idea-004 · 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.
This is a prescription, home-use wearable LED phototherapy wrap for newborns of at least 35 weeks gestational age with uncomplicated indirect hyperbilirubinemia whose bilirubin sits above the treatment threshold but below the escalation threshold — an alternative to the inpatient stay whose only purpose today is holding a baby under a light. Its distinguishing feature is an onboard radiometer that logs delivered irradiance and time, tied to the billing record, and it deliberately excludes any bilirubin measurement claim to stay on the existing device-class route. The launch regulatory shape is settled on verified record: Class II, 21 CFR 880.5700, product code LBI, via 510(k), with predicate availability confirmed rather than merely assumed [claim 1: verified] [claim 2: verified]. Since this narrative was last built, the file has grown by eighteen claims — a set of unit-economics and go-to-market assumptions drawn from a business case, plus four more surfaced as unsubstantiated messages from an investor, a user, and a payer — and every one of them is tagged unverified [claim 50: unverified] through [claim 67: unverified]; none has moved the composite. Almost everything below the regulatory line remains open — whether the reimbursement code pays enough to run a rental fleet, whether the dosimetry differentiator can actually be built and trusted, whether freedom-to-operate has ever been checked, and whether any discharging clinician would send the baby home rather than keep it in hospital. The one verified competitive fact in the file continues to cut against its own de-risking story: the closest cleared analog already covers this candidate's whole target population and more, with no dose record of its own [claim 2: verified].
Phototherapy itself is not in question — it is the established first-line treatment for neonatal hyperbilirubinemia. What this candidate proposes to change is where it happens: instead of an inpatient admission or readmission whose sole purpose is holding a newborn under a light, a soft LED-wrap garment, rented to the family by a DME supplier, worn for most of the day, with a built-in radiometer recording how much light actually reached the skin and for how long [claim 26: unconfirmed] [claim 27: unconfirmed]. The launch indication was written deliberately narrow — gestational age 35+ weeks, in-band bilirubin, treatment-only — specifically to keep the device on the existing Class II / 21 CFR 880.5700 / product code LBI / 510(k) route rather than the separate submission a bilirubin-measurement claim would trigger; that carried second route is now itself classified, on two different generic types (Class II general test system, Class I neonate-specific), neither of which reaches a non-invasive skin measurement, and its own transcutaneous variant remains unidentified after a twelve-variant name sweep [claim 31: verified] [claim 32: unconfirmed]. Predicate availability is not the constraint: four LBI clearances are confirmed by name, two of them wrap/wearable form factors, and the closest one's Indications for Use were read directly from the FDA summary PDF — hospital and home use, caregiver-administered, 1.5-5.5 kg / 40.5-57.5 cm, with no gestational-age floor and no bilirubin-band ceiling, and no claim to dosimetry [claim 2: verified]. So the narrowed indication this file treats as a de-risking lever buys no distinct regulatory lane the closest analog doesn't already occupy.
Below the regulatory line, almost nothing has a verified answer. Reimbursement rests on HCPCS E0202, whose existence, descriptor, and DMEPOS allowable are all unconfirmed for want of a data/ cache file that has never been supplied for this candidate [claim 3: unconfirmed] [claim 4: unconfirmed] [claim 5: unconfirmed] — and the descriptor's "with photometer" language is, on the file's own account, the most load-bearing line in it, since the onboard photometer is framed as a condition of billing the code at all. Newborns are not Medicare beneficiaries, so the payers that actually decide this — Medicaid and commercial plans — sit outside what this repo's Medicare-anchored connectors can reach at all [claim 6: unconfirmed]. The one real payment figure in the file came from a state Medicaid schedule found by hand rather than a national cache: Louisiana pays $33.71 for E0202, with a second, lower entry of $30.90 [claim 44: unconfirmed]. TAM is built from four decomposed, entirely unconfirmed components — a births figure, a phototherapy-rate share, a home-eligible share, and a revenue-per-episode figure pegged to the still-unknown E0202 allowable [claim 7: unconfirmed] [claim 8: unconfirmed] [claim 9: unconfirmed] [claim 10: unconfirmed] — though the births figure is now a sourced number rather than a round estimate: 3,628,934 US births in 2024 [claim 36: unconfirmed], behind two US birth-cohort studies establishing part of the eligible band [claim 37: unconfirmed] [claim 38: unconfirmed].
