idea-012 · 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.
The chair: the engineers who have to make a haired-skin adhesive hold a subcutaneous glucose sensor, and make a calibration algorithm correct for feline and canine physiology rather than the human physiology the underlying chemistry was built against.
How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-012.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 this. See docs/deck-spec.md.
Would have to be true: A sensor placed subcutaneously by a licensed veterinarian or veterinary technician would have to deliver continuous interstitial glucose readings, in clinic and at the owner's home, accurate enough for both insulin-dose titration and hypoglycemia detection, as an alternative to serial blood-glucose-curve sampling by venipuncture or lancet.
Where it stands: That is the intended-use statement as written in the candidate frontmatter, but no performance specification exists anywhere in the file to turn it into requirements — no target MARD, no sensor-life requirement, no sampling-interval requirement [no claim]. The mechanism it is built on is the decades-old glucose-oxidase amperometric sensing chemistry, whose specific historical attribution (Clark and Lyons 1962, Updike and Hicks 1967) is itself unconfirmed by the one review retrieved at abstract depth [claim 5: unconfirmed], [claim 16: unconfirmed].
What would settle it: A specification document naming target accuracy, sensor life, and sampling interval, ideally benchmarked against the cleared human-analog predicates' accepted parameter ranges — none of which exists yet in this file.
If it's false: There is no spec to design or test the calibration algorithm and adhesive against, and both open engineering conditions below stay open indefinitely.
Would have to be true: The species-specific calibration algorithm would have to correct two documented biases in the underlying off-label sensing technology when applied to dogs and cats: overestimation of blood glucose in marked hypoglycemia (<55 mg/dL) in cats, and skin-thickness-dependent accuracy loss in dogs.
Where it stands: Both source biases are real and sourced, though unconfirmed at Stage 3 depth. A hyperinsulinemic-hypoglycemic clamp study in 7 purpose-bred cats found the off-label sensor underestimates glucose in euglycemia and mild hypoglycemia but overestimates it in marked hypoglycemia (<55 mg/dL) [claim 12: unconfirmed]. A separate study of 14 client-owned diabetic dogs found clinical accuracy only in dogs with skin thickness above 5mm, with low accuracy requiring cautious interpretation below that [claim 14: unconfirmed]. A broader claim reports the class is reliable at normal/high glucose but more variable at low glucose generally [claim 4: unconfirmed]. Whether this candidate's own calibration algorithm corrects either bias is this deck's own write-back and entirely untested [claim 30: unverified].
What would settle it: A bench or clamp study — directly reusing the design of the cited feline clamp study [claim 12: unconfirmed] — run on this candidate's actual calibration algorithm, specifically across the hypoglycemic range and across a range of skin thicknesses in dogs.
If it's false: The differentiating engineering claim of this candidate — a calibration algorithm tuned for feline and canine kinetics rather than human ones — reduces to relabeling a human sensor without demonstrated correction, and the desirability deck's vet-trust condition [claim 27: unverified] fails by construction.
Would have to be true: Ordinarily, regulatory class sets the build spec — bench standards, biocompatibility protocol, software documentation level. Here, the question is whether anything forces one at all.
Where it stands: Verified, and unusual: because this device is veterinary-only, no FDA premarket review pathway attaches to it — established via independent verification of four CFR sections (807.65(b), 807.20(a), 807.81(a), 814.1(a)) all scoping their duties to "human use" or exempting veterinary-only manufacturers outright [claim 1: verified]. That means nothing in this file forces a bench-testing standard, a biocompatibility protocol (such as ISO 10993), or a software documentation level — adopting a human-CGM-class standard such as IEC 60601-2-10 would be a voluntary engineering choice, not a submission requirement [no claim].
What would settle it: Nothing further regulatorily — this reading is independently confirmed across four CFR sections. What remains open is which standard, if any, the company voluntarily adopts as its own bench-evidence bar, which the file does not record.
If it's false: If this candidate turned out to require an FDA pathway after all, every cost and timeline assumption on the viability deck's capital-to-first-dollar slide [claim 29: unverified] would need to be rebuilt from a materially higher floor.
Would have to be true: The species-specific haired-skin adhesive would have to hold a subcutaneous filament sensor against clipped or fur-parted skin for most of its labeled wear period, across home, barn, and clinic environments, without the detachment problems documented for the off-label analog's human-designed adhesive.
Where it stands: The off-label baseline is real and unfavourable: 70% (46/66) of FreeStyle Libre sensors stopped functioning in situ before the 14-day label end in diabetic cats, median functional life 8.3 days [claim 13: unconfirmed]. This candidate's own adhesive performance is entirely unestablished [claim 26: unverified], and its biocompatibility for repeated haired-skin contact is separately unestablished [claim 31: unverified].
