Balance Meter — Invention Brief for Patent Discussion¶
Confidential — Attorney Work Product / Pre-Filing Strategy · Do not distribute Date: 2026-06-02 · Prepared by: Gene Lang, PharmD — Director of Pharmacy Operations Named inventors (CONFIRM with counsel): Gene Lang; Matilda [_]; others as contributed Filing status (CONFIRM): ☐ no filing yet ☐ provisional drafted ☐ provisional filed — date _
Purpose: give counsel a clean, plain-language description of what the Balance Meter is, what is genuinely novel about it, and where it sits relative to prior art — so we can decide what to claim, what to hold as trade secret, and what to file before any further disclosure.
1. The problem¶
Falls are the leading cause of injury in older adults, and clinicians have validated static screens (e.g., CDC STEADI) and known fall-risk-increasing-drug (FRID) lists for decades. But the existing technology splits into three lanes, and none of them delivers a personalized, continuously-calibrated, wellness-safe signal of change from a person's own normal:
- Consumer wearables (Apple Watch, Fitbit) — one-size-fits-all fall detection after a fall, plus generic activity. No personalization to individual fall-risk context; no pre-fall change signal.
- PERS / medical-alert devices — emergency response after an event. No prediction, no baseline.
- Clinical fall-risk assessment (STEADI, Tinetti, TUG) — accurate but static, professional-administered, and regulated as clinical decision-support when consumerized (see SeniorLife warning letter, §5). Putting a "fall-risk score" in front of a consumer crosses into medical-device territory.
The unmet need: a device a senior wears every day that learns their normal, reads change from it, and tells them something useful without making a clinical claim, naming a disease, attributing a symptom to a drug, or implying it is a fall predictor.
2. The invention, in one paragraph¶
A wearable "balance meter": a personal-baseline movement barometer whose sensitivity is dynamically calibrated by (a) a provenance-weighted clinical risk index built at intake and (b) medication-change events — presenting the wearer a non-clinical wellness signal (a calm band state plus a gamified steadiness score) while reserving the underlying clinical index for a licensed professional (the pharmacist). The same measured gait deviation produces a different member-facing state depending on the wearer's calibrated risk — but the risk itself is never shown to the wearer.
3. The candidate novel elements¶
Each is described as: what it is · why it is non-obvious · prior-art contrast.
3.1 Dynamic sensitivity calibration of a personal-baseline gait barometer¶
What: The live signal is gait/movement deviation from the wearer's own rolling baseline (cadence, gait variability, speed proxy). A clinical risk index does not add to the member-facing reading — it sets the threshold and window at which a deviation moves the band. Higher calibrated risk → smaller deviation / fewer sustained days trips a state change; lower risk → more tolerance. Non-obvious: The risk model modulates sensitivity, not output — the same 15% cadence drop is "Some change" for one wearer and "Steady" for another, by design, and the wearer is never told why. Prior art contrast: Apple/Fitbit apply uniform thresholds to all users; STEADI produces a fixed score independent of continuous sensor data. We found no art that uses a clinical risk index to tune the sensitivity of a personal-baseline movement signal while hiding the index from the user.
3.2 Medication change as a transient sensitivity event¶
What: Adding/changing a medication does not add static "points." It temporarily tightens sensitivity (≈2 weeks, compared against a shorter "recent normal" window) so post-change drift surfaces sooner — then relaxes. Non-obvious: Treats a medication event as a time-limited modifier of a detection threshold, not a risk score; nothing is attributed to the wearer (no "your new drug caused this"). Prior art contrast: FRID lists flag drugs statically; no art (to our knowledge) converts a medication event into a transient sensitivity window on a personal movement baseline.
3.3 Provenance-weighted intake feeding the calibration¶
What: Every intake input carries a provenance tag and a weight: self-reported ×0.5 · proxy/caregiver ×0.6 · claims/Surescripts-confirmed ×0.85 · pharmacist-verified ×1.0. Lower- confidence inputs move the calibration less. Non-obvious: Encodes measurement reliability (self-reported conditions carry up to ~50% error) directly into how much an input is allowed to influence the gauge — so unreliable self-report can never dominate. Prior art contrast: Standard risk tools treat all inputs as equally true. Provenance- weighting of risk inputs by source-reliability, feeding a sensitivity calibration, is the novel combination.
