idea-019 · narrative
A synthesis across this candidate's claims, dossier and decks — every statement here traces to something that already exists elsewhere in the knowledge base. A narrative never invents a fact, upgrades a tag, or recommends anything.
How to read this. This is a synthesis, not a new source of truth: every claim it cites carries the tag the candidate file gives it right now, copied not paraphrased, and every line quoted from a deck is quoted verbatim. See docs/narrative-spec.md.
What it was built against. The candidate file as of updated: 2026-09-29: 142 claims, of which 54 are verified, 80 unconfirmed, 7 unverified (claims 136-142, appended 2026-09-29) and 1 refuted. It also reads the dossier neurogenic-orthostatic-hypotension, the seed medicare-coverage-at-authorization, and all three decks. check_decks.py reports all three decks current, and none is STALE. Three sources are behind the tags, and that is disclosed here rather than corrected. The candidate's own Verified Composite and Gate Check were written 2026-09-28 over claims 1-87 only. Its Wrap section still reads "none refuted", calls the DMEPOS (durable medical equipment, prosthetics, orthotics and supplies) fee schedule uncached, and quotes an earlier viability load-bearing condition. Dossier §6 still says no code, rate or volume was looked up. All four predate the Verifier's 2026-09-29 re-checks. None of those sources was edited.
This is a wearable for adults whose blood pressure falls on standing because of Parkinson disease, multiple system atrophy (MSA (multiple system atrophy)) or pure autonomic failure (PAF (pure autonomic failure)). An abdominal bladder, a small pump and a posture sensor inflate on standing and vent on sitting or lying, so no pressure is applied while the wearer is supine. That is the property the design sets against the supine hypertension caused by pressor drugs claim 31unconfirmed. The mechanism has a published precedent: a servo-controlled 40 mmHg (millimetres of mercury) binder reduced symptom burden comparably to midodrine in 19 patients claim 5verified. That binder was inflated by hand before a short stand, not triggered by posture claim 73verified, and this device has not been built claim 8unconfirmed. The candidate was designed to be paid through the existing Medicare durable-medical-equipment (DME (durable medical equipment)) benefit, and the premise that fell on 2026-09-29 is part of that base case: the pneumatic-compressor codes E0650 (a HCPCS code for durable medical equipment)-E0652 are not capped-rental items, so a Medicare placement is paid once as a purchase rather than as a monthly rental claim 117refuted. Allowables are now published for the family claim 107verified. But no coverage policy reaches this device claim 116verified, which code it would bill under is open claim 140unverified, and the three decks, all rebuilt on the purchase model, each name a different condition that everything else rests on: daily wear, the posture classifier, and cost of goods against a one-time allowable.
The problem. Neurogenic orthostatic hypotension (nOH (neurogenic orthostatic hypotension)) is a fall in blood pressure on standing, caused by failed sympathetic vasoconstriction. In this domain it rides on a neurodegenerative disease that also impairs dexterity, posture and cognition [dossier: neurogenic-orthostatic-hypotension]. The populations are partly sized. Parkinson disease is 572 per 100,000 among people aged 45 or older: about 680,000 US individuals in 2010, with a projection (not a count) of about 930,000 in 2020 claim 81verified. The file's headline figure of about one million is not confirmed claim 21unconfirmed. Pooled OH (orthostatic hypotension) prevalence in Parkinson disease is 33%, with the neurogenic subtype at 26% and symptomatic OH at about 20%. The Verifier's note on that claim records I²=96% and a publication-bias-adjusted 24.9% claim 22verified. MSA is 12.4 per 100,000 (about 41,133 people) on a one-claim definition and 5.7 on two claims claim 80verified. PAF has no figure, and whether a usable published estimate exists is open claim 133unconfirmed. Among patients with Parkinson disease, those who also have nOH were older (mean 77 versus 74) and more often had a medically attended fall (30% versus 21%) claim 83verified. The dossier adds survey context that has not become a claim: 57% of 363 patients reported at least one fall due to nOH symptoms in the prior year, and only 36% of those meeting criteria reported a formal diagnosis [dossier: neurogenic-orthostatic-hypotension].
