idea-019 · feasibility deck
Every slide states a condition that would have to be true, then reports where it stands using the candidate file's own claim and its own confidence tag. A deck never upgrades a tag, invents a number, or recommends anything, and there is no ask slide.
The chair: engineering. The people who would have to build a closed-loop pneumatic garment that knows when its wearer is standing, prove it is safe on a bench before anyone wears it, and produce the evidence a reviewer will ask for on a pathway whose regulation does not yet exist.
How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-019.md — the number is the claim's position in its ## Claims list, and the tag is copied from that claim as it stands right now, never adjusted here. A [no claim] marker means nothing in the file speaks to the condition at all. See docs/deck-spec.md.
The starting point, stated plainly. No device of this design exists as a prototype, and no bench, animal or human data exist for it — every performance property attributed to it in the candidate file is design intent [claim 8: unconfirmed]. What is verified is a published result for somebody else's investigational device [claim 5: verified]. The gap between those two sentences is this deck.
Would have to be true: The device would have to detect a transition to upright, inflate a splanchnic bladder to a prescribed set pressure under closed-loop servo control, hold that pressure for as long as the wearer is upright, and vent to zero on a transition to seated or supine — worn under clothing, in a home, by someone who may not be able to adjust it.
Where it stands: The set pressure has a published anchor: the investigational analog used sustained servo-controlled compression at 40 mmHg activated only on standing [claim 5: verified], and the registered sham-controlled trial used approximately 40 mmHg active against approximately 5 mmHg sham, inflating every time the participant stands up [claim 6: verified]. Almost every other quantity a requirements document needs is absent from this file: no time-to-pressure, no pressure-tolerance band, no battery life, no pump duty cycle, no garment mass, no maximum inflations per day [no claim]. The one that matters most clinically — that inflation is fast enough after standing to blunt symptom onset — is now on record as an assumption [claim 45: unverified].
What would settle it: The methods section of PMID 27271310 and the full protocol record of NCT03482297 for the investigational rig's own control parameters — literature-detail and trial-detail reach both identifiers, which are already in the file. Everything past that is a requirements exercise an engineering team writes, not a lookup.
If it's false: If the device cannot reach pressure in time, it is a garment that inflates after the patient has already sat down or fallen. The intended-use statement — reduce orthostatic symptom burden and increase upright standing time — is the claim the pathway is scoped against, so a latency failure is a Factor 1 problem as much as an engineering one.
Would have to be true: Compression of the splanchnic bed at the prescribed pressure, applied only while upright, would have to reduce orthostatic symptom burden and increase standing tolerance by a margin that matters — and it would have to do so with this device's bladder geometry and control loop, not just the Vanderbilt rig's.
Where it stands: The mechanism is the best-evidenced thing in the file. PMID 27271310 reports symptom burden falling 21.9±3.6 → 16.3±3.1 (P=0.032) with the binder versus 19.6±3.5 → 20.1±3.3 (P=0.756) on placebo, orthostatic tolerance improving similarly with binder and midodrine (AUC-SBP 195±35 and 197±41 versus 19±38 mm Hg×minute for placebo, P=0.003), and only midodrine raising seated SBP, in 19 autonomic-failure patients in a single-blind crossover [claim 5: verified]. A randomized, double-blind, sham-controlled trial of the same approach completed at Vanderbilt with 28 participants enrolled — but hasResults=false, so it verifies a design, not an outcome [claim 6: verified], and a separate Austrian study used an elastic binder, a weaker mechanism precedent [claim 7: verified]. Transfer to this candidate's own hardware is not evidence, it is an assumption, and it is now on record [claim 46: unverified] — because nothing has been measured for this device at all [claim 8: unconfirmed], and the Generator's own mechanism reasoning is labelled reasoning rather than finding [claim 31: unconfirmed].
What would settle it: Bench-equivalence work: instrument this device's bladder against the published rig's pressure profile, then a human haemodynamic study measuring standing SBP and symptom burden against the same endpoints claim 5 reports. The first is bench; the second is clinical and Tier 3.
If it's false: Mechanism & clinical risk is the strongest row in the file's Stage 2 draft table, and it is carried by claim 5 — a finding about a different device. If the effect does not transfer, that row's evidence base does not describe this candidate, and there is nothing else in the file holding it up.
Would have to be true: The launch route would have to be one whose evidence demands the programme can actually meet — and on the route the file carries first, that means a De Novo with special controls written from scratch, almost certainly including a clinical study.
