Narcolepsy · Type IIWakefulnessSleep attacks + difficult mornings
PREDICTA · ACCESSIBILITY UX CASE STUDY
Predicting better days by designing for the hardest ones.
Predicta is a symptom prediction app co-designed with a chronically ill graduate student to forecast high-risk days before they happen—with passive data and almost effortless logging.
Predicta
Tuesday · 7:40 AMHighrisk today
Today’s forecast
Plan for a gentler day.
Sleep disruption, weather change, and a full calendar are raising your symptom risk.
Report how I feelRemind me later
Sleep continuity42%
Weather shift+18%
Calendar loadHeavy
Recent symptoms2 days
Built to explain risk before asking for effort.
ROLE
Co-designer
Team of 4. Research, concept direction, prototype screens, and independent v2 development.
COURSE
INFO-H 581
Experience Design & Evaluation of Access Technologies · IU Indianapolis
TYPE
Accessibility UX
Co-design and assistive technology
OUTPUT
Mobile Product
Validated mid-fidelity prototype
OVERVIEW
What if a difficult day could be predicted before it begins?
Sam lives with Type 2 narcolepsy, chronic migraines, and an autoimmune condition. Energy, pain, and wakefulness can shift without warning, so the design challenge was direct: create something that still works on the day it is needed most.
Co-design with lived experience
Sam’s routines, workarounds, and feedback shaped the product from the first conversation.
Predict with passive signals
Wearable, sleep, calendar, and weather data reduce the burden of constant manual logging.
Validate with the person affected
Screen-by-screen reviews turned Sam’s feedback into concrete interaction decisions.
THE ACCESSIBILITY TEST
The hardest day cannot require the most effort.
01
Predict before symptoms
02
Ask for one tap
03
Save the record
Predicta shifts the burden from the person to the system.
OPPORTUNITY
Turn unpredictable symptoms into a day Sam can prepare for.
The opportunity was not another tracker. It was a low-effort prediction layer that reads passive signals, explains risk, and lets Sam respond when energy is available.
Predict before symptoms
A simple risk forecast helps Sam plan before energy, pain, or wakefulness shifts.
Ask for less
One-tap reporting and deferred reminders preserve data without demanding sustained attention.
Explain every signal
Showing why risk is elevated supports trust instead of presenting a mysterious score.
CASE STUDY SUMMARY
01
Problem
Existing health tools demand sustained logging precisely when symptoms leave Sam with the least energy to give.
02
Approach
Co-design, competitive research, affinity synthesis, rapid concepts, and screen-by-screen validation with Sam.
03
Outcome
A validated mid-fidelity forecasting concept and an independent v2 roadmap for reminders, offline logging, and trust evaluation.
04
My role
Team of 4. Research, concept direction, prototype screens, and independent v2 development.
02 / PROBLEM
Every existing app asks for effort at the moment Sam has none left.
Mornings require medication and a wall of alarms. Afternoons can bring sudden sleep attacks. Migraine episodes make light, noise, typing, and complex navigation painful or impossible. The tracking tools meant to help become another burden when symptoms peak.
WAKEFULNESS
Unpredictable
Narcolepsy makes waking up—and staying awake—unreliable.
MIGRAINES
Sensory overload
Bright screens and interaction-heavy tools can make logging impossible mid-episode.
TRACKING
Too much effort
Existing apps demand sustained input precisely when Sam has the least energy to give.
Lived system · interacting conditionsComplex disability is not a list. It is a system.
Autoimmune disorderEnergyFatigue + joint pain
MigrainesSensory loadLight + noise sensitivity
InflammationPainReduced movement + dexterity
One person · one dayComplex disabilityWakefulness · energy · pain · sensory load
Old modelRemember → type → interpret → repeat
The person carries the tracking burden during the hardest moment.
Predicta modelSense → predict → explain → ask once
The system does more work before requesting attention.
