idea-004 · business case

Who pays, from which budget, how much, and 2-4 commercial routes to first revenue — every figure below traces to a copied claim or a declared assumption, and the routes are laid out unranked. Choosing a commercial strategy is a Tier 3 act, so this page presents options and stops.

Home-use wearable LED phototherapy system with onboard irradiance dosimetry for outpatient neonatal hyperbilirubinemia, coded to E0202 — business case

Intended use: For use in the home under prescription to deliver phototherapy to newborns of at least 35 weeks gestational age with uncomplicated indirect hyperbilirubinemia whose total serum bilirubin is above the phototherapy threshold but below the escalation-of-care threshold, as an alternative to inpatient readmission for phototherapy.

Who pays

The budget-holder chain has two cheques in it, and they are written by different people for different things. That is the structural fact that shapes everything below: the party who buys the device is not the party who reimburses the therapy, and the second one sets the ceiling on the first.

Writes the cheque (for the device): the DME supplier who rents the unit to the family and bills the payer — not the hospital, not the pediatric practice, and not the family directly [claim 22: unconfirmed]. This is a deck-surfaced assumption in the candidate that adoption returned unverifiable against; the domain dossier likewise records that no source located names the actual purchaser of a home phototherapy unit in the US market. So the identity of the buyer — the first link in the chain — is itself unresolved.
Writes the cheque (for the therapy): a state Medicaid program or a commercial plan. Newborns are not Medicare beneficiaries, so the Medicare-anchored connectors in this repo reach only part of the question [claim 6: unconfirmed]. One state Medicaid program is on record paying E0202 at all [claim 44: unconfirmed]; whether the largest US birth states publish a coverage policy for home phototherapy for newborns is unresolved [claim 23: unconfirmed].
Out of which budget line: the DME supplier's rental-fleet capital budget, recovered from a DMEPOS allowable billed under HCPCS E0202. Every leg of that code is unresolved in this repo: the descriptor [claim 3: unconfirmed], whether "with photometer" is binding for payment rather than merely descriptive [claim 4: unconfirmed], and whether a published fee-schedule allowable exists nationally [claim 5: unconfirmed]. Re-run this build: hcpcs E0202, payment E0202 and coverage E0202 each returned "No local cache at data/…" — nothing was searched, which is a missing input, not a negative finding.
Against which comparator: the fiberoptic phototherapy blanket already sitting in that supplier's fleet (BiliBlanket / neoBLUE lineage), and now two cleared wrap/wearable garments in the same product code [claim 47: unconfirmed]. The closest of those, SnugLit (K251308), carries a cleared indication broader than this candidate's — hospital and home, caregiver-administered, no gestational-age floor and no bilirubin-band ceiling [claim 2: verified]. The comparator is therefore not a gap in the market; it is an incumbent plus two recent entrants, one of which already occupies the whole target population.
Stops buying instead: at the supplier, a fleet slot — one wearable displaces one blanket, so this is substitution spend inside an existing budget rather than new spend. At the payer, the thing displaced is an inpatient stay or a readmission whose only purpose is holding the baby under a lamp [claim 17: unconfirmed]. Note that the published evidence on that displacement points in two directions at once: a US biliblanket cohort found lower readmission with home treatment [claim 38: unconfirmed], while a 2024 meta-analysis found readmission higher with home phototherapy [claim 39: unconfirmed]. Nobody in this chain stops buying anything on the strength of the dose record specifically — no claim in the file establishes that a dose record has a budget line of its own [claim 33: unconfirmed].

The chain's weakest link is its first one. A device whose buyer is an unverified assumption and whose payer is a set of state programs this repo's connectors cannot reach is not a device with a hard commercial problem; it is a device whose commercial problem has not yet been located.

The money model

Decomposed, not asserted. Every line resolves to a copied claim or a declared assumption, and the composite at the bottom is arithmetic over unconfirmed inputs — not evidence.