3.4 Wellness / clinical bifurcation architecture¶
What: One engine, two surfaces. The wearer sees a warm band state (no number, no attribution) plus a gamified steadiness/engagement score (wellness, not a fall-risk number). The professional sees the full clinical index, % deviation, provenance, and likely contributors. Non-obvious: A single computation deliberately split so the consumer surface stays inside FDA general-wellness enforcement discretion while a licensed professional retains the clinical view — the architecture is the regulatory boundary. Prior art contrast: SeniorLife exposed fall-risk identification to consumers and drew an FDA warning letter. Our bifurcation is the structural answer to exactly that failure mode.
3.5 Quarterly polypharmacy reconciliation via pharmacy-network cross-check¶
What: A periodic (≈quarterly) check-in reconciles the wearer's self-entered list against pharmacy-network/claims data (Surescripts, Medicare Part D) to surface medications filled elsewhere and repair an incomplete list — then re-calibrates. Non-obvious: Targets a documented gap — seniors using multiple pharmacies under-report their own lists — and closes it on a cadence chosen to be inoffensive (meds, not "are-you-afraid-of-falling") and low-burden on the pharmacist. Prior-art contrast / evidence: ~38% of Medicare Part D seniors use multiple pharmacies (national sample, n≈927k); concurrent multi-pharmacy users average ~11 unique medications vs ~8 for single-pharmacy users — i.e., the self-entered list is systematically short, and a network cross-check is what repairs it.
3.6 (Method) User-initiated secondary reminder — wellness-safe adherence¶
What: Instead of the system designating drugs "time-critical," the wearer may opt a given medication into a second reminder ("would you like a second reminder so you don't miss it?"). User-configured; the device never classifies clinical criticality or assumes the duty. Non-obvious / why it matters: Keeps adherence support inside the wellness lane and avoids liability for system-assigned criticality, while still serving high-stakes meds (e.g., insulin) for the user who chooses it.
4. Prior art and how we are distinct¶
| Prior art | What it does | How Balance Meter differs |
|---|---|---|
| Apple Watch / Fitbit fall detection | Uniform post-fall detection + generic activity | Pre-fall change signal, personalized + risk-calibrated, wellness-framed |
| PERS / medical alert | Emergency response after an event | Baseline-relative monitoring, no clinical claim |
| CDC STEADI / Tinetti / TUG | Static, professional clinical screen / score | We derive plain-language items but never reproduce a consumer-facing clinical score; STEADI becomes calibration, not output |
| SeniorLife Technologies | Consumer "fall-risk identification" → FDA warning letter | Our wellness/clinical bifurcation is the structural fix for that exact violation |
| Prior-company medication-entry build (Patrick Ford, in Confluence) | Earlier intake/med-entry UI | Internal prior art for the intake UI component — distinct from the algorithm claims here; flag for FTO review (§6) |
5. Wellness-lane positioning is a feature of the IP, not a limitation¶
The constraints Matilda has held us to — no disease naming, no symptom-to-drug attribution, no severity language, no "fall-risk score" to the wearer, no implication the band is infallible — are not just compliance. They define the boundary the claims live within. The novelty is precisely how much useful, personalized signal we deliver while staying on the wellness side of that line. The boundary is the moat.
6. Open items for counsel¶
- Inventorship — confirm named inventors and contribution records (Gene; Matilda; tech contributors to the live-gait method).
- Filing status & timing — provisional vs non-provisional; file before any further external disclosure (the algorithm package was shown to the clinical reviewer and uploaded to an internal shared drive — confirm that does not constitute a public disclosure).
- Freedom to operate — the prior-company med-entry build (Patrick Ford / Confluence): ownership, assignment, and whether it reads on §3 claims or only the intake UI.
- Claim vs trade secret — recommend claiming the architecture (§3.1–3.5) and holding the exact thresholds, weights, and calibration constants as trade secret (they require the pilot to set and are easier to protect unpublished).
- Publication timing — coordinate any validation-study publication with filing.
- CPT / acquirer pathway — the pharmacist clinical index is the bridge to a future CPT-reimbursable clinical product an acquirer could pursue through FDA; we retain the wellness IP. Confirm this framing supports valuation without creating a device claim now.
7. One-line summary for the room¶
The Balance Meter is a personal-baseline movement barometer whose sensitivity is dynamically calibrated by a provenance-weighted clinical risk index and medication events — delivering a personalized, wellness-safe signal of change to the wearer while reserving the clinical view for a pharmacist. The wellness/clinical split is both the safety boundary and the novelty.