THE DESIGN QUESTION
How might we help Sam anticipate a difficult day and log symptoms without demanding energy, typing, or attention?
03 / OUTCOME PREVIEW
Predicting risk before the day begins.
Before unpacking the process, here is the experience the research led to: a low-effort forecast that explains risk, supports one-tap reporting, and lets Sam defer logging until energy returns.
9:41
PredictaToday
Tuesday’s outlookHigh risk
Prepare before symptoms peak.
Why this forecast
Interrupted sleep
Weather change
Heavy afternoon
A forecast that respects fluctuating capacity
Support arrives before the difficult moment.
NightPassive signals
Sleep and wearable data update without asking Sam to log.
MorningExplain risk
A clear forecast shows what changed and why it matters.
EpisodeOne-tap response
Capture the moment without typing through pain or fatigue.
LaterComplete when ready
A gentle reminder preserves context after energy returns.
04 / EVIDENCE
Six pain points became non-negotiable design requirements.
A one-hour semi-structured interview covering daily routines, assistive technology, workarounds, and frustrations showed how much infrastructure Sam had already built alone—from math alarms to weather tracking.
1:1primary co-designer
“The logging tool cannot become another symptom.”Research synthesis
Starting the day already requires a system of medication and alarms.
Fatigue and sleep attacks make follow-through unpredictable.
Typing becomes hardest at exactly the moment the data matters.
Explaining an episode can demand more energy than Sam has.
Light, noise, and unfamiliar environments can trigger symptoms.
Risk is discovered after the day is already in motion.
Research translated into requirements
Reduce interactionPredict earlierExplain the scoreSupport deferralLower sensory load
01
Waking is unreliable
Starting and sustaining the day requires medication, alarms, and repeated effort.
02
Fatigue arrives without warning
Sudden sleep attacks make planning and follow-through difficult.
03
Logging competes with symptoms
Typing and navigation are least possible at the moment data matters most.
04
Communication breaks down
Mid-episode, even explaining what is happening can become too demanding.
05
Sensory input triggers episodes
Light, noise, and screens can bring on a migraine without warning.
06
Unfamiliar places carry risk
Sensory conditions in new environments are unknown until Sam is already there.
05 / MARKET GAP
The market confirmed the same failure: every tool expects energy.
We reviewed Bearable, Flo, Migraine Buddy, and Alarmy alongside passive signals already available through Apple Watch, weather data, and Sam’s TENS unit. None combined low-effort logging with meaningful prediction.
Competitive opportunityThe gap was not tracking. It was asking less.
Four products solved pieces of the problem. None shifted enough work away from the person.
CapabilityBearableFloMigraine BuddyPredicta
Passive signalsPartialMissingMissingDesigned in
Predictive forecastMissingMissingMissingDesigned in
One-tap loggingPartialPartialPartialDesigned in
Explainable riskMissingMissingMissingDesigned in
Log laterMissingMissingMissingDesigned in
Low-sensory defaultPartialPartialPartialDesigned in
WHAT WORKS
Low-effort UI
Flo showed that a clean, focused interface can reduce daily friction.
WHERE THEY FAIL
Input-heavy tracking
Bearable and Migraine Buddy require sustained logging that Sam has already abandoned.
OPPORTUNITY
Passive prediction
Combine wearable and weather signals with one-tap reporting to forecast risk before symptoms peak.
06 / FLOW DECISION
Ask the system to predict before asking Sam to log.
Affinity mapping identified unpredictability as the upstream problem. Five concepts converged on a prediction-first flow: passive data before active input, one decision per screen, and no typing at the worst moment.
Affinity synthesis27 observations → 3 clusters → 1 root problem
Reminders + routine
The day depends on scaffolding
Symptoms + unpredictability
The upstream problem
Selected root problem
Morning wake-up
Highest effort before the day starts
Design priorityReduce unpredictability first; routine and follow-through become easier downstream.
