idea-001 · 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.

Home Stress-Incontinence Stimulator (pivot P2, two-channel trigger) — feasibility

The chair: engineering. The people who would have to build the probe, the controller, the reference lead and the detector, test them, make them, and produce the evidence an FDA (US Food and Drug Administration) reviewer will ask for.

How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-001.md. The number is the claim's position in that file's ## Claims list. The tag beside it is copied from that line exactly as it stands today, never softened and never upgraded. A bracketed "no claim" marker means nothing in the file speaks to the condition; no condition on this deck needs one. See docs/deck-spec.md.

Rebuilt 2026-10-01 from the candidate file after pivot P2 the discriminating trigger. The earlier 2026-10-01 build is retired, not amended. Its load-bearing condition asked whether a single vaginal EMG (electromyography) channel could tell a voluntary contraction from a Valsalva or brace. P2 the discriminating trigger is the shape built on the answer that it cannot be assumed to. The candidate's mechanism: now reads: stimulation "triggered by a two-channel detector - pelvic floor surface EMG from the probe plus an abdominal-muscle surface-EMG reference from a skin electrode the woman places herself each session - so that each stimulation burst fires on a voluntary pelvic floor contraction and is withheld when the detector sees a Valsalva manoeuvre or abdominal brace before the burst (the decision is taken before each burst; stopping a burst already running when a strain begins is not part of this mechanism)".

A note on how much this deck says about the detector. The withhold logic is the element the file's IP (intellectual property) claims treat as possibly new claim 198unconfirmed, claim 199unconfirmed. This deck describes it at the level of the mechanism: field and the claims already on file, and adds no implementation detail of its own.

Slide 1 — What the device actually has to do

Would have to be true: The device would have to do four things.

Deliver intravaginal NMES (neuromuscular electrical stimulation) within output ranges already cleared in the KPI pool.
Before each burst, read a pelvic floor EMG channel from the probe and an abdominal reference channel from a self-placed skin electrode, then fire on a voluntary pelvic floor contraction and withhold on a Valsalva or brace.
Never let the reference cause a burst the pelvic floor channel did not trigger.
Log enough session data to show use on at least 16 days in 30.

Two quantities the spec needs are blank in the file: the discrimination performance the detector must hit, and the number of sessions in a course.

Where it stands:

Output: whether the envelope can sit inside cleared KPI ranges claim 83unconfirmed. The nearest triggered predicates deliver up to about 90 to 100 mA into 500 ohms claim 206unconfirmed. K201014 (an FDA 510(k) clearance number)'s preloaded stress-incontinence ETS (electrical tissue stimulation) programme is 45 Hz, 200 microseconds, 6 s work and 10 s rest claim 207unconfirmed.
Decide before the burst: the mechanism takes the decision before each burst. A strain that begins during a burst is not caught, and catching it would reopen the concurrent-recovery problem claim 207unconfirmed. Cleared envelopes leave at least 2 s off between bursts on the three summaries read claim 152unconfirmed.
Reference can only remove: if the detector fires only when the pelvic floor channel crosses its own threshold and the reference condition is also met, no reference fault can cause an untriggered burst claim 205unconfirmed.
Data capture: limited by her use rather than by engineering claim 108unconfirmed.
Blanks: no claim states a target sensitivity, specificity or false-withhold rate. The file says only that the false-withhold rate on correct contractions has to be measured in women claim 177unconfirmed. Sessions per course are stated nowhere claim 183unconfirmed. Electrodes per session would be two or three claim 208unconfirmed.

What would settle it: A design-input document that sets the discrimination targets and the session schedule. Both are design decisions a human makes, and no source supplies them.

If it's false: Without targets there is nothing for bench verification (Slide 5) to verify against, and no basis for the performance claim the pathway would need (Slide 3).

Slide 2 — The mechanism that has to hold

Would have to be true: Two things would have to hold. First, two channels would have to carry enough information to separate a voluntary pelvic floor contraction from a Valsalva or brace, in women who have failed an exercise trial, by a margin a detector can use. Second, withholding stimulation on the wrong manoeuvre would have to change something that matters clinically compared with firing on it. The magnitude of either is unknown.