What patients are given today. Midodrine and droxidopa. Their approval status and label text are unconfirmed only because this repo has no drug-label connector claim 11unconfirmed claim 61unconfirmed claim 62unconfirmed. FDA (US Food and Drug Administration) determined that brand PROAMATINE was not withdrawn for safety or effectiveness claim 63verified. The drug side cuts both ways. Persistence is imperfect. Median persistence was 303 days on droxidopa and 172 on midodrine in US pharmacy claims, though the Verifier notes the midodrine cohort is OH generally claim 64verified. Droxidopa continuation was 75% at six months claim 66verified. A quarter of midodrine episodes ended in discontinuation in a Medicaid cohort, and the causes were not established claim 65verified. Supine hypertension, the harm this design is built to avoid, is common: 30.4% across alpha-synucleinopathies and 46.5% in MSA with OH claim 82verified. Yet on droxidopa, supine systolic pressure above 180 mmHg was 4.9% against 2.5% on placebo claim 69verified, and blood-pressure-related discontinuation was about 2.5% claim 67verified, so the pressor drugs' supine liability is measurable but not large in trials. The newest drug, ampreloxetine, missed its primary endpoints in both phase 3 trials claim 87verified. Abdominal binders already have guideline standing: a GRADE-based review gives them a strong recommendation among non-pharmacological measures claim 78verified, and the European EFNS guideline lists elastic binders in initial management claim 79verified.
The mechanism, and how far the evidence reaches. The strongest result is claim 5verified, a single-blind crossover in which servo-controlled 40 mmHg compression reduced symptom burden like midodrine and unlike placebo claim 5verified. The full text shows what it did not test: the inflator was activated manually less than 2 minutes before a stand of 10 minutes or less claim 73verified. One of 19 subjects had mild discomfort at 40 mmHg that eased at 20 mmHg claim 74verified. Other binder studies found less:
The one sham-controlled study of the automated binder, NCT03482297 (a ClinicalTrials.gov registration number), completed with 28 enrolled and has posted no results claim 6verified. Whether it shows benefit is open claim 98unconfirmed. Tolerability data on record are short-exposure only:
The only posture-triggered analog on record reached about 11 mmHg, and it was tested in healthy volunteers claim 86verified. The dossier adds context that is not a claim. The analog's own authors write that binders "are difficult to use at compression levels that are effective" and that most patients "may require assistance applying" them. A manikin study suggests not exceeding 25 mmHg. And no home-use adherence figure exists for any abdominal compression product in nOH [dossier: neurogenic-orthostatic-hypotension].
The device. It does not exist claim 8unconfirmed. Every engineering property is a chair's condition on an unbuilt design:
The latency budget those 15 seconds must fit inside is newly on record and unchecked claim 142unverified. The population makes posture sensing hard: 15.6% of one large MSA cohort had camptocormia claim 84verified. The nearest sensing precedent used three sensors and reported no numeric accuracy claim 85verified.
The regulatory route. 21 CFR (Title 21 of the Code of Federal Regulations) 870.5800 is a Class II (FDA device class II) regulation with product code JOW claim 1verified. Its text says "limb" and says nothing of the abdomen or of orthostatic hypotension claim 2verified. It already houses non-limb codes: LHX anti-shock trousers claim 51verified and PPS head-and-neck sleeves claim 52verified. FDA has cleared trunk and chest garments under it by 510(k) (premarket notification), none for OH claim 104verified. No granted De Novo (De Novo classification request) matches the relevant device names claim 55verified, but a name search is not proof that no predicate exists claim 3unconfirmed. A Class II De Novo is the regulatory chair's stated condition claim 102unconfirmed, and whether an IDE is needed is open claim 47unconfirmed. The claim-stripped fallback is weaker than it looks. The Class I truncal-orthosis regulation is musculoskeletal claim 54verified. A powered garment stripped of the OH claim would plausibly be a Class II 510(k) object claim 105unconfirmed. And the text of 21 CFR 801.4 runs against selling it to nOH patients without an intended use being read into it claim 127unconfirmed.
