User Testing Synthesis — Pre-Pilot Field Findings¶

Prepared for: Elliot — review with Peyman Author: Gene Lang, PharmD — Director of Pharmacy Operations Date: 2026-04-23 Dataset: 11 participants across two days at the anchor clinical site · 4/20 (n=8) + 4/21 (n=3)

What we did¶

Over two days we ran pre-screener field conversations with 11 older adults and caregivers. This document is the consolidated read.

The big four¶

1. The concept resonates universally — friction is everywhere else¶

Zero of 11 participants rejected the product outright. Two said "not there yet" (future pipeline, not rejection). The rest ranged from moderate future interest to 5/5 enthusiasm. The pitch (wristband + app + caregiver support circle) landed without confusion in every conversation.

2. The 2-week gap story reframes caregiver-visibility from convenience to safety¶

One elderly participant reported going two full weeks without taking any of her medications while her adult daughter was out of town. She's a watch-enthusiast and loves the concept.

This is the single strongest adherence-risk signal in the dataset. It turns caregiver-visibility from a nice-to-have convenience into a clinical-safety feature. Any caregiver-dependent design must function during caregiver absence — travel, illness, hospitalization — not as an edge case but as an explicit product state.

3. Reminder-alone is not enough — fatigue beats forgetting for a real segment¶

A caregiver for her mother described the existing reminder infrastructure: four alarms a day, all hearing-audible, all working as designed. The mother still misses doses. From the field notes: "she forgets even with the alarm system" — and the reason isn't memory. The caregiver described her mother as "tired or too fatigued" to act on the alarms.

From a different angle the following day: a dementia patient has two phone alarms. When they fire, she calls her daughter to ask what they're for. Every alarm produces a caregiver interrupt, zero medications taken. The daughter absorbs the load.

These are not reminder-system failures. They are cognition-and-action failures that a reminder system cannot fix alone. The minimum viable adherence product for these cohorts is a closed loop: alarm → identify medication → instruct patient → confirm dose taken → escalate to caregiver if unconfirmed. Not a better alarm. A complete encounter.

4. $35/month is the emerging price anchor¶

Two 4/21 participants independently cited $35/month across different motivating features: the dementia-track caregiver for GPS + care loop, and the self-managing patient conditional on blood glucose and blood pressure monitoring. The 4/20 pharmacist-literate participant anchored at $30–35 for the device plus ~$5/month subscription — different pricing structure but the same order of magnitude.

Two participants raised Medicare coverage unprompted as the dominant pricing mental model. The pricing conversation is already a payer conversation for this demographic. GTM should address the payer pathway, not lead with out-of-pocket.

What surprised us¶

These ran counter to the assumptions we went in with:

Patients are already running their own adherence systems. Five of 11 had self-built systems before we pitched anything: one pharmacist-literate patient used a label-maker on his pill-box compartments, one caregiver stacked four alarms a day, another patient poured tomorrow's doses into a small cup every night, another lined bottles up on the counter with no pill box at all, and one 4/21 encounter used a tray to pre-stage doses. The existing system is the product landscape. The onboarding conversation should open with "walk me through how you manage your medications today" rather than pitching features against a default of "you probably forget."
The dementia archetype is not a wellness use case — it's a care-coordination use case. The daughter of a dementia patient asked for GPS/location first, fall alerting second, and medication guidance (with audio + visual instruction at dose time) third. Adherence reminders were almost secondary. The family had already placed an Apple AirTag in the patient's cane for location tracking — a DIY workaround showing pull for the feature is already in the field. For this cohort, the product is a care-coordination device.
Caregiver = buyer for the high-acuity cohort. The patient is not the decision-maker. The adult child is the buyer, the evaluator, and the operator. Onboarding, pricing, and distribution conversations need to address the caregiver directly.
Clinical authority is the adoption unlock — both ways. Pharmacist credibility surfaced 4/20: "If pharmacist tells patient don't use the boxes, keep it in bottle, they listen better than the doctor." Prescriber-as-distribution-channel surfaced 4/21: a self-managing patient asked whether her prescriber could provide or recommend the device before she'd acquire it. Same underlying dynamic from two angles — this demographic does not adopt direct-to-consumer. Acquisition runs through a clinical trust layer.
Older-adult tech literacy is higher than expected. The 80+ male is an iPhone + Apple Watch user with strong opinions across the full feature set. The dementia family uses a cane-mounted Apple AirTag. We are not landing in a low-baseline-tech market.

