idea-010 · 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: engineering. The people who have to make a single-use, three-channel electrochemical strip work: an ion-selective potassium channel, an enzymatic creatinine channel, and a hemoglobin channel whose only job is to suppress the potassium result on a hemolysed sample. They also have to produce the evidence a reviewer will ask for and manufacture the strip at volume.
How to read this: every slide is a condition, not a conclusion. A bracketed claim reference points into knowledge-base/candidates/idea-010.md. The number is the claim's position in its ## Claims list, and the tag is copied from that file as it stood on 2026-09-29 and never adjusted here. [no claim] means nothing in the file speaks to the condition. Claims 65–68 were appended by this deck's build. See docs/deck-spec.md.
What changed since the last build: claim 8verified moved to verified on 2026-09-29. FDA (US Food and Drug Administration) operates the Dual 510(k) (premarket notification) and CLIA (Clinical Laboratory Improvement Amendments) Waiver by Application route and has a final guidance on it. The same claim is explicit that this settles the route's existence, not the study package. None of the bench-physics claims this lens turns on moved, and no connector in this repo can reach them.
Would have to be true: From one fingerstick capillary drop, collected by trained clinic staff, the strip would have to report potassium and creatinine as plasma-equivalent values. It would have to do so within a bias and imprecision allowance fit for a RAASi dose-titration decision (the allowance is blank: no claim sets it). It would also have to report nothing for potassium when the sample is hemolysed. And the three channels would have to fill from that single drop without a second puncture.
Where it stands: The creatinine half has a verified format precedent. A cleared strip-and-meter measures creatinine at point of care in capillary whole blood and is "calibrated to provide plasma equivalent results to laboratory methods" claim 5verified. The potassium half has none in this format claim 7unconfirmed. Whether any FDA-recognised or published criterion defines the allowable potassium error for a titration decision is unchecked, so the core accuracy spec has no number claim 39unconfirmed. Whether one drop fills all three channels is unchecked. This build put it on record claim 65unverified.
What would settle it: A human reading the CLSI and FDA-recognised consensus standards for potassium analytical performance, together with published total-allowable-error criteria. None of these is reachable by a connector. Then a fill-volume study on a prototype strip measuring the drop volume clinic staff actually obtain.
If it's false: Without a defined allowable error there is no pass/fail line for Slide 5's bench study, and Factor 6 cannot be scored even after bench data exists. If three channels need more than one drop, the device needs a second stick, and the desirability deck's Slide 4 gets worse.
Would have to be true: Once hemolysed fingerstick samples are excluded, the remaining capillary potassium results would have to agree with venous potassium closely enough to support a titration decision. An ISE potassium channel calibrated to plasma-equivalent output would have to hold that calibration across the hematocrit range seen in HF and CKD (chronic kidney disease) patients.
Where it stands: The mechanism claim is what decides the candidate. Capillary sampling lyses red cells, hemolysis releases intracellular potassium, and the resulting false elevation is the file's named reason a strip-based fingerstick potassium meter does not already exist. It comes from the Generator's reasoning, with no source read and no connector able to reach it claim 22unconfirmed. Whether the published record shows that non-hemolysed capillary potassium tracks venous closely enough is unchecked. This build put it on record claim 67unverified. Hematocrit robustness of the plasma-equivalent calibration is bench physics with no connector claim 41unconfirmed. The regulatory text sets the matrix question: the potassium generic type names serum, plasma and urine, and not whole blood or capillary samples claim 1verified. The creatinine precedent shows that plasma-equivalent calibration is how a whole-blood device has sat inside a plasma-worded generic type claim 5verified.
What would settle it: First and cheapest: a human-read literature search for paired capillary-versus-venous potassium studies in adults, reporting bias with and without hemolysed samples. The Scores table notes this "may already be decisive". A literature scan returns pointers only, never a tag. Definitive: a paired capillary-versus-venous bench comparison across a hematocrit range.
If it's false: If non-hemolysed capillary potassium still disagrees with venous by more than a titration decision tolerates, suppressing hemolysed samples does not rescue the device. The mechanism fails, Factor 6 goes to the floor, and Slides 3–7 have nothing to build on.
Would have to be true: The engineering programme would have to produce the evidence for two generic types and two predicate arguments in one submission. It would also have to carry the CLIA-waiver studies with intended waived-site operators at multiple sites, and a review of the hemolysis suppression function as part of the device, not as an unreviewed internal safeguard.