Why now — and why the file does not lean on it. CMS (Centers for Medicare & Medicaid Services) has published a Medicare coverage-at-authorization notice (RAPID) for Class II and III Breakthrough devices entering at IDE presubmission claim 12verified. The notice requires the device to be separately payable claim 13verified and excludes IVDs claim 14verified. It is a notice whose comment period closes 2026-10-13 claim 15verified. It paused the older TCET pathway claim 16verified, which CMS had capped at about five candidates a year claim 17verified. The instrument before that, MCIT, was repealed before it ever operated claim 18verified. RAPID admits a 510(k) device only through a recent De Novo predicate claim 19verified. The Verifier's note on claim 13verified adds that RAPID excludes a device already under a controlling NCD (national coverage determination). The Breakthrough statute itself is unconfirmed, for a tooling reason claim 20unconfirmed. The seed's reading of this thread is that "the destination is stable and the instrument is not" (seed context, not evidence). The seed also records this candidate as the one built closest to its design levers while keeping the pathway as upside only (seed context, not evidence). The candidate's base case needs no pathway in force on any date.
The money, and what changed on 2026-09-29. The base case is Part B DME. That the device falls in the DME benefit and is separately billable is an assumption claim 32unconfirmed, as is its meeting DME-definition conditions (3)-(5) claim 121unconfirmed. The controller's three-year life is a design commitment claim 119unconfirmed. Whatever Part B decides also binds Medicare Advantage (Medicare Advantage) basic benefits claim 120verified. That most users are Medicare beneficiaries is unconfirmed claim 44unconfirmed. The problems on record were already these:
On 2026-09-29 the Verifier read the July 2026 DMEPOS fee schedule and utilisation files:
The business case and decks then appended the purchase-model conditions, none of them yet checked:
The earlier capped-rental version of that supplier condition stays on file with its premise contradicted claim 124unconfirmed. So does the garment-replacement question, whose capped-rental framing no longer applies to the compressor family claim 122unconfirmed. Whether the manufacturer can bill as a supplier itself is open claim 125unconfirmed. No cost figure exists anywhere in the file: cost of goods against the allowable claim 96unconfirmed, margin against the fixed programme cost before a 510(k) follower claim 123unconfirmed, and garment life claim 129unconfirmed are all open.
IP (intellectual property) and competition. The Vanderbilt publication US2015313608 is real, covers abdominal venous-pooling compression, and is live on a naive term to 2033-11-04. Its grant status and claim scope are unknown claim 26verified. Whether any family member is granted and in force claim 88unconfirmed, whether it discloses posture triggering claim 90unconfirmed, and whether it is licensable claim 43unconfirmed are all open. No landscape search has been run claim 89unconfirmed claim 27unconfirmed. Isotech's ISO (International Organization for Standardization)-101 has a registered OH study with unknown status claim 60verified. The Generator's null competitor search is not an absence claim 34unconfirmed.
What has never been checked. Every statement about what a person would do. That covers:
The Verifier has recorded each of these as answerable only by primary research. The dossier adds that no price for any garment or drug was searched [dossier: neurogenic-orthostatic-hypotension].
Only verified claims. Several are published findings (evidence class 2), which the candidate's Verified Composite section leaves out of its arithmetic pending a human decision; they are verified as what the paper reports, not as appraised.
Refuted, said first. E0650, E0651 and E0652 are not capped-rental items. The July 2026 DMEPOS fee schedule assigns all three category IN on every pricing line, so a placement in this family is paid as a purchase (or an IN rental), not as a 13-month rental stream claim 117refuted. This is the most decision-relevant line in the file, because the business model was drawn around a rental stream. It changes the payment cadence and nothing in the coverage half: a purchase allowable is still not coverage claim 116verified.
Unverified — appended 2026-09-29 on the purchase model, not yet put to the Verifier:
Unconfirmed — the money:
Unconfirmed — regulatory, clinical and engineering:
Unconfirmed — people:
Unconfirmed — IP and competition:
Dossier context that has not become a claim (context, not evidence):
All three decks exist, were built 2026-09-29 against the candidate's current updated:, and pass check_decks.py. None is STALE. All three were rebuilt after claim 117refuted was refuted, and each says so in its opening. They agree on the shape of the file: nothing verified is about this device, because it has not been built, and the verified claims under each lens are analog findings, classification facts, or obstacles. They disagree on which condition sits underneath the others. Each makes an explicit case that its own condition does. None of them treats coverage as the first thing to check any more.