Two conversion pathways — neither adherence-driven¶

From 11 conversations, two distinct pathways have emerged:

Caregiver track — converts on GPS + fall alerting + closed-loop medication guidance. Buyer is the adult-child caregiver. High acuity (dementia, elopement, cognitive impairment). Price tolerance demonstrated at $35/month. 30-minute-daily time burden is the anchor for the caregiver value proposition.

Self-managing biometric track — converts on blood pressure and blood glucose monitoring. Patient is self-adherent and has no reminder problem — no biometrics means no reason to wear the device. Distribution preference runs through the prescriber, not DTC. Price tolerance $35/month conditional on biometrics.

Our current adherence-reminder core is necessary floor but not sufficient ceiling for either cohort.

What's next¶

Day 2+ anchor-site data continues this week; target n=15–20 by Friday. Interview methodology is shifting to hypothesis-driven + measurable metrics per the 4/21 product team direction. IRB exemption submission in progress (10-minute format, end-of-week target). Findings 2 and 3 above have direct implications for TestFlight v1 on May 1.

Supporting detail — signals by category¶

Lighter version of the 8-category synthesis underneath the executive brief. Source documents have the full per-participant detail.

1. Adherence & behavioral patterns¶

Initial-denial pattern. First answer is often "I don't miss it"; followup probing reveals soft non-adherence. Timing drift (8 AM dose at 10 AM) shows up when a binary "did you take it" gate misses.
Skip-don't-double-dose is the patient intuition. "If forget skips it till next day." App design should not prompt make-up doses.
Self-perceived adherent ≠ actually adherent. Three patients reported no need for reminders. Anchor-site aggregate data shows 22% late refills despite most patients self-identifying as adherent.
Fatigue as an independent root cause. Alarm fires, patient hears it, is too tired to act. Reminder-only insufficient.
Cognition-to-action gap in dementia. Alarm fires, patient doesn't recognize what it's for, calls caregiver. The reminder has become a caregiver-interruption system.

Product implications: Avoid defensive-denial triggers in reminder language ("staying on track" works, "are you missing doses?" does not). Build a confirmation loop for fatigue-driven misses. For dementia, build a full encounter — alarm + medication image + spoken instruction + confirmation + caregiver escalation.

2. Medication management friction¶

Bottles > pill boxes is a pattern. Two participants explicitly prefer bottles; one told a pill-box failure story — it opened during travel and pills mixed up: "Fortunately he knew which is which but ordinary guy doesn't know."
Patient-engineered workarounds. Nightly bottle-to-cup transfer, hand-labeled pill-box compartments, lined-up bottles, tray pre-staging, 4-alarm stacks.
Generic-name pronunciation friction. Patients can't pronounce generic names, "sometimes request wrong medication." Brand names in the app UI.
NDC change risk. White pill becomes blue pill between refills. Flag appearance changes proactively.
Dementia-specific spec. Reminder notification must include medication image + spoken instruction at dose time. Silent vibration or a generic beep is net-negative — generates caregiver calls without producing adherence.

3. Reminders & memory systems¶

Location-based cueing works. Bathroom sink cabinet, first thing seen in the morning. Habit-stacked.
Family cueing is already happening. Adult children verbally reminding parents — formalized app version viewed positively.
The 4-alarm baseline is the bar to beat. "Alarms are good but could be better." Incumbent reminder infrastructure is already in the home.
Modality preference is bimodal. Vibration (privacy, no public embarrassment) vs. voice (companionship-feel). Must be configurable; do not default-ship one.
Advance-notice window preferred, not just at-dose-time.
"Not there yet" is not rejection. Capture as future-pipeline, track over time.

4. Refill & sync experience¶

Refill anticipation is a high-demand feature for pharmacy-literate users — 2–5 days before monthly refill due, 2 weeks for 90-day fills.
Most patients don't raise refill unprompted. Formal screener should ask directly; otherwise the signal is lost.
Refill cadence aligns with anchor-site supply distribution (roughly 60% 30-day, 18% 90-day).