Line	Figure	Backing
Annual US births (denominator)	3,628,934 in 2024, up 1% from 3,596,017 in 2023	[claim 36: unconfirmed]
Share of those births at ≥35 weeks gestation, so the two rows can be multiplied	not established anywhere in the file	[assumption A1]
Share of newborns ≥35 weeks who receive phototherapy	8-12%	[claim 8: unconfirmed]
Independent read on the post-discharge slice specifically	1.8% (2,623 of 148,162 infants, KPNC 2012-2017) exceeded the phototherapy threshold after discharge	[claim 37: unconfirmed]
Share of phototherapy-treated newborns clinically eligible for home treatment	25-40%, the Generator's own assumption	[claim 9: unconfirmed]
Addressable home-phototherapy episodes per year (US)	~72,600 at the low end to ~174,200 at the high end; ~118,000 at midpoints	arithmetic over [claim 36: unconfirmed] × [claim 8: unconfirmed] × [claim 9: unconfirmed], commensurability per [assumption A1]
Medicaid share of that pool (the fraction priced at Medicaid rather than commercial rates)	the file cites "roughly 40%+" and states the figure is itself uncited	[claim 6: unconfirmed], quantified as [assumption A14]
Allowable to the supplier, one state	$33.71 (Louisiana Medicaid DMEPOS, type-of-service modifier 09, rate effective 2021-09-01) and a second entry at $30.90 (modifier 07, effective 2012-07-01)	[claim 44: unconfirmed]
Billing unit of that allowable — per day of rental, per rental month, per episode, or purchase	unresolved; the file records the rate and not its unit	[assumption A2]
Episode length	2-4 days plus setup	[claim 10: unconfirmed]
Gross revenue per episode to the supplier	$33.71 read per episode; $67.42-$134.84 read per day over a 2-4 day episode	[claim 44: unconfirmed] × [claim 10: unconfirmed] under [assumption A2]
Supplier-side revenue pool, whole US, at that one state's rate	~$2.4M/yr (low volume, per-episode reading) to ~$23.5M/yr (high volume, per-day reading)	arithmetic over the rows above — [claim 36: unconfirmed] × [claim 8: unconfirmed] × [claim 9: unconfirmed] × [claim 44: unconfirmed] × [claim 10: unconfirmed], under [assumption A1] and [assumption A2]; no claim in the file asserts this composite
Supplier's fully-loaded cost per episode — same-day delivery, in-home setup, parent training, retrieval, reprocessing	not established	[assumption A3]
Whether any margin survives that cost	asserted, not measured	[claim 24: unconfirmed]
Price a DME supplier pays per fleet unit	not established	[assumption A4]
Episodes per fleet unit per year (turns), which decides whether a unit ever repays its price	not established	[assumption A5]
Share of addressable episodes served on this device by year 3	not established	[assumption A6]
Gross margin on the device at that price	not established	[assumption A7]
Incremental payment earned by the dose record itself — a modifier, a higher-paying code, or a payer documentation requirement	not established; the file's moat rests on this [claim 12: unconfirmed]	[assumption A8]
Reprocessing between rental patients, without which the rental model has no unit economics at all	unresolved for both the reprocessing route and the single-patient-BOM route	[claim 29: unconfirmed]

Read the largest figure first. It is the supplier-side revenue pool, and it is arithmetic over one state's fee schedule, two Generator assumptions and an unresolved billing unit. Its upper bound is smaller than the lower bound of many single-hospital capital budgets. The manufacturer's share of it is smaller again by [assumption A4] and [assumption A7], and this candidate would be competing for that share against an incumbent blanket and two cleared wearables [claim 47: unconfirmed], [claim 15: unconfirmed].

One published finding sits directly on the money model and is worth reading beside it: a 2026 mixed-methods systematic review of 31 studies [claim 40: unconfirmed] found no studies at all evaluating cost-effectiveness of home phototherapy, and rated evidence quality low for all outcomes. The economic argument this business case would need does not exist in the published base.

The two connectors that would ordinarily carry a business case both refuse, and the verdicts go on the page rather than being routed around:

python3 -m connectors market "US home phototherapy episodes for neonatal hyperbilirubinemia represent an addressable market of 3.6M births x 8-12% phototherapy rate x 25-40% home-eligible share x the E0202 rental allowable per episode" → unverifiable — "Market-size claim (...) has no free, authoritative, machine-queryable source. Stays unconfirmed permanently unless a human attaches a licensed report and edits the claim by hand."
python3 -m connectors adoption "A DME supplier stocking a home phototherapy rental fleet will choose a dosimetry-equipped wearable over the already-cleared SnugLit wearable (K251308)" → unverifiable — "Adoption/preference claim (...) asserts what clinicians, payers, or patients would do. No free, authoritative, machine-queryable source answers that — it needs primary research (interviews, survey, published preference study). Stays unconfirmed permanently unless a human attaches real research and edits the claim by hand. Do NOT substitute a model's own estimate of stakeholder behaviour: a synthetic stakeholder opinion is a fabricated claim."

The connector's own suggested rescopes — coverage for the comparator, procedures for volume — are unavailable here. Both returned missing-cache unconfirmed on this build, and procedures is structurally inapplicable regardless, because this population has no Medicare exposure [claim 6: unconfirmed].