Where it stands: The classification analog is confirmed to exist and confirmed to be a limb device: 21 CFR 870.5800 classifies a compressible limb sleeve as Class II under product code JOW, with real cleared devices whose indications for use are DVT prophylaxis and venous/edema indications [claim 1: verified], and the section's three sentences contain "limb" three times and neither "abdomen" nor "orthostatic" [claim 2: verified]. Whether FDA would find an abdominal, posture-triggered device with an OH treatment claim substantially equivalent to that regulation is an open question the regulation text alone does not settle, and a Q-Sub is the only instrument that settles it [claim 2: verified]. No cleared predicate or De Novo grant for this indication was identified, with the De Novo half not checked at all because data/fda_denovo.csv is absent [claim 3: unconfirmed]. The Breakthrough statutory criteria the narrowed indication is scoped against cannot be read here either — no connector in this repo reads the US Code or designation status [claim 20: unconfirmed]. The RAPID footnote that admits 510(k) devices requires a De Novo predicate authorized within 18 months, which does not reach this device on the JOW route [claim 19: verified]. That an IDE clinical investigation will be required rather than bench, biocompatibility and human-factors data alone is now on record as an assumption [claim 47: unverified], and no claim in this file names a single test standard, special control, or software safety level [no claim].
What would settle it: A Q-Sub pre-submission with FDA — human-run, multi-month, and the only instrument that resolves the SE question, as claim 2's own note says. Before that, De Novo decision summaries for comparable Class II powered therapeutic devices show what special controls and bench batteries have been demanded in practice; data/fda_denovo.csv has to land before the De Novo half of claim 3 can be checked at all.
If it's false: If FDA reads this as PMA-scale, or demands a pivotal trial larger than the 28-participant scale of the academic work, the capital and calendar in the viability deck's Slide 4 change shape entirely. If instead SE to JOW were somehow granted, the build spec gets much smaller — the uncertainty runs in both directions and neither is settled here.
Would have to be true: A torso-mounted accelerometer or inclinometer would have to tell standing from sitting and from lying down, reliably, in people whose disease produces stooped posture, camptocormia, freezing, festination and falls — and it would have to almost never inflate while the wearer is recumbent.
Where it stands: Posture-detection error is the first named failure mode in the Generator's own list, alongside pressure-sensing or vent failure, skin and abdominal-wall injury, interference with respiration or with abdominal stomas and feeding tubes, and reduced effectiveness at high BMI — all uncharacterized in both directions because the device does not exist [claim 31: unconfirmed] [claim 8: unconfirmed]. The sensing requirement is now on record as its own assumption [claim 48: unverified]. What makes it load-bearing rather than merely hard: the candidate's whole clinical differentiator is that supine exposure to the applied pressure is zero by construction — a property that is not a property of the bladder at all but of the posture classifier. The published comparison supports the underlying premise, in that only midodrine increased seated SBP while the binder did not [claim 5: verified]; whether patients and prescribers would treat that as decision-changing is the desirability deck's Slide 5 [claim 30: unconfirmed].
What would settle it: A sensing study before a garment exists: torso-mounted inertial recording in adults with PD and MSA across a full day, scored against observed posture, with the false-inflation rate reported as its own endpoint. No connector reaches this; it is bench plus an observational protocol. A MAUDE sweep of product code JOW would give adjacent-device harm patterns for the pneumatic failure modes, but MAUDE is a window scan with no denominator and produces no confidence tag, as claim 31's note says.
If it's false: A device that inflates while its wearer is lying down delivers exactly the supine pressure exposure the design exists to avoid, which converts the candidate's central safety argument into its central hazard. Factor 6's device-level risk is entirely unretired today; this is the part of it that cannot be engineered around by choosing a better pump.
Would have to be true: A bench rig would have to show, before any patient wears this: set pressure reached and held within tolerance across garment sizes and leak conditions; venting to zero on single-fault conditions — pump runaway, pressure-sensor failure, loss of the posture signal — fast enough that no wearer is compressed while recumbent; interface pressure mapped across the abdominal wall; and pump duty cycle and battery endurance across a realistic day of sit-stand transitions.
Where it stands: None of it has been done, and the file says so twice over: no bench, animal or human data exist for this device [claim 8: unconfirmed], and the named failure modes are uncharacterized [claim 31: unconfirmed]. No acceptance criterion for any of these tests appears anywhere in the file [no claim]. The single-fault venting requirement is now on record as an assumption [claim 49: unverified].