01Daily readiness
02Smart wake-up
03Zero-effort log
04Environment layer
05Invisible support
Selected directionPredicta
Prediction + one-tap logging + deferred follow-through
01Passive before active
02One decision per screen
03No typing at the worst moment
Sleep, weather, calendar, and wearable signals update quietly.
Predicta identifies elevated risk before Sam begins the day.
The interface shows which signals changed and what they mean.
Sam reports with one tap or chooses to be reminded later.
The system builds a useful pattern history without daily burden.
Accessibility guardrail
Every step must remain understandable when attention, dexterity, light tolerance, or wakefulness is reduced.
07 / VALIDATION
Co-design turned feedback into the next iteration.
After an initial design critique, we walked Sam through the prototype screen by screen. This was not usability testing with a stranger; it was a design review with the person the product exists for.
Co-design review
“If I’m not good five hours from now, can I do a second remind?”Sam · primary co-designer
BeforeOne reminder
A single snooze assumed energy would return on schedule.
AfterStackable reminders
Repeatable, configurable intervals let Sam decide when to return.
01Less forced input
02More user control
03Clearer recovery path
SAM ASKED
“If I’m not good five hours from now, can I do a second remind?”
Reminders became stackable and repeatable, with configurable snooze intervals instead of forcing immediate logging.
08 / ACCESSIBILITY-FIRST
Dark-first is an accessibility decision, not an aesthetic one.
Bright screens trigger Sam’s migraines, so the visual system prioritizes low stimulation, clear contrast, static surfaces, gentle language, and large, forgiving controls.
#0E1713Primary surface
#7BAE8AAction + status
#DDEFE3Low-stimulus wash
#F7F8F5Quiet background
Plan for a gentler day.
Clear hierarchy · sentence case · plain language44px+minimum target
4.5:1contrast goal
0required typing during an episode
1primary decision per screen
09 / FINAL SOLUTION
Predicta forecasts risk before the day begins.
Sleep, weather, energy, calendar, and wearable signals become a simple risk forecast. Predicta explains why risk is high, supports one-tap symptom reporting, allows “remind me later,” and builds a doctor-ready history without daily burden.
Final experience · reserved mediaThree moments define the product.
These frames are intentionally ready for the final prototype screenshots.
01
Today’s forecast
Replace with final prototype screenshot
+02
Why risk is elevated
Replace with final prototype screenshot
+03
Report now or later
Replace with final prototype screenshot
+Forecast before effortExplain before askingLog when ready
01
Prediction first
Today’s risk level leads the screen so Sam can plan before getting out of bed.
02
Explainable
Predicta shows why risk is elevated and offers gentle, actionable guidance.
03
Effortless
Report now and log later—without typing through pain or fatigue.
10 / IMPACT
A concept project, measured honestly.
Predicta has not shipped, so there are no product metrics to claim. The prototype established a validated direction; these are the measures I would use next to evaluate effort, comprehension, follow-through, and trust.
Effort
Taps + timeCan Sam log on a high-symptom day?
Comprehension
Reason recallDoes the forecast explain itself?
Follow-through
Deferred completionDoes ‘later’ become a completed log?
Trust
Forecast alignmentDoes lived experience match prediction over time?
Prototype outcome
A validated direction and a measurable roadmap—not a claim of shipped impact.
11 / REFLECTION
Accessibility means designing for the worst day, not the average one.
I used to think accessibility meant contrast, readability, and simpler navigation — a checklist. Sam taught me it means designing for someone’s worst day, not their average one. The tasks I’d normally call easy can be impossible during a migraine.
The reframe stuck: stop asking how to make something usable, and start asking how it works when the user has nothing left to give. If I continued, I’d recruit participants with accessibility needs earlier — accessibility became a core design principle, but it was underrepresented in the first round of research.
Built with Ruby, Clara, and David — and most importantly, with Sam.
NEXT PROJECT · 02