5. Health context & living situation¶

Adult-child caregivers are the primary informal support layer. Four of 11 memos feature an adult child as the active caregiving agent.
Quantified caregiver burden: 30 minutes/day managing meds for a parent.
5-layer alert-circle aspiration: self → sibling → doctor → pharmacist → hospital. Broader than current spec probably assumes.
Caregiver-absent fragility is a real failure mode. Travel, illness, hospitalization — the product must function during these gaps.
Existing consumer wearables already deployed. Apple AirTag in a cane for location tracking. Mother wearing a fall necklace. We are replacing or integrating with incumbent solutions, not landing in a blank field.
Dual-role archetype (self-managing patient who also caregives) is a distinct persona.

6. System failures & transition points¶

NDC / manufacturer changes cause dangerous confusion between refills.
Pharmacist > doctor trust for this demographic. Onboarding voice should be pharmacist-led, not clinician-led.
Travel and caregiver-away are both explicit failure points. Not edge cases — product states.
Prescriber as gatekeeper to adoption. Self-managing patients will not acquire independently without prescriber endorsement or distribution.
Phone alarms as active liability in dementia. Every alarm generates a caregiver call. The status quo is a daily failure mode.

7. Adoption readiness & motivation¶

Concept never rejected — 0 of 11. Friction is timing, pricing, feature specifics.
Medicare coverage is the dominant pricing mental model. Two independent unsolicited mentions. Pricing conversation is already a payer conversation.
Caregiver-driven adoption for high-acuity. Adult-child daughter is "very interested" — at $35/month with GPS, fall alert, and medication guidance, not price-sensitive.
Biometrics as activation feature for self-managing archetype. Without BP + BG, no reason to wear it.
Segmented pipeline: active system-builders (high engagement now) vs. future-need (track over time).

8. Technology & product fit¶

Older-adult tech literacy is higher than assumed. iPhone + Apple Watch ownership at 80+.
Criticality-tiered UI ask ("red star" for high-priority meds) maps to the existing Medication Risk Database tiers.
In-context medication guidance is an unsolicited ask — "take with food or not, basic consultation tips." Validates an AI medication-info layer.
Vital-signs monitoring asks — blood pressure, temperature, balance. Scope question: in or out?
Hardware asks: waterproof (surfaced unprompted), privacy-by-default on voice announcements.
$35/month price anchor holds across days and archetypes. Medicare-coverage expectation recurring. Blue Button 2.0 deductible integration surfaced as a user ask.
GPS is the #1 feature for caregiver-track. Existing Apple-AirTag-in-cane workaround confirms pull.
Audio + visual medication instruction is a spec requirement for dementia-track users.
Biometric monitoring (BP + BG) is the sole conversion feature for the self-managing archetype.
Prescriber-channel distribution is an explicit patient ask.

Appendix — Full signal detail per category¶

The executive summary above distills the themes. This appendix preserves signal-level detail with verbatim quotes and per-participant attribution for readers who want to trace findings back to source. Tags: date + participant ID. Source files referenced at the end.

Participant reference¶

Tag	Date	Age	Gender	Track	Interest	Key signal
P-NOTES-01	4/20	80+	M	Dual (A + B)	5 / 5	Pharmacist-literate; strong opinions across the full feature set
P-VOICE-01 (4/20)	4/20	~80s	F	A	High (4–5)	Missed-dose admission after initial denial
P-VOICE-02 (4/20)	4/20	—	F (est.)	A	Moderate (future)	"Not there yet" on doctor-appointments feature
P-VOICE-03 (4/20)	4/20	73	F	A	Future (2–3)	Self-assessed adherent; future interest
P-NOTES-02	4/20	—	F	B (caregiver)	4 / 5	4-alarm system; fatigue > forgetting; 30 min/day burden; 5-layer alert circle
P-NOTES-03	4/20	Elderly	F	A	High (4–5)	2-week non-adherence when daughter was traveling
P-NOTES-04	4/20	Elderly	F	A	Low-now / future	Self-built pill-cup nightly system
P-NOTES-05	4/20	—	M	B	—	Vitals ask (BP, temp, balance); waterproof; no pill box
P-VOICE-01 (4/21)	4/21	Elderly	F	B (daughter reporting)	5 / 5	Dementia + elopement; GPS is #1 ask; phone alarm system non-functional
P-VOICE-02 (4/21)	4/21	—	—	A	Insufficient	Tray-organization system; brief encounter
P-VOICE-03 (4/21)	4/21	Elderly	F	A	Conditional 4 / 5	Self-adherent; biometric monitoring (BG + BP) as conversion