Where it stands: The regulatory skeleton is verified. Potassium is 21 CFR (Title 21 of the Code of Federal Regulations) 862.1600, Class II (FDA device class II) claim 1verified. Creatinine is 21 CFR 862.1225, Class II, with an identification naming only plasma and urine claim 2verified. The product codes are CEM for potassium ISE claim 3verified and CGL for creatinine claim 4verified. The creatinine channel has a cleared predicate whose indication is broader than this device's claim 5verified. The Dual 510(k) and CLIA Waiver by Application route exists claim 8verified. The potassium channel has cleared whole-blood point-of-care predicate candidates (K200865 (an FDA 510(k) clearance number), K042270) but none in this form factor claim 7unconfirmed. A later De Novo (De Novo classification request) lookup found no potassium, creatinine or electrolyte entry in the cached listing, which the file records as context that moves no tag claim 26unconfirmed. What the waiver study package requires for this device is unchecked, and this build put it on record claim 66unverified. Whether FDA will review the suppression function inside the application is FDA review practice with no connector claim 43unconfirmed.
What would settle it: A human reading the named guidance, "Recommendations for Dual 510(k) and CLIA Waiver by Application Studies", to write the waiver study protocol. A pre-submission (Q-Sub) meeting with FDA on predicate adequacy for the potassium channel and on how the suppression function is reviewed. Whether K200865 or K042270 is an adequate predicate is FDA's substantial-equivalence judgement, not a lookup.
If it's false: If FDA does not accept a whole-blood point-of-care potassium predicate for a capillary strip, the potassium channel's route and calendar change. That lands in the viability deck's Slide 4 and in Factor 7. If the suppression function has to be validated as a claimed feature, the hemolysis detection threshold (Slide 4) becomes a reviewed performance claim with its own acceptance criteria.
Would have to be true: The on-strip hemoglobin channel would have to detect free hemoglobin in one fingerstick drop at the concentration where potassium becomes materially biased, and suppress the result before display. And the share of fingerstick attempts that end suppressed, by real clinic staff on real HF and CKD patients, would have to be low enough that the device stays usable. That threshold is blank. The file's only anchor is that one draw in four is not usable.
Where it stands: Detection at the relevant threshold is bench physics with no connector claim 40unconfirmed. The invalid-result rate is the file's own named real performance metric, measured on ordinary clinic samples and not only on clean ones. The Verifier recorded no view on its likely size claim 23unconfirmed. The two pull against each other. A lower detection threshold protects accuracy and raises the invalid rate, and a higher one does the reverse. No claim says where the usable window sits, or whether one exists [no claim].
What would settle it: A bench study spiking hemolysate into whole blood to set the detection threshold against a defined potassium bias. The bias limit comes from Slide 1's blank. Then the multi-operator field study in Slide 5, with the suppressed-result rate as a primary endpoint.
If it's false: If there is no threshold that keeps both potassium bias and the invalid rate acceptable, the device either reports wrong potassium or reports too few results to use. The first is a clinical risk. The second destroys the desirability deck's Slide 4 and turns every suppressed strip into an unbilled cost in the viability deck's Slide 7 claim 63unverified. This is the load-bearing condition for this lens (Slide 9).
Would have to be true: Before any patient-facing study, a bench and method-comparison programme would have to show four things: (a) paired capillary-versus-venous agreement within the allowable error, (b) reliable detection under a deliberate lysis challenge, (c) calibration holding across hematocrit, and (d) a suppressed-result rate across multiple ordinary operators that fits the usable window from Slide 4.
Where it stands: None of it exists. The file names this study as the only thing that settles the mechanism: paired capillary versus venous, across operators, with a deliberate lysis challenge claim 22unconfirmed. The file also names the endpoint it must report claim 23unconfirmed. The detection, hematocrit and allowable-error legs are open as well claim 40unconfirmed, claim 41unconfirmed, claim 39unconfirmed.
What would settle it: This is Stage 7 bench work that no agent here can do. The file's Diligence Package slot, "minimum viable bench-evidence before any animal study", is where a human Diligence Lead would draft it. A human-read literature review (Slide 2) comes first and may narrow what the bench needs to show.
