Home wearable LED phototherapy with onboard dosimetry — feasibility

The chair: engineering. The people who have to build a light-emitting garment that sits on a newborn for days, measure what it delivers, survive a rental fleet, and produce the evidence a 510(k) (premarket notification) 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-004.md — the number is the claim's position in its ## Claims list, and the tag is copied from that line and never adjusted here. A [no claim] marker means nothing in the file speaks to the condition at all. See docs/deck-spec.md.

What changed since the last build. The E0202 (a HCPCS code for durable medical equipment) descriptor is now verified as naming a photometer claim 4verified, so the onboard sensor is a descriptor requirement of the code the file plans to bill, as well as the differentiator — verified as descriptor wording, not as a payment rule. The business case put on record whether the dosimetry subsystem could ride into an already-cleared host as a module claim 61unconfirmed. This build put two new conditions on record: how FDA (US Food and Drug Administration) would read the dose record claim 74unverified, and whether a battery wrap stays within skin-temperature limits claim 75unverified. Rebuilt from the file, not patched.

Slide 1 — What the device actually has to do

Would have to be true: A wrap-form LED array in the ~450-475 nm band would have to deliver the guideline-specified spectral irradiance over the specified fraction of body surface, on battery, for a full prescribed home episode, across every infant size it is labelled for — where the irradiance target, the body-surface fraction and the episode length are blanks: no claim in the file carries a number for any of them.

Where it stands: The spec is on record and unconfirmed claim 26unconfirmed. The size range that matters is the cleared analog's, 1.5-5.5 kg and 40.5-57.5 cm claim 2verified, and whether one wrap holds emitter-to-skin and sensor-to-skin geometry across it is unconfirmed claim 35unconfirmed. The regulation's identification text — lamps emitting a specific spectral band, under which an infant is placed — matches a treatment-delivery-only use claim 1verified.

What would settle it: The 2022 AAP guideline's irradiance specification and IEC 60601-2-50, both read by a human, to fill the three blanks; then a design specification with numbers in it.

If it's false: There is no device to test. Slides 2 to 5 have nothing to measure against.

Slide 2 — The mechanism that has to hold

Would have to be true: Phototherapy works; what has to hold is that this delivery format, worn at home with no nurse adjusting it, keeps delivered irradiance high enough that in-band infants do not continue to rise — at a rate no worse than the inpatient comparator.

Where it stands: The device-specific risk is under-dosing at home, as distance and body-surface coverage fall claim 14unconfirmed. The one observation of a wearable: 10 of 12 neonates completed within 48 hours, and 2 were switched to conventional intensive phototherapy because bilirubin kept rising claim 41unconfirmed — Dutch, hospital-based, a different device. A US trial of a competing wearable completed with no results posted claim 48unconfirmed. The home-versus-inpatient readmission evidence conflicts claim 39unconfirmed, claim 40unconfirmed, claim 38unconfirmed.

What would settle it: A human read of the Dutch pilot for delivered irradiance in the two failures; the NCT03599258 (a ClinicalTrials.gov registration number) registry entry read directly. Beyond that the evidence is clinical, and a study of infants is not an agent's decision.

If it's false: Factor 6 — the highest Stage 2 draft score in the file, now resting on nothing verified — falls, and the missed-escalation path becomes the device's defining failure.

Slide 3 — What the pathway forces you to build

Would have to be true: A 510(k) against an LBI predicate would have to be achievable on bench photometric data, with the dose record read by FDA as a feature of a phototherapy unit — not as a bilirubin measurement or clinical-decision function that pulls the device onto a second classification.

Where it stands: The launch route is verified: Class II (FDA device class II), 21 CFR (Title 21 of the Code of Federal Regulations) 880.5700, LBI, 510(k), and the verification lapses if any bilirubin measurement claim is added claim 1verified. The measurement route is on a different review panel and different regulations claim 31verified, and the transcutaneous version is unidentified claim 32unconfirmed. How FDA would read a dose-and-adherence record sent to the ordering clinician was never on record; this build put it there claim 74unverified. Clearance on bench data alone is unconfirmed claim 30unconfirmed, and the claim notes the closest analog has a broader indication and claims no dosimetry. The eCFR section carries no enumerated special controls claim 1verified, so there is no special-controls text to design to. The OEM-module route is open claim 61unconfirmed.

What would settle it: A pre-submission meeting with FDA, which is a human act. Before it: a human read of the two 2025-2026 wearable 510(k) summaries for their performance testing, and of FDA's device-modification guidance for the module question.

If it's false: A clinical study enters the plan, Factor 7's cost and calendar change, and if the record is read as measurement, the carried second route becomes the launch route.

Slide 4 — The hardest unknown

Would have to be true: An onboard radiometer would have to measure irradiance actually delivered at the skin — in a garment that moves against a neonate, at both ends of the size range — within a tolerance a reviewer and a payer would accept as a dose record. The tolerance is a blank: no claim carries a number for it.

Where it stands: Unconfirmed claim 27unconfirmed, and the geometry half of it is unconfirmed too claim 35unconfirmed. Whether bench testing alone could settle the accuracy question is unconfirmed claim 64unconfirmed. It now carries weight on two decks: the descriptor of the code the file plans to bill names a photometer claim 4verified, and the cleared wearable claims no dose record claim 2verified.

What would settle it: A bench rig: a reference photometer, a neonatal phantom at the two ends of the size range, the garment in motion. No subjects, no regulator, no payer.

If it's false: The differentiator is gone and what remains is a commodity LED wrap in a class with two cleared wearables. See slide 9.