Where it stands:

Why one channel is not enough. Vaginal pelvic floor EMG rises during Valsalva and bracing as well as during a voluntary contraction, and symptomatic women show more of it on Valsalva claim 150unconfirmed. Vaginal pressure did not separate the two manoeuvres in one small study claim 151unconfirmed.
Why two channels are not obviously enough. Abdominal EMG also rises during a correct contraction claim 177unconfirmed, and intra-abdominal pressure rises with an effortful correct contraction, so the decision cannot be a simple veto on the reference alone claim 203unconfirmed. No study reports how well any two-channel detector discriminates claim 204unconfirmed.
Whether withholding matters. No study compares a two-channel with a single-channel trigger on any continence outcome claim 178unconfirmed. Triggered stimulation through a vaginal probe for incontinence is buildable and clearable claim 165verified, but its margin over open-loop stimulation has never been measured claim 91unconfirmed. That stimulation during a strain causes no harm beyond a lost training effect, which is what makes the decide-before-burst limit a performance issue rather than a hazard, is on record and unchecked claim 209unverified.
Who the trigger leaves out. Women who cannot contract cannot trigger it claim 96unconfirmed. Women who co-activate the abdomen would receive fewer stimulated contractions claim 179unconfirmed.

What would settle it: The separation question is answered by recordings in women (Slide 4). The outcome question needs a clinical comparison against a single-channel trigger, and committing to one is a human decision.

If it's false: If the channels do not separate, P2 the discriminating trigger has no function to sell, and the differentiation slides on both other decks fall. If they separate but withholding changes nothing, the feature is real and clinically empty. Factor 6.

Slide 3 — What the pathway forces you to build

Would have to be true: The pathway would have to stay a KPI 510(k) (premarket notification) cleared on non-clinical testing. That makes the build list: electrical safety and EMC, including isolation of the stimulator from the abdominal lead; software and cybersecurity documentation; biocompatibility of the probe for its cumulative contact; usability for a self-placed electrode; and bench data on the withhold logic sufficient for whatever the labelling says about it.

Where it stands:

The class is settled. KPI is Class II (FDA device class II) under 876.5320, a 510(k) code claim 1verified, claim 2verified. The triggered predicate K243079 was cleared on non-clinical testing alone claim 153verified.
P2 the discriminating trigger's route is not settled. Whether the reference channel is a technological difference needing no De Novo (De Novo classification request) and no clinical data where bench testing verifies the withhold logic claim 188unconfirmed. Whether the device is equivalent to K243079 on the triggered characteristic claim 155unconfirmed. A labelling claim that it withholds on the wrong manoeuvre needs test data in women; a functional description could rest on bench verification claim 189unconfirmed.
The electrode. The reference electrode as a recording cutaneous electrode under 882.1320 adding no second submission claim 190unconfirmed. Cleared surface-EMG skin electrodes are available from more than one manufacturer claim 184verified.
Usability. If a lifted or misplaced reference only ever withholds or falls back to single-channel triggering, placement may not be a critical task claim 192unconfirmed. The claim that no home ETS device's summary shows a usability file to borrow from is refuted: K220161 is OTC and lists IEC 62366 usability work claim 191refuted.
Electrical, biological, software. Isolation and auxiliary-current testing under IEC 60601-1 and 60601-2-10 claim 206unconfirmed. A biocompatibility category set by cumulative contact claim 162unconfirmed, with test articles in the condition of use claim 161unconfirmed. Software and cybersecurity content claim 164unconfirmed.

What would settle it: An FDA pre-submission on the reference channel and the intended labelling. Then a human read of FDA's software, cybersecurity, human factors and ISO (International Organization for Standardization) 10993-1 guidance, which no connector reaches.

If it's false: If FDA wants clinical data for the withhold, the build list gains a study. Factor 1 holds, but Factor 7 moves.

Slide 4 — The hardest unknown

Would have to be true: A two-channel detector, with the reference electrode placed by the woman herself and not by a clinician, would have to separate a voluntary contraction from a Valsalva or brace with usable specificity and an acceptable false-withhold rate, session after session, unsupervised. Displacement of the electrode would have to degrade it detectably rather than silently.

Where it stands: Never measured, in any configuration.

No study reports two-channel discrimination, nor how it changes when the abdominal electrode is displaced claim 204unconfirmed. Single-channel discrimination has not been measured either claim 105unconfirmed.
Contact-impedance monitoring catches a detached or lifted electrode, but not one adhered over the wrong muscle, which presents normal impedance and has to be caught from the signals claim 205unconfirmed.
Whether she can place it without critical use errors claim 172unconfirmed.
Whether one probe geometry serves both stimulation and discriminating EMG claim 131unconfirmed.

