29) — never checked; recorded so the Verifier can see it

claim 46unverifiedUSAsk ↗

FDA accepts radiographic union of an established long-bone nonunion as satisfying the 21 CFR 890.5870(b)(1) special control that imaging data demonstrate fusion at the treatment site