Slide 5 — Bench evidence before anything lives

Would have to be true: Before any infant wears it, a bench would have to show irradiance mapping across the body-surface fraction, sensor agreement across the size range, battery endurance for a full episode, and skin-contact temperatures within the limits that apply to neonatal phototherapy equipment — all with the wrap in its worst plausible fit.

Where it stands: Irradiance and endurance claim 26unconfirmed; sensor agreement claim 27unconfirmed, claim 35unconfirmed; bench sufficiency claim 64unconfirmed. The thermal condition was never on record for a powered array worn against a newborn's skin for a full episode; this build put it there claim 75unverified. The one trial of a competing wearable tracked skin irritation and temperature control as safety endpoints and posted no results claim 48unconfirmed.

What would settle it: IEC 60601-2-50 read by a human for the test methods and limits; the bench programme itself. What follows a bench is Tier 3: no agent in this repo sets up a study in infants.

If it's false: A thermal or mapping failure sends the emitter layout back to design, and slide 1's spec with it; the under-dosing risk on slide 2 claim 14unconfirmed is what a cooler, dimmer wrap would trade into.

Slide 6 — Making it, and using it

Would have to be true: A rental unit would have to be reprocessed between newborns to a neonatal-skin-contact standard — or be cheap enough to be single-patient — and survive repeated flexing at its connections, and be fitted correctly by a parent at night.

Where it stands: Reprocessing or single-patient use is unconfirmed claim 29unconfirmed. The incumbent blanket was the subject of two recalls for degradation at the pad-to-box connection claim 46unconfirmed — the wear-and-flex mode a garment inherits. Parent operation is unconfirmed claim 18unconfirmed. The whole service model — delivery, setup, training, retrieval — is unconfirmed claim 13unconfirmed.

What would settle it: A cleaning-validation plan and ISO (International Organization for Standardization) 10993 assessment for neonatal skin contact; a flex-cycle test on the connection; a human-factors study with parents, which is a human-subjects decision.

If it's false: The rental model the viability deck assumes cannot run, and a single-patient bill of materials has to fit a payment nobody has established per episode.

Slide 7 — What FTO forbids

Would have to be true: No live, enforceable US patent would have to cover a wearable LED phototherapy garment with integrated delivered-dose measurement — or, if one does, a workaround would have to exist that keeps the dose record.

Where it stands: Unconfirmed, and in substance unexamined: the claim names no patent, and the patent lookup could not run without a PATENTSVIEW_API_KEY claim 28unconfirmed. It is the only claim in the file under the rubric's one automatic-kill factor, in a form factor with two recent cleared entrants claim 2verified.

What would settle it: A PATENTSVIEW_API_KEY supplied by a human, then a landscape search on wearable phototherapy, garment enclosures and onboard radiometric dosimetry, scoped to the named assignees; a legal FTO (freedom to operate) opinion is the eventual settling source and is Tier 3.

If it's false: A blocking patent over the garment is a design constraint; one over the dose record removes the differentiator; and FTO blocked is an automatic kill at Stage 5.

Slide 8 — Where this deck outruns the file

Every condition above with nothing verified behind it — all seven. The verified claims on this deck (the launch classification, the measurement route, the analog's size range and indication, the descriptor wording) describe what the pathway permits and what the code says; none shows that anything here can be built.

• The spec itself: irradiance, body-surface fraction, battery, episode — claim 26unconfirmed — the guideline and IEC 60601-2-50, then a design spec.
• One wrap across the size range — claim 35unconfirmed — optical bench across the range.
• No under-dosing at home — claim 14unconfirmed, claim 41unconfirmed — the pilot read by a human; clinical evidence is downstream.
• FDA reads the dose record as a phototherapy feature — claim 74unverified — put on record by this build; a pre-submission meeting.
• Bench data alone clears — claim 30unconfirmed, claim 64unconfirmed — the wearable 510(k) summaries; a pre-submission.
• The module route — claim 61unconfirmed — FDA modification guidance, read by a human.
• The radiometer measures delivered dose within tolerance — claim 27unconfirmed — a bench rig. See slide 9.
• Skin-contact temperature within limits — claim 75unverified — put on record by this build; a thermal bench test.
• Reprocessing, connection durability, parent fit — claim 29unconfirmed, claim 46unconfirmed, claim 18unconfirmed — validation and human-factors work.
• FTO — claim 28unconfirmed — an API key, then a search.

This is the shape of the minimum viable bench-evidence list Stage 7 would ask for: most of it needs a bench or a human read, not a connector.

Slide 9 — The load-bearing condition

If only one thing from this chair could be checked: whether an onboard radiometer can measure delivered irradiance at the skin, in a garment that moves against a neonate, within a tolerance a reviewer and a payer would accept as a dose record claim 27unconfirmed.

It holds from the last build, and the week's one verified move on this deck adds to it rather than displacing it: the descriptor of the code the file plans to bill names a photometer claim 4verified, so the sensor is now a coding requirement as well as the only differentiator the cleared wearable leaves open claim 2verified. It is also the safety instrument: a wearable has been seen failing to control bilirubin in 2 of 12 infants claim 41unconfirmed, and a trustworthy delivered-dose reading is how a clinician would tell "this infant needs escalation" from "this wrap is not delivering". If it fails, the moat claim 12unconfirmed and the under-dosing mitigation claim 14unconfirmed fail with it. It needs a bench, not subjects, a regulator or a payer — and a measured tolerance is also what a pre-submission on how FDA reads the record would need in hand claim 74unverified.

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