1. Adherence & behavioral patterns — full signals¶

Initial-denial pattern [P-VOICE-01, 4/20] — First said "I don't miss it" then admitted "oh last week I missed the whole thing, I didn't take it at all in the morning." Behavioral signal: 5/5, surprise flag. Interviewers must probe past first-answer denial.
Timing drift [P-VOICE-01, 4/20] — Patient takes 8 AM dose at 10 AM when she forgets. Soft non-adherence that a binary "did you take it?" gate misses.
Catastrophic non-adherence when caregiver is unavailable [P-NOTES-03, 4/20] — Patient missed 2 full weeks of medications when her daughter was out of town. Highest single adherence-risk signal across the 11-memo dataset.
Skip-not-double-dose rule [P-NOTES-01, 4/20] — "If forget skips it till next day." Validates app design: do not prompt make-up doses.
Self-perceived adherent ≠ actually adherent [P-VOICE-03 4/20, P-NOTES-04 4/20, P-VOICE-03 4/21] — Three patients report no need for reminders. Pattern matches anchor-site aggregate data (22% late refills despite most patients self-identifying as adherent).
Fatigue > forgetting as the cause of misses [P-NOTES-02, 4/20] — "Her mom misses taking meds because she's tired or too fatigued" and "she forgets even with the alarm system." The alarm fires, she hears it, she is simply too fatigued to act.
Complete alarm-to-action failure [P-VOICE-01, 4/21] — Dementia patient has two phone alarms (morning and evening) but when they fire she doesn't recognize what they're for. She calls her daughter to ask. The daughter then has to intervene. This is not forgetting — it's a semantic breakdown between the cue and the required action.
Tray pre-staging behavior [P-VOICE-02, 4/21] — Patient takes medications out of their containers and places them on a tray before taking them. Self-cueing without technology.
Self-adherent, high confidence [P-VOICE-03, 4/21] — Patient reports never forgetting to take her medication. Habitual integration.

2. Medication management friction — full signals¶

Bottle vs. pill box philosophy [P-NOTES-01, P-NOTES-05, 4/20] — Two participants strongly prefer bottles over pill boxes. P-NOTES-05 explicitly: bottles lined up, no pill box. P-NOTES-01's rationale: readable instructions, prevents mixups.
Pill box failure anecdote [P-NOTES-01, 4/20, verbatim]: "One time he had the box with seven meds, he put it inside, and when he went to trip it opened and it all mixed up. Fortunately he knew which is which but ordinary guy doesn't know."
Hybrid bottle-to-cup workaround [P-NOTES-04, 4/20] — Patient pours next day's doses from bottles into a small cup each night. Self-engineered "soft pill box."
Generic-name pronunciation friction [P-NOTES-01, 4/20] — Patients can't pronounce generic names → "sometimes request wrong medication." Product ask: brand names in the app UI.
Color/shape/appearance change risk [P-NOTES-01, 4/20] — NDC changes between refills create confusion.
Label-hack workaround [P-NOTES-01, 4/20] — Participant labeled pill-box compartments himself.
Instruction gap at point of action [P-VOICE-01, 4/21] — Patient needs to be shown a picture of the medication AND told what it is and what to do at the time of the alarm. Verbatim: "instructions talking to her to say what to do and what to take." Without that, the alarm produces a phone call to the daughter rather than a medication taken.
Tray system removes friction [P-VOICE-02, 4/21] — Pre-staging medications on a tray handles the "which one, when" question before the dose time.

