idea-004 · viability 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: the investor, and the operator who has to turn a rental fleet, a van and a same-day delivery promise into a business.
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, the tag is copied from it and never adjusted here. A [no claim] marker means nothing in the file speaks to the condition at all. No number appears here that is not in a claim, and there is no ask slide. See docs/deck-spec.md.
What changed since the 2026-08-25 build — read this first. The file now contains its first real payment figure for E0202 [claim 44: unconfirmed], and an implementation study that names lack of reimbursement options as one of the two main barriers to home phototherapy [claim 42: unconfirmed]. Together they move this deck's load-bearing condition; see slide 9. Rebuilt from the file, not patched.
Would have to be true: A DME supplier would have to rent this to a family and bill a payer under a durable medical equipment benefit — meaning the company's customer is the supplier, not the hospital, the practice, or the family.
Where it stands: The premise is on record and unchecked [claim 22: unconfirmed]. The structural note under it is the one that shapes everything: newborns are not Medicare beneficiaries, so the payers that matter are Medicaid and commercial plans [claim 6: unconfirmed] — which is why the national CMS caches, the tool this repo is best at, answer less here than they do elsewhere in the portfolio. The first payment evidence to arrive came from a state Medicaid schedule accordingly [claim 44: unconfirmed].
What would settle it: Two or three DME operator interviews on who buys, who stocks and who carries the working capital; a state Medicaid DME supplier manual for who may bill.
If it's false: The go-to-market is a hospital capital sale or a cash-pay consumer product, either of which changes the TAM arithmetic and the moat argument entirely.
Would have to be true: A code has to exist, it has to pay an amount that supports a rental episode, and a payer has to cover the service. All three, separately.
Where it stands: One is now partly answered and it is the one that hurts.
What would settle it: The CMS HCPCS Level II and DMEPOS payment files in data/ for the descriptor and the national allowable — read the missing-cache results as missing inputs, never as negative findings. Then state Medicaid DME coverage criteria and a named commercial DME medical policy, by hand, which is where the answer for this population actually lives.
If it's false: Factor 2 has no floor to stand on. Note the difference from the last build: the risk is no longer only that nothing has been checked. One figure has been, and it is small.
Would have to be true: Enough eligible episodes a year, at enough revenue per episode, to be worth a rental fleet.
Where it stands: The denominator is now sourced — 3,628,934 US births in 2024 [claim 36: unconfirmed] — replacing the Generator's estimate [claim 7: unconfirmed]. The two shares that convert births into episodes are still unbacked [claim 8: unconfirmed], [claim 9: unconfirmed], though two US birth cohorts now exist to derive them from [claim 37: unconfirmed], [claim 38: unconfirmed]. Revenue per episode was defined as being on the order of the E0202 rental allowable [claim 10: unconfirmed]; the first observed allowable is $33.71 [claim 44: unconfirmed]. The published base will not help with the economics: the 2026 review of 31 studies found no studies evaluating cost-effectiveness [claim 40: unconfirmed].
What would settle it: A Medicaid claims extract for the home-eligible share; payment E0202 once the cache exists, for the national figure to set beside the state one.
If it's false: A modest, service-heavy revenue pool — and TAM is one of the three hard kill conditions, so this is where a bad answer is fatal rather than discouraging. Nothing has fired: the components are unmeasured, not adverse.
Would have to be true: The regulatory path would have to be cheap and short and the commercial path would have to not need a fleet — and the file's own reasoning says the second half is unlikely.
Where it stands: The estimate is unbenchmarked [claim 16: unconfirmed]. The regulatory half is the best-supported thing in the file — Class II, product code LBI, 510(k) [claim 1: verified], with predicate availability explicitly not the constraint [claim 2: verified] — and the research lane has now enumerated the pool: 10 distinct LBI clearances in the ten years to 2026-08-27, including two wrap/wearable garment entrants cleared in 2025-2026 [claim 47: unconfirmed]. Whether clearance comes on bench photometric data alone is still unchecked [claim 30: unconfirmed]. The commercial half — a rental fleet with same-day delivery, in-home setup, parent training and retrieval — is the part with no evidence at all [claim 24: unconfirmed], and the part the payment figure now presses on.
What would settle it: Read the 510(k) summaries for the two 2025-2026 wearable clearances for whether clinical data was required; a DME operator's unit economics — fleet cost, utilisation, delivery cost per episode, days-to-payment — from an interview or a licensed report. No connector in this repo returns submission cost or review time.
If it's false: A cheap clearance in front of an expensive, slow-to-scale logistics business — the shape investors most often misread as a device play.
Would have to be true: Something other than the LED array would have to be defensible — and something other than the narrowed indication too, because the narrowing is not available as a barrier.