What would settle it: A sensing-only study recording vaginal and abdominal EMG with a cleared KPI/HCC device in women being treated for incontinence, with candidate detectors applied offline. The file records that this uses the device within its labelling claim 166unconfirmed, and that the sensing hardware for it may be purchasable claim 149unconfirmed. Otherwise the schedule is set by a prototype probe build claim 114unconfirmed. Running any study with women is a human decision.

If it's false: This is the load-bearing condition (Slide 9). P2 the discriminating trigger exists to provide this discrimination, and without it the device is a single-channel triggered device with an extra electrode.

Slide 5 — Bench evidence before anything lives

Would have to be true: Before any woman uses it, a bench rig would have to show four things. The reference can only remove a trigger. Every reference fault ends in withholding or single-channel fallback. No stimulation current reaches the abdominal electrode in normal condition or single fault. And the decision completes inside the off-time before each burst.

Where it stands:

The fail-safe follows from its premise, and the premise is a design that does not exist yet claim 205unconfirmed, claim 192unconfirmed.
Isolation or disconnection of the abdominal input during each burst, verified by patient auxiliary current and leakage measurement claim 206unconfirmed.
At least 2 s off between bursts on the three summaries read claim 152unconfirmed. Whether a triggered burst runs its whole work time is read from one summary's wording claim 207unconfirmed.
Discrimination itself cannot be benched. It needs recordings in women, which is Slide 4's sensing-only study claim 166unconfirmed.

What would settle it: Fault-injection and IEC 60601 testing on a works-like controller, and the offline detector study. Tier 3 sits immediately downstream: no agent here greenlights any study in women.

If it's false: A reference fault that can fire stimulation makes placement a critical task (Slide 6) and a safety question, not a usability one.

Slide 6 — Making it, and using it

Would have to be true: The system would have to be makeable and supportable as at least five objects with different lifetimes, one of them a per-session skin electrode with a shelf life. And a woman alone at home would have to set up the probe, the electrode and the app correctly, every session, without a skin reaction that ends the course.

Where it stands:

Five objects. A single-patient-use probe, a returnable controller, the app and cloud pathway, connectivity, and a per-session electrode whose demand scales with sessions claim 183unconfirmed, claim 111unconfirmed.
The electrode supply. Cleared electrodes are purchasable claim 184verified. A commodity snap would mean no proprietary electrode claim 208unconfirmed. Labelled shelf life and storage, and a recall for temperature exposure, make kitting and replenishment a supply-chain item claim 185unconfirmed.
Skin. Skin reaction is the most frequent home injury narrative for cutaneous electrodes in MAUDE (Manufacturer and User Facility Device Experience), with no rate claim 186unconfirmed.
Probe and controller. A single-patient probe in a returnable system claim 84unconfirmed; controller reprocessing and redeployment claim 113unconfirmed, claim 141unconfirmed; no terminal sterilisation, shown for one family only claim 117unconfirmed.
Use at home. Self-placement of the reference claim 172unconfirmed. Setting up the probe and app unsupervised claim 65unconfirmed.

What would settle it: A formative usability study with a works-like prototype. A supplier quote and storage spec for the chosen electrode. A skin-contact biological evaluation of its adhesive.

If it's false: Supply failures and skin reactions land on the woman (desirability Slide 4) and on complaint handling. Use errors land on Slide 3's usability file.

Slide 7 — What FTO forbids

Would have to be true: P2 the discriminating trigger keeps EMG triggering, so it does not design around the four live Reider-family grants. It would have to fall outside their issued claims as written. The withhold logic would also have to be unclaimed by anyone else, including the pending Femme application.

Where it stands:

Reider family. Four live US grants on EMG-triggered pelvic floor stimulation, scope never read claim 21unconfirmed. A design that avoids them while keeping the EMG trigger claim 85unconfirmed; P2 the discriminating trigger keeps it. Independent claim 1verified of US10307589 may not recite triggered stimulation at all, but the claim text was never read claim 195unconfirmed. The family's term may run to about 2036 rather than the naive dates claim 196unconfirmed. The holder may be an individual inventor claim 200unconfirmed. Cleared ETS devices post-date most of the filings, and one predicate may have done ETS earlier claim 201unconfirmed.
Femme. A published, possibly pending US application on intravaginal stimulation fired by intra-abdominal pressure, whose claims can still change claim 197unconfirmed.
The withhold logic. No US patent or application claims it claim 198unconfirmed, and no prior publication discloses it claim 199unconfirmed. Both are absence claims that no sweep can verify.