3. Reminders & memory systems — full signals¶

Location-based cueing [P-NOTES-01, 4/20] — Bathroom sink cabinet, first thing seen in morning. Habit-stacked to wake routine.
Family cueing (informal) [P-VOICE-01 4/20, P-NOTES-03 4/20] — Adult children verbally reminding parents. Already happening; formalized app version viewed positively.
Multi-alarm baseline [P-NOTES-02, 4/20] — 4 alarms/day is the existing adherence infrastructure. The user: "Alarms are good but could be better."
Advance-notice preference [P-NOTES-01, 4/20] — Configurable pre-alert window.
Modality preference is bimodal — P-NOTES-01 (4/20) strongly prefers vibration (privacy). P-NOTES-02 (4/20) prefers voice / talking ("but all are good").
Enthusiastic reaction to watch-based reminder [P-VOICE-01 4/20, P-NOTES-03 4/20].
"Not there yet" posture [P-VOICE-03 4/20, P-NOTES-04 4/20].
Phone alarm as active liability [P-VOICE-01, 4/21] — Two alarms per day are set and working but produce zero adherence action. The reminder system has become a caregiver-interruption system.
No reminder system used [P-VOICE-03, 4/21] — Self-reported perfect adherence without external cues.

4. Refill & sync experience — full signals¶

Refill anticipation signal (unsolicited) [P-NOTES-01, 4/20] — 2–5 days before monthly refill due, 2 weeks for 90-day fills. Explicit pharmacist-lead-time rationale.
Refill cadence aligns with anchor-site supply distribution (60.7% on 30-day, 17.9% on 90-day).
Refill signals sparse elsewhere — surfaced primarily in P-NOTES-01 (pharmacist-literate). Most patients did not raise refill unprompted. Formal screener must actively probe.
No refill data surfaced in 4/21 batch.

5. Health context & living situation — full signals¶

Dual-role participant [P-NOTES-01, 4/20] — Self-managing + caregiving simultaneously. Distinct archetype.
Adult-child care circle dominates [P-VOICE-01 4/20, P-NOTES-02 4/20, P-NOTES-03 4/20, P-VOICE-01 4/21] — Primary caregivers across 4 of 11 memos.
Caregiver time burden quantified [P-NOTES-02, 4/20] — 30 minutes/day planning meds and tracking intake.
Extended care-circle aspiration [P-NOTES-02, 4/20] — "Self, sister, doctor, pharmacist, and even hospital if possible" — 5-layer alert circle.
Fragility of caregiver-dependent adherence [P-NOTES-03, 4/20] — Non-adherence when caregiver is away.
Existing fall-alert wearable baseline [P-NOTES-02, 4/20] — Mother already wears a fall necklace. Existing PERS incumbent.
Cannot be alone [P-VOICE-01, 4/21] — Caregiver-dependent by necessity. Adult daughter manages medication oversight. Pre-existing location solution: family placed Apple AirTag in patient's cane. Elopement risk active — "she runs away." Cognitive impairment (dementia) is the underlying condition.
Independent self-manager [P-VOICE-03, 4/21] — Living independently. Monitoring multiple chronic conditions (blood glucose, blood pressure). Wants to involve her prescriber in the product acquisition decision.

6. System failures & transition points — full signals¶

NDC / manufacturer changes [P-NOTES-01, 4/20] — White pill becomes blue pill between refills.
Pharmacist > doctor trust [P-NOTES-01, 4/20] — "If pharmacist tells patient don't use the boxes, keep it in bottle, they listen better than the doctor."
Travel as failure point [P-NOTES-01, 4/20] — Pill box failed during travel.
Caregiver-away as failure point [P-NOTES-03, 4/20] — 2-week gap.
Phone alarm as active liability [P-VOICE-01, 4/21] — Existing reminder infrastructure doesn't just fail; it creates a burden loop. Family compensated with an Apple AirTag in the cane (location workaround), but the medication loop has no working solution.
Prescriber as gatekeeping authority [P-VOICE-03, 4/21] — Patient wants to know if her prescriber can provide or recommend the device before she will acquire it.