What would settle it: A wear-time and adhesion bench-and-field study on the candidate's actual adhesive, directly benchmarked against the 8.3-day / 70%-failure baseline; ISO 10993 biocompatibility testing on the wetted and adhesive-contact materials.
If it's false: The product inherits the off-label analog's most common real-world failure mode, and the desirability deck's Slide 2 (what they'd have to do differently) collapses — there is nothing left for the species-specific form factor to have improved.
Would have to be true: Before any dog or cat wears this sensor, a bench rig would have to show the calibration algorithm correcting the documented biases (Slide 2) and the adhesive holding to spec (Slide 4).
Where it stands: No bench claim exists in the file for this candidate [no claim]. Both conditions this deck ties bench evidence to remain unverified: [claim 30: unverified], [claim 31: unverified]. What exists instead is the off-label analog's field literature — real, sourced, but about a different, human-calibrated, human-adhesive product [claim 11: unconfirmed] through [claim 16: unconfirmed].
What would settle it: IEC 60601-2-10 and ISO 10993 define most of the test battery if voluntarily adopted (Slide 3); the off-label analog's clamp-study design [claim 12: unconfirmed] is a directly reusable protocol for testing this candidate's own calibration algorithm.
If it's false: Any move to placing this device in a live animal is premature. The boundary is explicit: what happens after bench evidence is a Tier 3 decision, and no agent here makes it.
Would have to be true: The sensor, reader, and app would have to be manufacturable at a cost consistent with a premium-but-affordable owner price, sterilizable or single-use as appropriate, shelf-stable, and usable by an owner at home or a vet tech in a barn or exam room without specialist support.
Where it stands: No manufacturing, sterilization, shelf-life, or human-factors claim exists in the file [no claim]. The nearest cost signal is a competitor's single-clinic retail price, $75.08 per 14-day sensor [claim 19: unconfirmed] — a market price, not a manufacturing cost, and not this candidate's own figure. Biocompatibility for the actual use environment is this deck's own write-back and unverified [claim 31: unverified].
What would settle it: A contract-manufacturer bill-of-materials and quote; an ISO 10993 testing plan; a human-factors study in a real veterinary-practice or home setting.
If it's false: Even a working sensor fails on the operational side — cost, shelf life, or usability outside a lab — which the viability deck's capital-to-first-dollar slide already flags as entirely unbacked [claim 29: unverified].
Would have to be true: The sensing chemistry, adhesive/insertion mechanism, and calibration algorithm would have to be buildable without infringing a live, enforceable patent — plausible on the base chemistry's age, but not established for adjacent process, membrane, or manufacturing patents.
Where it stands: Open, not cleared, for two independent reasons: no specific patent has been identified to search against, and the one connector probe attempted is blocked by a missing PATENTSVIEW_API_KEY, so it returns unconfirmed unconditionally regardless of input [claim 6: unconfirmed]. The base chemistry traces to 1960s research and is plausibly off-patent on age alone, but the Generator explicitly did not check whether Abbott, Dexcom, or Medtronic hold newer process, membrane-chemistry, or manufacturing-process patents [claim 5: unconfirmed].
What would settle it: Setting PATENTSVIEW_API_KEY and running a real patent search on glucose-oxidase sensing, membrane chemistry, adhesive/insertion mechanisms, and manufacturing process, held by Abbott, Dexcom, and Medtronic.
If it's false: FTO blocked is an automatic kill under the Stage 5 gate, regardless of every other score on this file — the one condition on any of these three decks with that property.
Every condition above with nothing verified behind it — the [no claim] markers, plus the unverified claims this deck itself put on record. Read this slide first.
The pattern is stark: the only verified claim in the file is regulatory scope [claim 1: verified] — that no FDA premarket pathway attaches to a veterinary-only device. Not one engineering condition specific to this device's own performance is backed by anything better than unconfirmed, and several are this deck's own unverified write-back.
If only one thing from this chair could be checked: freedom to operate on the sensing chemistry, the adhesive/insertion mechanism, and the calibration algorithm [claim 6: unconfirmed].
It is the one condition on any of these three decks with an automatic-kill property under the Stage 5 gate, and it is currently open for a tooling reason rather than a genuine unknown: the one connector call that could move it is blocked by a missing PATENTSVIEW_API_KEY, not by an unresolvable question. It is also, once the credential gap is closed, the cheapest item left on any of the three decks to actually settle.
Naming it is not a recommendation, a gate, or a kill.