Where it stands: The moat rests on the dosimetry record and the billing tie [claim 12: unconfirmed]. Whether the descriptor's photometer requirement binds for payment [claim 4: unconfirmed] is the mechanism that would convert that record from a feature into a condition of payment — and it is the single most valuable unchecked line on this deck, because it is the only route by which the dose record becomes structural rather than promotional. FTO was never attempted and names no patent number [claim 28: unconfirmed]. The competitive picture is now enumerated rather than asserted [claim 47: unconfirmed], [claim 15: unconfirmed], and it includes two cleared wearables in this exact form factor.
What would settle it: The primary HCPCS descriptor; a payer policy or Medicaid criterion requiring documented delivered irradiance for payment; a patent landscape read on dosimetry logging, which needs credentials supplied before it can run at all.
If it's false: A commodity product in a channel owned by incumbents, competing against a cleared device with a broader label. Note where the damage lands: Factor 5 is blank in the file rather than low, and the file is explicit that adjacent verified evidence was deliberately not converted into coverage for it — so this slide reports adverse context, not a score.
Would have to be true: The wearable home segment would have to be open — or occupied by someone this product can beat on something a payer or a clinician will notice.
Where it stands: It is not open. FDA recorded 10 distinct 510(k) clearances under LBI in the ten-year window, including two wrap/wearable-garment entrants cleared in 2025-2026, BiliWrap (K243372) and SnugLit (K251308) [claim 47: unconfirmed] — and the cleared analog's indications-for-use cover hospital and home use, caregiver-administered, across a wider population than this candidate targets [claim 2: verified]. Whether that incumbent already holds the DME distribution relationships is unbacked [claim 25: unconfirmed]. Two adjacent signals are worth carrying: the Natus neoBLUE blanket was the subject of two FDA recalls for discoloration/degradation of the fiberoptic bundle at the pad-to-box connection [claim 46: unconfirmed] — an incumbent stumble in exactly the wear-and-flex failure mode a garment inherits — and a completed US randomised trial of a competing home phototherapy device posted no results on ClinicalTrials.gov [claim 48: unconfirmed].
What would settle it: A licensed source or human market scan for funding and traction, which no connector covers, and a read on which DME suppliers hold the home-phototherapy rental relationships regionally.
If it's false: Read this one inverted. The risk is not that the field turns out to be occupied — the verified claim already says it is, by a device with a broader cleared indication. The differentiation argument has to be carried entirely by the dose record.
Would have to be true: A single afternoon of human work would have to be able to move the money question from assumption to fact.
Where it stands: Partly spent, and the remainder is now better targeted. Two claims are unconfirmed for a missing local cache rather than a negative answer [claim 3: unconfirmed], [claim 5: unconfirmed] — and those caches were populated on 2026-08-30 and used to re-verify nine claims on idea-001 the same day. This candidate was never re-run against them. The harder truth is that they will answer less here than elsewhere: newborns are not Medicare beneficiaries [claim 6: unconfirmed], so the national DMEPOS file gives a reference price, not the price this business would be paid. The state schedule that does [claim 44: unconfirmed] was found by hand.
What would settle it: Restore the CMS caches per data/README.md and re-run hcpcs E0202 and payment E0202. Then one large-birth-state Medicaid home-phototherapy policy read by hand — which is the leg that actually decides this candidate — and credentials supplied so the FTO line can be attempted at all.
If it's false: Nothing here is false — this slide is a work order, and the largest block of unresolved claims is still blocked on inputs rather than on adverse findings.
Every condition above with nothing verified behind it.
Of the six rubric factors that live in this lens, exactly one carries any verified claim — competitive intensity, scored off the predicate record [claim 2: verified]. The one factor with verified evidence is the one that reports who else is already in the room.
If only one thing from this chair could be checked: whether the E0202 allowable leaves a DME supplier any margin after same-day delivery, in-home setup, parent training and retrieval [claim 24: unconfirmed], [claim 44: unconfirmed].
This is a change from the 2026-08-25 build and the reason is new evidence rather than a new opinion. That build named the supplier's device choice as load-bearing — dosimetry-equipped versus the broader-labelled cleared wearable [claim 34: unconfirmed]. That question is downstream of a more basic one, and two findings put it there. A state Medicaid schedule now gives a real number for what E0202 pays [claim 44: unconfirmed], and it is small. And an implementation study of home phototherapy names lack of reimbursement options as one of the two main barriers to the service existing at all [claim 42: unconfirmed] — not device preference, not clinical doubt, not parent willingness.
If the allowable does not support the service model, the supplier's device choice never arises, because there is no supplier. This deck states the tension and stops there: $33.71 is one state's rate, unconfirmed, and the national allowable has never been retrieved [claim 5: unconfirmed]. Whether it is enough is an operator's unit-economics question that no connector in this repo can answer, and no claim in this file addresses.
Naming it is not a recommendation, a gate, or a kill.