7. Adoption readiness & motivation — full signals¶

5/5 power user [P-NOTES-01, 4/20] — Pharmacist-literate, would recommend broadly.
4/5 caregiver [P-NOTES-02, 4/20] — Active system-builder; would benefit materially.
4–5 patient enthusiasts [P-VOICE-01 4/20, P-NOTES-03 4/20] — "Loves the idea"; explicit Medicare-coverage wish.
"Not there yet" [P-VOICE-03 4/20, P-NOTES-04 4/20].
Concept never rejected — 0 of 11 participants rejected the concept outright.
Medicare-coverage wish is explicit and recurring [P-NOTES-01 4/20, P-NOTES-03 4/20] — Two separate participants independently raised Medicare as a value-unlock.
Caregiver-driven adoption [P-VOICE-01, 4/21] — Daughter is "very interested in the device." Motivated by location/GPS for elopement, fall alert, and closing the medication loop. Not price-sensitive at $35/month for those features.
Biometrics as the activation feature [P-VOICE-03, 4/21] — Only conditionally interested — if the device can read blood glucose and blood pressure. Without that, no reason to wear it (she doesn't have an adherence problem). $35/month acceptable if biometric monitoring included.
Insufficient data [P-VOICE-02, 4/21] — Brief encounter.

8. Technology & product fit — full signals¶

Older-adult tech literacy is higher than assumed [P-NOTES-01, 4/20] — 80+ male already uses iPhone + Apple Watch.
Criticality-tiered UI ask [P-NOTES-01, 4/20] — "Red star" for high-priority meds → maps to the Medication Risk Database tiers.
Double-confirmation for critical meds [P-NOTES-01, 4/20] — Two-signal UX pattern.
In-context guidance ask [P-NOTES-02, 4/20] — "Something told her, take with food or not, or basic consultation tips." Direct validation of an AI medication-info feature.
Vital-signs monitoring asks [P-NOTES-05, 4/20] — BP, temperature, balance.
Waterproof requirement [P-NOTES-05, 4/20] — Hardware spec ask.
Privacy on voice announcements [P-NOTES-01, 4/20] — Public embarrassment risk.
Pricing anchors (4/20): Device $30–35 acceptable [P-NOTES-01]; subscription ~$5/month [P-NOTES-01]; Medicare coverage expectation recurring.
Medicare / Blue Button integration ask [P-NOTES-01, 4/20] — Deductible visibility.
GPS/location is the #1 feature for caregiver-track [P-VOICE-01, 4/21] — Daughter already solved this with an Apple AirTag in the cane.
Audio + visual medication instruction is a dementia-specific must-have [P-VOICE-01, 4/21] — Caregiver explicitly described needing the device to show a picture of the medication and speak an instruction at dose time.
$35/month cited by two of three 4/21 participants [P-VOICE-01 caregiver, P-VOICE-03] — Independently, across different motivating features (elopement vs. biometrics). Three-session recurring anchor across both dates.
Biometric monitoring as sole conversion feature [P-VOICE-03, 4/21] — Blood glucose and blood pressure monitoring.
Prescriber channel as distribution path [P-VOICE-03, 4/21] — Patient asked whether her prescriber could provide the device to her.
Existing location wearable in use [P-VOICE-01, 4/21] — Apple AirTag placed in patient's cane is already in the field.

Interviewer calibration notes (for field team)¶

Observations for ongoing field interviewer training:

Expect initial denial on missed-dose questions. Probe gently; don't accept first "I never miss it" as final.
"Not there yet" is not rejection. Score as moderate future interest; capture future-trigger language.
Ask about existing systems — "Walk me through how you manage your medications today" is a higher-yield opener.
Probe caregiver-away scenarios — "What happens when your usual helper isn't around?"
Family members come up organically; don't force the caregiver question.
Pharmacist-credibility framing resonates — ground the concept pitch in pharmacist input.
Older-adult participants are more tech-literate than expected.
Price ceiling appears anchored at $35/month.
Patient-track demographics lean elderly female — not sampling bias; population data supports it.
Fatigue is a valid adherence answer.
Dementia-track encounters require a different protocol — pivot to the caregiver quickly.
The biometrics question is live in the field; interviewers need guidance.
Apple Watch / existing consumer wearables are showing up — ask about existing wearable use early.
Prescriber involvement is a real friction point for independent adopters.

Source documents¶

4/20 synthesis: Pharmacy & Clinical/Daily_Testing_Results/2026-04-20/categorized_analysis.{md,docx} + scrubbed raw notes
4/21 synthesis: Pharmacy & Clinical/Daily_Testing_Results/2026-04-21/categorized.md + handoff README
Anchor-site aggregate medication context: 03_Pilot_Deliverables/Rx_Only_Analysis_April2026.{md,docx}

Raw voice transcripts and raw field notes remain local per PHI workflow.