Desirability. The user's chair finds the literature strong and the user thin. Every statement about what a patient, carer, prescriber or supplier would do is unconfirmed, and the adjacent verified evidence leans against easy adherence: stockings rejected over donning and stigma claim 71verified, and no real-life activity change after a week of binder use claim 75verified. The purchase model changed one thing on this lens. Abandonment no longer costs a supplier a rental stream; it costs the patient their one paid device within its useful lifetime claim 139unverified. The deck argues this condition sits beneath prescriber willingness and price, because a prescriber will not keep ordering a device that patients take off.
"whether adults with Parkinson disease, MSA or PAF would actually wear and operate a powered abdominal garment during every upright period, day after day, for months" — the load-bearing condition of the desirability deck (built 2026-09-29).
Viability. The investor's chair is where the refutation landed hardest. The deck's previous build named coverage as load-bearing. It moved because paying once as a purchase changed what coverage would buy. The code this device most plausibly takes, E0650, has the lowest allowable in the family claim 140unverified and the smallest Medicare purchase volume claim 25verified. A controller is paid at most once per beneficiary in five years claim 139unverified. No cost figure exists claim 96unconfirmed. The deck says this can be tested before coverage is argued, through a coding enquiry and a desk-costed bill of materials, and that coverage stays necessary, since both verified coverage facts say current policy does not reach this device claim 9verified claim 116verified. The candidate's own Wrap section still quotes the earlier, coverage-based condition.
"whether the one-time Medicare purchase allowable of the code this device would actually bill under — E0650, the lowest in the family, if a single non-segmental bladder takes the non-segmental descriptor — exceeds the fully loaded cost of one patient's controller plus the garments it consumes" — the load-bearing condition of the viability deck (built 2026-09-29).
Feasibility. Engineering's chair points out that the candidate's reason to exist, set against pressor drugs, is zero supine exposure "by construction", and the classifier is the construction. The population is the hard case: camptocormia is common in MSA claim 84verified, and the only sensing precedent gives no numbers claim 85verified. The deck takes the purchase model into account and holds its position. A tighter cost envelope does not displace a classifier that might inflate on a recumbent patient with supine hypertension claim 82verified. It also notes that the classifier can be tested on sensor and firmware alone, before the pneumatic system or cost target is final.
"whether a single torso-worn posture classifier can keep false-upright inflation of a recumbent wearer below 1 in 1,000 recumbent periods in Parkinson disease, MSA and PAF patients, including those with camptocormia and dyskinesia" — the load-bearing condition of the feasibility deck (built 2026-09-29).
Now that a Medicare placement is paid once, as a purchase claim 117refuted — most plausibly at the lowest allowable in the family claim 140unverified, and for one controller per beneficiary in five years claim 139unverified — would a patient with Parkinson disease, MSA or PAF keep a posture-triggered abdominal garment on through every upright period for months claim 28unconfirmed, so that the one payment buys sustained use rather than a controller set aside after the first weeks?
This is a synthesis of the three lenses, not a choice among them. The refutation of claim 117refuted is what joins them. Under a rental stream, a patient who stopped wearing the device ended a payment. Under a single purchase, the payment has already been made. The viability deck's cost-against-allowable arithmetic claim 96unconfirmed is then worth only as many placements as turn into sustained wear. And the desirability deck's wear question is asked about a device the patient may not get a second paid chance at. The question does not answer the feasibility deck's condition, which comes before it: nobody should wear a device that inflates on a recumbent patient claim 108unconfirmed. The answer would have to come from device-logged wear in a home-use study claim 99unconfirmed, which needs a prototype, and before that from structured interviews that no connector can run. Nothing on record about this device answers it today.
Naming it is not a recommendation, a gate, or a kill. It is the answer to "if a human could only read one more thing about this candidate, what question should it answer."
_shared/plain-language.md