Path to first revenue

First dollar comes from: the first DME supplier that buys or leases fleet units — a device sale into a rental channel, not a reimbursement event [claim 22: unconfirmed]. The payer's dollar arrives later and only sets the ceiling.
What must exist first, each attributed:
- 510(k) clearance under 21 CFR 880.5700 / product code LBI [claim 1: verified], against a predicate pool that is not the constraint [claim 2: verified] — but on the untested premise that FDA accepts bench photometric data alone for a narrowed indication carrying a dosimetry feature the predicate does not claim [claim 30: unconfirmed].
- A radiometer that measures delivered dose at the skin within a defensible tolerance [claim 27: unconfirmed], holding emitter- and sensor-to-skin geometry across the 1.5-5.5 kg / 40.5-57.5 cm range the cleared analog spans [claim 35: unconfirmed].
- A billable code with a confirmed descriptor and a confirmed allowable [claim 3: unconfirmed], [claim 4: unconfirmed], [claim 5: unconfirmed], and at least one payer on record paying it [claim 44: unconfirmed].
- A reprocessing or single-patient-BOM answer, without which no rental fleet exists [claim 29: unconfirmed].
- An FTO position: no patent search has ever been attempted, on the rubric's one automatic-kill factor, in a form factor that has just drawn two entrants [claim 28: unconfirmed].
Calendar to first dollar: 12-24 months [claim 16: unconfirmed] — the file's own estimate for a 510(k) against an existing LBI predicate with bench photometric testing and no clinical study.
Capital to first dollar: low single-digit millions on the same estimate [claim 16: unconfirmed]. That estimate explicitly excludes the go-to-market leg — rental fleet, logistics, days-to-payment — which the file states has no public source at all.
Fleet capital before the first episode is billed: units × price per unit [assumption A4], sized by turns [assumption A5].
Lag from first fleet delivery to first paid claim — payer enrollment, prior authorization, claims cycle: not established [assumption A9].
Cost and calendar to build DME channel coverage in the target states: not established [assumption A10]. The incumbent-relationship question cuts against a standing start: the cleared wearable's manufacturer may already hold or be building the distribution relationships this candidate needs [claim 25: unconfirmed].

Routes

Four, laid out and not ranked. Choosing a commercial strategy is a Tier 3 act — the same class of decision as choosing what to shortlist — so this section presents options and stops.

Route A — Sell fleet units into the existing DME rental channel under E0202

The model the candidate file already describes: a device company selling hardware to DME suppliers who rent to families and bill Medicaid or commercial plans [claim 22: unconfirmed].

Requires: 510(k) clearance [claim 1: verified]; a confirmed E0202 descriptor and a confirmed allowable [claim 3: unconfirmed], [claim 4: unconfirmed], [claim 44: unconfirmed]; a reprocessing answer for a rental fleet [claim 29: unconfirmed]; and distribution relationships with suppliers who already rent biliblankets [claim 25: unconfirmed].
Costs: 12-24 months and low single-digit millions to clearance [claim 16: unconfirmed], plus fleet unit price [assumption A4], turns to payback [assumption A5], and channel build [assumption A10].
Killed by: the E0202 allowable, on whichever billing unit applies [assumption A2], leaving the DME supplier no margin over delivery, setup, training and retrieval at achievable episode volumes — i.e. [claim 24: unconfirmed] resolving false against [assumption A3]. If the channel cannot make money renting it, the channel does not stock it, and there is no other buyer in this route.

Route B — Own the service: enrol as the DMEPOS supplier and bill payers directly

Vertically integrated — build the fleet, run the logistics, bill the state programs. Treats the file's own assumption that the binding constraint is the service model rather than device performance [claim 13: unconfirmed] as the thing to own rather than the thing to sell into.

Requires: DMEPOS supplier enrollment and accreditation, plus per-state Medicaid enrollment in the target birth states; a published or negotiable coverage pathway for home phototherapy in those states [claim 23: unconfirmed]; and family out-of-pocket exposure low enough that episodes are not declined at the door [claim 21: unconfirmed].
Costs: everything in Route A [claim 16: unconfirmed], [assumption A4], [assumption A5], plus enrollment and accreditation [assumption A11], the claims-cycle lag before cash arrives [assumption A9], and per-episode service delivery [assumption A3].
Killed by: the largest US birth states publishing no coverage policy for home phototherapy for newborns and operating no route to one — [claim 23: unconfirmed] resolving false. Without a payer policy there is nothing to bill, and this route has no hardware-sale fallback because the fleet is on its own balance sheet. A qualitative implementation study naming lack of reimbursement options as a main barrier to home phototherapy existing at all [claim 42: unconfirmed] is the nearest evidence to that finding, and it is non-US.