What would settle it: Stage 7 counsel reading the issued claims, file histories, assignment records and maintenance status of the four grants, the Femme US claims, and the two Chinese two-channel leads in full. This is billable legal work, not an agent task.

If it's false: A Reider claim that reads on EMG-triggered stimulation as such is not designed around by adding a condition to the trigger. Factor 4 is then the factor that could fire a kill on evidence, which is Stage 5's arithmetic and not this deck's.

Slide 8 — Where this deck outruns the file

Every condition above has nothing verified behind it. Verified claims appear on Slides 2, 3 and 6, and they settle sub-conditions only. Triggered stimulation through a vaginal probe is clearable. KPI is a Class II 510(k) code. The triggered predicate was cleared without clinical data. Cleared EMG skin electrodes can be bought. None of them says the two channels discriminate, the route holds for P2 the discriminating trigger, or a woman can place the electrode.

The spec has targets. No claim sets them claim 177unconfirmed. Sessions per course are a visible blank claim 183unconfirmed. Settled by a design-input decision.
Two channels separate the manoeuvres. Rests on claim 150unconfirmed, claim 203unconfirmed and claim 204unconfirmed. Settled by a sensing-only study in women.
Withholding changes an outcome, and firing on a strain is harmless. Rests on claim 178unconfirmed and claim 209unverified. Settled by a clinical comparison and a safety literature read.
P2 the discriminating trigger stays a non-clinical 510(k). Rests on claim 188unconfirmed, claim 155unconfirmed and claim 189unconfirmed. Settled by a pre-submission.
Discrimination survives self-placement. Rests on claim 204unconfirmed, claim 205unconfirmed and claim 172unconfirmed. Settled by the sensing study plus formative usability.
The fail-safe and isolation hold on the bench. Rests on claim 205unconfirmed and claim 206unconfirmed. Settled by fault-injection and IEC 60601 testing.
Five objects can be made and supported. Rests on claim 183unconfirmed, claim 185unconfirmed, claim 186unconfirmed and claim 141unconfirmed. Settled by supplier quotes and design verification.
FTO (freedom to operate). Rests on claim 21unconfirmed, claim 195unconfirmed and claim 198unconfirmed. Settled by counsel.

This deck appended no claims today: every condition it states already had a line in the file. That is not good news. Most of those lines are unconfirmed because no connector can reach a bench result, an FDA judgment or a patent's claim text.

Slide 9 — The load-bearing condition

If only one thing from this chair could be checked: whether a two-channel detector, pelvic floor surface EMG from the probe plus an abdominal reference from a skin electrode the woman places herself, separates a voluntary pelvic floor contraction from a Valsalva or abdominal brace with usable specificity, and keeps doing so when she places the electrode imperfectly claim 204unconfirmed.

Why this one: P2 the discriminating trigger is the response to finding that the closed loop is already sold claim 167refuted, and its only claimed difference is this discrimination. Every other slide on this deck builds, clears, makes or protects a function that this condition says has never been measured. It is also the cheapest of the feasibility questions to start on. The file records a sensing-only route that uses a cleared device within its labelling and delivers no stimulation claim 166unconfirmed.

Naming it is not a recommendation, a gate, or a kill.

Words on this page12 abbreviation(s) and code(s), with what each stands for — copied from _shared/plain-language.md

510(k): premarket notification — the FDA route for a device shown to be as safe and effective as one already on the market; months, no clinical trial as a rule
Class II: FDA device class II — moderate-risk devices, usually cleared by 510(k)
De Novo: De Novo classification request — the FDA route for a new, lower-risk device with no comparable product to point at; longer than a 510(k), usually needs clinical data
EMG: electromyography — reading the electrical activity of a muscle
ETS: electrical tissue stimulation — the FDA's name for the product family of stimulators
FDA: US Food and Drug Administration — the US regulator that clears or approves a device before it can be sold
FTO: freedom to operate — whether a product can be sold without infringing someone else's live patent
IP: intellectual property — patents and the other rights that let a company stop others copying it
ISO: International Organization for Standardization — the body whose numbered standards devices and laboratories are tested against
K######: an FDA 510(k) clearance number — the public record of one cleared device; the letter K and six digits
MAUDE: Manufacturer and User Facility Device Experience — the FDA's public database of reported device problems
NMES: neuromuscular electrical stimulation — making a muscle contract with a small electrical current

Generated from knowledge-base/decks/idea-001-feasibility.md by engine/scripts/build_review.py on 2026-10-04. This page is a view — edit the source file and rebuild, never edit here.