What would settle it: A bench programme — instrumented garment, pressure logging, fault injection — costed and scheduled as Stage 7 work. The Scores table says the same in its Mechanism row: device-level risk closes only with a bench programme. Nothing in this repo and no agent in it greenlights the animal or human work that sits immediately downstream; that is a Tier 3 decision.
If it's false: If single-fault venting cannot be demonstrated, the device cannot be argued safe for unsupervised home use by a population that sleeps in it or near it, and both the De Novo special-controls case and the human-factors case fail at the same time.
Would have to be true: The garment would have to be manufacturable in a small number of stock sizes that still deliver the prescribed splanchnic pressure across real body habitus, and it would have to be donned, charged and cleaned in a home by someone with a movement disorder — not fitted by a technician in a lab.
Where it stands: Sizing and manufacture are now on record as an assumption [claim 50: unverified]; the file's only adjacent statement is that effectiveness may be reduced at high BMI, which is listed as an uncharacterized failure mode [claim 31: unconfirmed]. The human-factors reality is the desirability deck's Slide 2, referenced rather than restated: daily donning and operation by a dexterity-impaired population is an explicit assumption [claim 28: unconfirmed], and tolerance of sustained compression through every upright period is another [claim 39: unverified]. Nothing about sterilisation, cleaning, shelf life or supply chain appears anywhere in the candidate file [no claim].
What would settle it: Formative human-factors sessions with PD and MSA patients and carers using a mock garment — cheap, early, and the standard FDA would expect for a home-use device anyway; plus anthropometric sizing work against the target population. Neither needs a working servo loop, so both can run in parallel with the bench programme.
If it's false: If the device needs per-patient custom manufacture, the DMEPOS unit economics in the viability deck's Slide 3 change against an allowable that is itself unknown [claim 23: unconfirmed]. If it cannot be donned unaided or with ordinary carer help, the addressable share in claim 24 shrinks to whatever fraction has daily hands-on support.
Would have to be true: The design space for a posture-triggered, servo-controlled abdominal compression device would have to be open — either the on-mechanism family does not cover this implementation, or a licence is available, or there is a workaround that does not compromise the mechanism.
Where it stands: One named family sits squarely on the mechanism: US2015313608, "Compression device, system, and method for decreasing abdominal venous pooling", filed 2013-11-04, naive 20-year term expiry 2033-11-04, not expired — with grant status, issued claim scope, current assignee and the Vanderbilt CTTC availability listing all undetermined, and with the candidate's own asserted filing date of 2015-04-30 contradicted by the record [claim 26: verified]. A pre-grant publication is not a granted patent and this is not an FTO conclusion in either direction; it is the trigger for one. Beyond that single number, no patent search was performed at all — no landscape of posture-sensing compression, servo pneumatic bladder control, or abdominal-compression-for-hypotension art, and no assignee search against IPC incumbents [claim 27: unconfirmed]; an incidental lookup confirms adjacent live servo/pneumatic-control art exists. Whether an exclusive licence is available is the viability deck's Slide 5 [claim 43: unverified].
What would settle it: A patent-office file-history pull on US20150313608A1 — grant status, issued claims, assignee, term — free and a few hours, and the single highest-value hour in this file. A real FTO opinion across the landscape is outside counsel work and Tier 3.
If it's false: FTO blocked is one of the three named Stage 5 kill conditions, and Factor 4 sits near the floor of the file's Stage 2 draft table for unexamined risk on a named on-mechanism family. If the granted claims cover posture-triggered abdominal compression and no licence is available, the design question is not how to build it but whether it can be built at all in this form.
Every condition above with nothing verified behind it. Read this slide first. The verified claims on this deck cluster in two places — the mechanism (claims 5-7) and the regulation text (claims 1-2) — and neither of them is about this device.
data/fda_denovo.csv has to land first.
If only one thing from this chair could be checked: whether a wearable inertial sensor can discriminate upright from seated and recumbent posture in adults with Parkinson disease or multiple system atrophy at a false-inflation rate low enough for a safety case built on zero supine pressure exposure — [claim 48: unverified].
Why this one rather than the others. The bladder, the pump and the servo loop are ordinary engineering, and the mechanism already has a published human result behind it [claim 5: verified]. The differentiator this candidate is built on is not the compression; it is when the compression is not applied. That property lives entirely in a posture classifier running on a population whose posture is abnormal by definition — stooped stance, camptocormia, freezing, falls — and it is the one unknown that can be attacked today, with an inertial logger and a scoring protocol, before any garment exists. The FTO file-history pull on Slide 7 is cheaper and should be done first because it is an hour; this is the condition the engineering programme turns on.
Naming it is not a recommendation, a gate, or a kill.