Route C — Licence the dosimetry subsystem into an already-cleared wearable

Ship the differentiator, not the garment: an OEM radiometer-plus-dose-record module into SnugLit (K251308) or BiliWrap (K243372), both already cleared and neither claiming dosimetry [claim 2: verified], [claim 47: unconfirmed].

Requires: a radiometer validated as a standalone module [claim 27: unconfirmed] that holds geometry across the indicated size range [claim 35: unconfirmed]; an FTO position on integrated dosimetry [claim 28: unconfirmed]; and a partner who believes the dose record is worth carrying — the moat premise [claim 12: unconfirmed].
Costs: no fleet capital and no channel build, and the rental reprocessing question [claim 29: unconfirmed] transfers to the host-device holder rather than disappearing; in their place, a regulatory scoping and module-validation spend [assumption A12] against the same clearance-cost baseline [claim 16: unconfirmed]. Revenue in this route is a royalty or transfer price set by a partner, not a device price, so [assumption A4] and [assumption A7] are replaced rather than answered.
Killed by: bench data showing the sensor cannot hold sensor-to-skin geometry across the 1.5-5.5 kg / 40.5-57.5 cm range, so a logged dose does not mean the same thing at both ends of it — [claim 35: unconfirmed] resolving false. There is nothing to licence if the record is not comparable between patients.

Route D — Sell the dose record to health systems as a readmission-avoidance instrument

Bypass the DME channel: contract with birth hospitals and health systems, where the budget line is avoided readmission rather than a rental allowable, and the device travels with the record.

Requires: a US comparative readmission figure attributable to home phototherapy with a dose record; discharging clinicians willing to substitute home treatment for inpatient treatment without an added home bilirubin measurement [claim 19: unconfirmed]; and ordering clinicians who act on the record rather than filing it [claim 33: unconfirmed] — in a population where only 24% of surveyed US birthing hospitals obtained the guideline's post-discontinuation bilirubin check [claim 45: unconfirmed].
Costs: a US comparative study, benchmarked against the one registered US trial in this space, NCT03599258, which enrolled 72 and posted no results [claim 48: unconfirmed] — scoped as [assumption A13], on top of [claim 16: unconfirmed].
Killed by: a US comparative study returning no readmission advantage for home phototherapy with a dose record. The pooled non-US base already points that way — 4.61x higher readmission with home phototherapy across 9 studies [claim 39: unconfirmed] — and the same evidence base contains a first-in-human wearable pilot in which 2 of 12 neonates were switched to conventional intensive phototherapy because bilirubin kept rising [claim 41: unconfirmed]. If the readmission differential is absent or negative, there is no health-system budget line to sell into.

Assumption register

Every assumption cited above, with what would settle it and who could run it. These are written back to the candidate as unverified claims — an assumption that lives only in a business case is invisible to the Verifier.