The file's most recent addition sits on top of that same reimbursement and TAM uncertainty rather than resolving it. A business case has drafted fourteen further assumptions, all tagged unverified and explicitly recorded as never checked: whether the all-births denominator and the 35-plus-week phototherapy rate can be multiplied without material error [claim 50: unverified]; whether the one observed Louisiana rate is paid per day, per episode, or per rental month [claim 51: unverified]; whether a DME supplier's fully-loaded episode cost sits below that allowable [claim 52: unverified]; whether the price a supplier would pay per fleet unit, the episodes a unit turns per year, and the gross margin at that price together support a device business [claim 53: unverified] [claim 54: unverified] [claim 56: unverified]; whether a material share of episodes would run on a dosimetry-equipped wearable by year three [claim 55: unverified]; whether the dose record itself earns incremental payment from a payer [claim 57: unverified]; whether the lag from first fleet delivery to first paid claim is short enough to finance [claim 58: unverified]; whether DME channel coverage and DMEPOS/Medicaid supplier enrollment in the largest birth states are affordable within a launch and clearance budget [claim 59: unverified] [claim 60: unverified]; whether the dosimetry subsystem could qualify as an OEM module into an already-cleared LBI device without a new full 510(k) [claim 61: unverified]; whether a US comparative study large enough to show a readmission differential is affordable and runnable in a useful calendar [claim 62: unverified]; and whether the Medicaid share of US births is large enough that Medicaid rates, rather than commercial rates, set the effective price of the addressable pool [claim 63: unverified]. Four more claims arrived the same day as unsubstantiated messages rather than a business case, one from an investor, one from a user, and one from a payer, in the file's own words: that the onboard dosimetry accuracy question can be settled by bench testing alone, without human subjects, regulatory interaction, or payer input [claim 64: unverified]; that per-episode delivered-dose records generated in routine use constitute usable clinical evidence for this device's own regulatory and payer submissions [claim 65: unverified]; that newborns excluded from the indication (under 35 weeks, complicated or hemolytic hyperbilirubinemia, at or above the escalation threshold) are appropriately excluded and are treated elsewhere [claim 66: unverified]; and that payer prior authorization can be obtained within the clinical window in which treatment must start [claim 67: unverified]. None of these eighteen claims has been attempted by a Verifier pass.
Competitively, the field is not empty: ten LBI 510(k) clearances were recorded in a ten-year scan window, including two wrap/wearable garment entrants cleared in 2025-2026 [claim 47: unconfirmed], and the incumbent fiberoptic-blanket lineage carries its own history of flex-point failure — two recalls on the Natus neoBLUE for discoloration/degradation at the pad-to-box connection [claim 46: unconfirmed], a failure mode a garment inherits and, by flexing more and carrying a sensor as well as emitters, arguably compounds. Freedom to operate has never been examined in substance: the file's one FTO claim asserts both a position and that a search was performed, and no search has actually run — PATENTSVIEW_API_KEY is unset [claim 28: unconfirmed] — on the rubric's single automatic-kill factor.
On demand, the chain of consent this candidate depends on is unbacked at every link: whether a material share of the eligible band is currently treated as inpatients at all [claim 17: unconfirmed], whether parents can sustain the prescribed dose unsupervised [claim 18: unconfirmed], and — the condition the file returns to most often — whether discharging clinicians and general pediatricians will actually substitute home treatment for inpatient care without an added home bilirubin measurement [claim 19: unconfirmed]. A qualitative implementation study independently names lack of reimbursement options, alongside the difficulty of reaching all relevant clinicians for education, as the two main barriers to home phototherapy happening at all [claim 42: unconfirmed], and adoption where it has been measured is low — 12% of surveyed Canadian paediatricians' centres offer it [claim 43: unconfirmed]. Guideline cover, the other plank of the adoption case, is itself inconsistent in practice: in a 2026 survey of US birthing hospitals, only 24% obtained the post-phototherapy bilirubin check the 2022 AAP guideline recommends [claim 45: unconfirmed]. And the device's central engineering promise now has its first direct evidence, and it cuts both ways: a first-in-human pilot of a wearable phototherapy device in 12 neonates completed treatment within 48 hours in 83%, while 2 of 12 were switched to conventional intensive phototherapy because bilirubin kept rising despite wearable treatment [claim 41: unconfirmed] — a small study of a different device, but the exact under-dosing failure mode this file names as its own device-specific risk [claim 14: unconfirmed], and also the clearest argument yet for why a trustworthy dose record would matter clinically rather than just for billing. The published evidence on home phototherapy generally is thin and mixed: one meta-analysis of 9 studies found higher readmission with home phototherapy than inpatient (RR 4.61) on evidence graded "very low" [claim 39: unconfirmed], while a larger 2026 review of 31 studies found a pooled readmission rate of 3.5% on evidence graded "low," and — bearing directly on both Reimbursement and TAM — found no included study that evaluated cost-effectiveness at all [claim 40: unconfirmed]. Beyond the candidate's own claims, the domain dossier records that almost all of the controlled clinical literature on home phototherapy outcomes is non-US, and that the one published cost-effectiveness analysis in the entire evidence base comes from a single Swedish RCT — a fact the candidate's own claims never picked up [dossier: neonatal-hyperbilirubinemia-home-phototherapy].