If it's false: If the bench cannot show agreement, detection and an acceptable invalid rate together, nothing downstream starts: no waiver study, no clinic pilot, no manufacturing scale-up. Tier 3 sits immediately after this slide, and nothing here proposes whether to run any of it.
Would have to be true: The three-channel strip would have to hold its calibration and its hemolysis-detection threshold across a commercially viable shelf life in ordinary clinic storage. It would have to be manufacturable at a yield that supports a strip price below the clinic's payment. And medical assistants and nurses in ordinary HF and nephrology clinics would have to get an adequate drop without milking the finger, from patients with edema, cool extremities or frequently lanced fingers.
Where it stands: Shelf life is unchecked bench stability, with no connector claim 42unconfirmed. The strip manufacturing yield curve appears in the file only as a claimed moat asset claim 24unconfirmed. The bill-of-materials margin it has to support is a business estimate carried in the viability deck claim 46unconfirmed. Operator sampling adequacy is a usability question with no connector claim 32unconfirmed. No claim says how much finger-milking contributes to the capillary hemolysis in the mechanism claim. If it does contribute, this condition feeds Slide 4's invalid rate directly [no claim].
What would settle it: Real-time and accelerated stability studies on prototype strips. Pilot-line yield data. A sample-collection usability study with clinic staff on the target population, reporting how often a drop was obtained without milking.
If it's false: If the strip drifts in storage, the detection threshold drifts with it and the Slide 4 window moves after release. If staff routinely milk fingers, the invalid rate rises regardless of the chemistry. Either one lands in Factor 6 and, through cost per reported result, in Factor 7.
Would have to be true: No in-force US patent would have to claim detecting sample hemolysis on a test strip or meter and suppressing a co-located potassium result. If one does, a workaround would have to exist that keeps suppression as the design, not a compromise of it. The potassium-selective membrane chemistry at whole-blood ionic strength would also have to be practisable, in dense ISE and glucose-strip art held by the same firms that hold current CEM and CGL clearances.
Where it stands: No patent search has been performed and no number is named. The field is unexamined, not clear, and PATENTSVIEW_API_KEY is unset claim 25unconfirmed. The suppression logic and membrane chemistry are listed as moat assets on the Generator's judgement only claim 24unconfirmed. Whether the suppression function is free of in-force claims is unchecked. This build put it on record claim 68unverified.
What would settle it: The file's own highest-value query: art on hemoglobin or hemolysis detection used to flag or suppress a co-located electrolyte result, searched by CPC class (G01N27/333, G01N33/49, G01N33/72, G01N27/327, C12Q1/00) and by assignee (Nova Biomedical, Abbott Point of Care, Roche, Siemens Healthineers, i-SENS) in Patents Public Search or Espacenet. Then connectors patent on the numbers found, once the key is set. Freedom to operate is only settled by a Tier 3 legal opinion at Stage 7.
If it's false: If suppression-on-hemolysis is claimed and cannot be worked around, the device's one distinguishing feature is off the table, and a blocked FTO (freedom to operate) is the rubric's only automatic kill. The design would fall back to correcting a hemolysed result instead of suppressing it. The file's mechanism is written explicitly to suppress, not correct.
Every condition above that has nothing verified behind its load-bearing half. Slides 1–3 cite verified regulatory claims (1–5, 8). In each case those are the known skeleton, and the unresolved half is what the condition turns on. So all seven are listed.
Six verified claims anchor this deck, the most of any lens. All six are regulatory-database facts about classification and a creatinine predicate. None speaks to whether the potassium channel works.
If only one thing from this chair could be checked: what share of fingerstick attempts, by ordinary clinic staff on real HF and CKD patients, end with the potassium result suppressed rather than reported, at a detection threshold that still keeps reported results accurate enough to titrate on claim 23unconfirmed, claim 40unconfirmed?
This one rather than the others because it is where the mechanism (Slide 2), the design choice to suppress rather than correct (Slide 4), and real-world sample collection (Slide 6) meet in one number. It also feeds the other two lenses directly: the suppressed-result burden on the desirability deck's Slide 4, and the unbillable-strip cost on the viability deck's Slide 7. It cannot be answered by any connector or database, only by bench and multi-operator data. The cheapest first look is a human reading the published capillary-versus-venous potassium literature claim 67unverified.
Naming it is not a recommendation, a gate, or a kill.
_shared/plain-language.md