#	Assumption	Falsifier	Owner
A1	The share of annual US births that are at ≥35 weeks gestation is high enough that the all-births denominator and the ≥35-week phototherapy rate can be multiplied without material error	NCHS/CDC WONDER natality final birth data for 2024, gestational-age distribution — free and authoritative, outside this repo's connector set, so a human attachment	Research lane (the same human attachment claim 7's note already names)
A2	The Louisiana Medicaid E0202 allowable is paid per day of rental rather than per episode, per rental month, or as a purchase	The Louisiana Medicaid DMEPOS Fee Schedule (Report RF-0-76D) read for its own unit column, plus the state provider manual on what type-of-service modifiers 07 and 09 each mean for a rental episode — the document is already cited in the candidate file	Reimbursement analyst
A3	A DME supplier's fully-loaded cost to deliver one home phototherapy episode — same-day delivery, in-home setup, parent training, retrieval, reprocessing — is below the allowable for that episode	Cost interviews or a P&L extract from 5-10 DME suppliers currently renting fiberoptic biliblankets in a target state	Commercial / business development
A4	The price a DME supplier would pay per fleet unit is high enough to support a device business at achievable volumes	Revealed-behaviour twin, not a willingness-to-pay survey: actual acquisition prices paid for the incumbent fiberoptic blanket and for the two cleared wearables (BiliWrap K243372, SnugLit K251308), from distributor price lists, state purchasing schedules or supplier invoices	Commercial / business development
A5	A fleet unit turns enough episodes per year to repay its acquisition price within the fleet's useful life	Units-on-hand against annual billed E0202 episodes, from one DME supplier's existing biliblanket fleet in a defined service area	Commercial / business development
A6	A material share of addressable episodes would be served on this device by year 3	Revealed-behaviour twin: fleet composition at named suppliers before and 12-24 months after the last two comparable LBI launches (BiliWrap, decided 2025-05-29; SnugLit, decided 2026-01-15), and the stated reason for each substitution	Commercial / business development
A7	Gross margin on the device at the achievable fleet price supports the business	Bill-of-materials cost roll-up from a contract manufacturer for the wrap, LED array, radiometer, battery and enclosure at stated annual volumes	Engineering + operations
A8	The logged irradiance-and-adherence record earns incremental payment — a modifier, a higher-paying code, or satisfaction of a payer documentation condition	The CMS HCPCS Level II Alpha-Numeric file plus the target-state Medicaid DMEPOS provider manual, read for any dose or documentation condition attached to E0202 and for any add-on	Reimbursement analyst
A9	The lag from first fleet delivery to first paid claim — payer enrollment, prior authorization, claims cycle — is short enough to finance	Days-to-payment from an operating DME supplier's aged-receivables report on E0202 claims in a target state, plus that state's published prior-authorization requirement for DMEPOS	Reimbursement analyst
A10	DME channel coverage in the largest US birth states can be built at a cost and calendar the clearance budget can absorb	Count of DME suppliers actually billing E0202 in a target state, from the state's Medicaid provider-billing data or T-MSIS, plus the commercial terms of a signed distribution agreement in a comparable DMEPOS category	Business development
A11	DMEPOS supplier enrollment and accreditation, plus per-state Medicaid enrollment in the target birth states, is affordable within a launch budget	Published accreditation standards and surety-bond requirements, the enrollment application fee schedule for each target state, and a quote from a CMS-approved accrediting organisation	Reimbursement / regulatory operations
A12	The dosimetry subsystem can be qualified as an OEM module into an already-cleared LBI device without a new full 510(k) for the host device	A regulatory consultant's written scoping opinion on whether adding a radiometer to a cleared LBI device is a letter-to-file, a special 510(k), or a new submission, plus a bench-validation quote for the module	Regulatory consultant + engineering
A13	A US comparative study large enough to produce a readmission differential attributable to home phototherapy with a dose record is affordable and can be run in a useful calendar	A CRO protocol scoping and quote against that endpoint, benchmarked to NCT03599258 (72 enrolled, completed 2021-05-27, no results posted)	Clinical
A14	The Medicaid share of US births is large enough that Medicaid rates, not commercial rates, set the effective price of the addressable pool	NCHS natality "source of payment for delivery" tabulation for 2024 — free and authoritative, a human attachment, and the direct check on the "roughly 40%+" figure the candidate itself flags as uncited	Research lane

Where this case outruns the file

Every figure with no claim behind it, gathered as work items.

The supplier-side revenue pool (~$2.4M-$23.5M/yr) — arithmetic over one state's fee schedule and two Generator assumptions. No claim in the file asserts a national pool, and market returns unverifiable on it permanently.
The addressable episode count (~72,600-174,200/yr) — a product of three unconfirmed claims, none of which is a floor, multiplied across denominators the file never establishes are commensurable (A1).
The billing unit of the E0202 allowable (A2) — the file records a rate and never its unit, and the difference between a per-day and a per-episode reading is roughly 4x on every revenue figure in this case.
The price a DME supplier pays per unit (A4), fleet turns (A5), gross margin (A7) — nothing in the candidate file speaks to any of the three, and together they are the entire device P&L.
Whether the dose record is paid for at all (A8) — the candidate's moat rests on the dosimetry record tied to the billing record [claim 12: unconfirmed], and no claim in the file establishes that any payer pays anything extra for it, or requires it.
Every service-delivery cost (A3), enrollment cost (A11), channel-build cost (A10), module qualification cost (A12) and study cost (A13) — the candidate's own capital estimate [claim 16: unconfirmed] covers the 510(k) and explicitly excludes the go-to-market leg, which is where every one of these sits.
Payer mix (A14) — the "roughly 40%+" Medicaid birth share is flagged as uncited in the file itself and is not carried forward here as a figure.
The lag to first paid claim (A9) — no claim in the file speaks to working capital or days-to-payment.

Generated from knowledge-base/commercial/idea-004.md by engine/scripts/build_review.py on 2026-09-09. This page is a view — edit the source file and rebuild, never edit here.