Smart Scheduling — Claim Architecture (inventor draft)¶
Date: 2026-06-04 · Owner: Gene (Pharmacy Ops) · For: patent counsel, as a starting artifact
⚖ THIS IS NOT LEGAL CLAIM LANGUAGE. This is the inventor's plain-language articulation of the system and its pieces — a structured starting point for counsel to draft formal claims. It deliberately avoids legal terms of art. Counsel owns claim scope, language, and the FTO determination. Every element below is cross-referenced to the prior-art read in
KB/03_Precedents_Tools_FTO.
The strategy in one paragraph¶
The prior-art corridor for "predict a miss and personalize the reminder" is crowded (Alarm.com US9070267B2, 2011, is the closest reference). So the independent claim leads on the integrated system — a wellness-lane scheduler whose reminder behavior is governed by a pharmacist-curated, drug-class PK/PD tolerance model, that carries dose-time context, that is provenance-tiered, that de-escalates as a routine proves out, and that shares a medication-change event with a fall-risk engine. Prediction and personalized timing appear only as dependent claims (implementations), never as the independent invention.
Independent Claim 1 — the system (plain language)¶
A medication-reminder system for an individual, comprising:
- (a) a drug knowledge store that holds, per drug class, a tolerance window derived from both pharmacokinetic (half-life) and pharmacodynamic (duration-of-effect) properties — i.e., how late a dose may be before lateness clinically matters — together with dose-time administration context (e.g., with-food / empty-stomach / separate-from-X);
- (b) a scheduler that generates reminders whose timing and escalation behavior are governed by that tolerance window — such that a long-tolerance medication is reminded as a multi-day window and a short-tolerance medication is reminded precisely — rather than by a fixed clock;
- (c) a confirmation capture that records whether the individual confirmed the dose, and that reduces or removes subsequent reminders for that dose once confirmed, and reduces reminder intensity over time as a routine is established (de-escalation);
- (d) a provenance tier attached to the regimen (self-entered / imported / clinician-or-pharmacist-verified) that sets the precision/confidence of the generated reminders;
- (e) an event interface that, upon a change to the individual's medication set affecting fall risk, emits a single medication-change event consumed by both the reminder scheduler (to open a heightened-attention period) and a separate fall-risk/balance engine (to open a sensitivity watch-window); and
- (f) a constraint that the individual-facing output is limited to reminders and confirmations and does not output dosing instructions, titration, or skip/double-dose guidance — clinically-sensitive conditions are routed to a human (pharmacist/caregiver).
Why it clears the closest art: Alarm.com (US9070267B2) covers (b)-adjacent withholding/adjusting within "therapeutic windows" and (e)-adjacent caregiver escalation — but not a tolerance window derived from curated PK+PD per drug class governing reminder behavior, not the shared-event coupling to a fall-risk engine, and not provenance-tiered precision. The novelty is the combination of (a)+(b)+(d)+(e) operating together.
Independent Claim 2 — the method (plain language)¶
A method of reminding an individual to take medication, comprising:
- retrieving, for a medication, a PK/PD-derived tolerance window and dose-time context from a curated store;
- generating a reminder whose timing and escalation reflect that window (wide window → day-ranged, gentle; narrow window → precise, escalating);
- capturing a confirmation and suppressing further reminders for that dose on confirmation;
- lowering reminder intensity over successive periods as confirmations establish a routine, and raising it only on a detected lapse;
- on a fall-risk-relevant medication change, emitting one event that simultaneously adjusts the reminder schedule and a separate fall-risk engine's sensitivity; and
- withholding any dosing/titration instruction from the individual, routing clinically-sensitive conditions to a human.
Dependent claims — strength-ranked (each adds to Claim 1 or 2)¶
| # | Adds… | Prior-art status | Role |
|---|---|---|---|
| D1 | the curated tolerance is pharmacist-authored/signed-off and each reminder decision is traceable to a versioned knowledge-base entry (auditable) | open — un-arted as a product feature | strong — anchors the "clinically-curated, not naive" distinction |
| D2 | the shared medication-change event specifically opens a fall-engine sensitivity window and a scheduling-attention window of bounded duration | open — no art pairs scheduling with a fall engine | strongest single piece |
| D3 | provenance tiers map to discrete reminder-precision levels; verified provenance unlocks the most precise tier | open / thin | strong — portfolio-coherent |
| D4 | the reminder carries the dose-time administration context to the individual at dose time (with-food, etc.) sourced from structured labeling | thin / white space (Springer 2020) | strong |
| D5 | de-escalation: reminder frequency/intensity decreases monotonically as a per-dose confirmation streak grows, distinct from escalation ladders | thin — everyone escalates | strong (contrarian) |
| D6 | classification of a medication as high-stakes/PRN changes behavior (explicit-confirm + human-routing for high-stakes; no scheduled reminder for PRN) | partly arted | medium |
| D7 | pre-emptive miss prediction from behavioral signals (snooze/late-confirm) triggers an earlier, gentler nudge | crowded (Alarm.com, RL/JITAI) | defensive only — implementation, not independent |
| D8 | learned personal timing anchored to the individual's confirmed-dose rhythm (meals/routine) | crowded (context-aware reminders) | defensive only |
| D9 | caregiver/Circle notification on lapse, opt-in | commodity | defensive only |
| D10 | on-device generation of the reminder with offline operation; learning data de-identified before leaving the device boundary | partly arted; privacy-framing useful | medium |
Drafting notes for counsel¶
- Keep D7/D8/D9 dependent. If any drifts toward the independent claim, FTO collides with Alarm.com + Medisafe JITI + the RL/JITAI corpus + the cloud-ANN patent (US 12040066). Search those head-on.
- The independent claim's center of gravity is (a)+(b)+(d)+(e). That four-way combination is what we believe is open. D1/D2/D3/D4/D5 are the strongest dependents to fall back to if the independent claim is narrowed.
- The † wellness/CDS line runs through (a)/(b): a PK/PD-curated tolerance window that governs reminders should read as scheduling convenience, not clinical decision support. The Jan-2026 FDA guidance widened this lane ("positioning, not capability, decides classification" —
KB/05); the member-facing limitation (f)/step 6 is the design guardrail that keeps it wellness. - Portfolio question: the shared medication-change event (D2) and the provenance model (D3) are common to the Balance Meter filing — argues for a coordinated filing or a shared specification.
- Enablement is supported: "drug forgiveness" is published science (Assawasuwannakit 2015, PMC4394614; AARDEX;
KB/02), so the tolerance-window concept is enabled; the novelty is its use as curated reminder behavior, not the underlying pharmacology. - Open dependency: D1's curated tolerance table (
KB/02, 27 classes) is preliminary — values are illustrative until a clinical review finalizes them.
Cross-references: deck Smart_Scheduling_Deck_2026-06-04 (S13 claim spine, S14 prior art) · Smart_Scheduling_Invention_Brief_2026-06-04 · KB/03_Precedents_Tools_FTO (the full FTO read + patent numbers) · KB/02_Drug_Forgiveness_Tolerance_Data.