intended_use:
[claim 1: verified]
intended_use: [claim 31: verified]
data/ cache rather than any adverse finding
[claim 3: unconfirmed] [claim 4: unconfirmed] [claim 5: unconfirmed]
unverified and
explicitly recorded as never checked:** whether the all-births-times-rate arithmetic is
materially correct [claim 50: unverified]; the billing unit behind the one observed Louisiana
rate [claim 51: unverified]; whether a supplier's fully-loaded episode cost sits below the
allowable [claim 52: unverified]; the fleet unit's purchase price, episode turns per year, and
resulting gross margin [claim 53: unverified] [claim 54: unverified] [claim 56: unverified];
the dosimetry-equipped wearable's addressable share by year three [claim 55: unverified];
whether the dose record itself earns incremental payment [claim 57: unverified]; the financing
lag to first paid claim [claim 58: unverified]; the cost and affordability of DME channel
coverage and DMEPOS/Medicaid supplier enrollment in the largest birth states
[claim 59: unverified] [claim 60: unverified]; whether the dosimetry subsystem could qualify as
an OEM module without a new full 510(k) [claim 61: unverified]; the affordability and calendar
of a US comparative readmission study [claim 62: unverified]; and whether Medicaid, not
commercial, rates set the effective price of the pool [claim 63: unverified]
All three decks report ok under python3 engine/scripts/check_decks.py — every claim reference they carry resolves and matches the tag the candidate holds right now, and their source_updated: frontmatter (2026-09-03) agrees with the candidate's updated:. That is a mechanical check, not a content one, and it is worth naming plainly here: the deck bodies themselves say they were rebuilt on 2026-08-30, and nothing in any of the three mentions or quotes any of the eighteen claims added on 2026-09-03 [claim 50: unverified] through [claim 67: unverified] — the business-case unit-economics chain and the four unsubstantiated messages. check_decks.py cannot and does not check completeness; a narrative built after those claims landed should say so rather than let a clean checker run stand in for currency. All three independently make the same observation about the file's evidentiary shape as it stood on 2026-08-30: every verified claim in it concerns permission — the launch classification, the predicate pool and its labels, and the classification of a measurement route the candidate is not currently taking — and none of them is evidence of capability or demand. That observation still holds: none of the new business-case claims is verified either.
The desirability deck turns on the same condition it turned on at its last build, sharpened by two findings surfaced by the research lane rather than displaced by them: a qualitative study naming clinician-education reach as a barrier, and a physician survey showing low measured adoption of home phototherapy generally.
"whether discharging clinicians and general pediatricians will substitute home phototherapy for inpatient treatment in the eligible band without an added home bilirubin measurement" — the load-bearing condition of the desirability deck (built 2026-08-30).
The viability deck's load-bearing condition changed at its last build — not from a new opinion but from new evidence. It previously turned on whether a DME supplier would choose this device over the already-cleared, more broadly-labelled competitor; a state Medicaid payment figure and an implementation study naming reimbursement as a top barrier pushed it to a more basic question underneath that one. The business-case claims added since then sit squarely underneath this same condition — supplier margin, fleet economics, channel-build cost, and financing lag are now each named as a separate unverified assumption rather than folded into one deck slide, but the deck itself has not been rebuilt to reflect that decomposition.
"whether the E0202 allowable leaves a DME supplier any margin after same-day delivery, in-home setup, parent training and retrieval" — the load-bearing condition of the viability deck (built 2026-08-30).
The feasibility deck turns on the device's only remaining differentiator once the predicate record is read in full: the cleared analog already covers this candidate's whole target population without a dose record, so the dosimetry claim is the entire case, and a first-in-human pilot of a different wearable device has shown the exact failure mode a trustworthy dose record would need to catch. One of the new unsubstantiated-message claims speaks directly to this condition — that bench testing alone, without human subjects, regulatory interaction or payer input, could settle the dosimetry-accuracy question [claim 64: unverified] — but the deck predates that claim and does not address it.
"whether an onboard radiometer can measure delivered dose at the skin, in a garment that moves, within a defensible tolerance" — the load-bearing condition of the feasibility deck (built 2026-08-30).
Where the three do not agree is on what breaks the idea first. The desirability deck's condition would render the device irrelevant regardless of how well it is built; the viability deck's condition would mean there is no supplier to build a business on top of, regardless of who prescribes it; and the feasibility deck's condition is the one thing among the three that is cheap to test on a bench, with no subject, no regulator, and no payer required — though one of the newest claims in the file is an unsubstantiated assertion that this is exactly sufficient [claim 64: unverified], which is a different thing from it having been shown.
If an onboard radiometer can be shown to measure delivered dose accurately enough to trust, and the E0202 allowable turns out to support a rental fleet after delivery, setup, training and retrieval — does that combination actually get a discharging clinician to send an in-band newborn home instead of keeping it in hospital, given that the closest cleared competitor already occupies the same home-wearable indication across a wider population with no dose record of its own, that reimbursement — not device performance — is the barrier independently named by the one implementation study in the file, and that the fleet economics beneath that margin question — unit price, utilization, channel-build cost, OEM-module regulatory status, and time to first paid claim — are now named as fourteen separate unverified business-case assumptions rather than settled numbers?