DHRUMILSavsani

MediSync, Healthcare Platform

Records, reminders and care, designed to feel clear and easy to read.

Project hero

The result, first

I cut appointment booking from 9 steps to 4 and took every screen to WCAG 2.1 AA, medication-reminder engagement rose 2.3×.
Product

MediSync, a patient-facing mobile app for health records, appointment scheduling, and medication reminders.

Business problem

Existing health apps assumed a young, tech-literate user; this provider’s cohort skewed 55–72 with real accessibility needs.

Target users

Older patients managing chronic conditions, low vision, motor impairment, and low smartphone confidence.

MediSync is a patient-centred mobile platform for health records, scheduling, and medication reminders. I designed it for a healthcare provider whose patients were being failed by apps built for a younger, more confident user.

One number set the tone for the whole project. 60% of the existing screens were failing WCAG 2.1 AA. So accessibility was not a finishing touch here, it was the reason we were redesigning the app at all.

I owned UX and UI across a three-month build, working alongside a clinical review team, an engineering lead, and an accessibility consultant.

I lead with the result because in healthcare the outcome, patients actually booking and taking medication, matters more than the craft that produced it.

Who I was designing for

The first decision I made was to stop testing with the usual young, tech-comfortable users. The people who needed MediSync the most were also the ones who would struggle the most with a badly designed screen. So I made sure the research participants actually matched the real patients, older adults with genuine accessibility needs, and not younger stand-ins.

This one choice brought out problems that no other test group would have shown me, and it became the standard I checked every later decision against.

About

Manages a health condition and a daily medication schedule, and would rather stay independent than ask family for help with an app.

Goals
  • Book and attend appointments without help
  • Keep medication on track
  • Read every screen without straining
Frustrations
  • Tap targets too small to hit reliably
  • Pinch-zoom calendars are unusable
  • Ends up calling the clinic instead

Large touch targets alone weren’t enough, adjacent targets within 8px still caused mis-taps, so spacing mattered as much as size.

About

Takes several medications a day at times that shift with work and meals, and needs prompts that fit the real routine, not a fixed clock.

Goals
  • Take every dose on time
  • Fit reminders around a changing day
  • Avoid missed or doubled doses
Frustrations
  • A fixed 8am reminder never matches reality
  • Notifications ignored once they feel wrong
  • No way to set personal dose times

Medication engagement depended on timing, not visibility; a fixed morning reminder missed people who dosed throughout the day.

About

Recently got a first smartphone and approaches new apps cautiously, worried that a single wrong tap will break something or lose progress.

Goals
  • Get set up once, correctly
  • Recover easily from a mistake
  • Feel safe exploring the app
Frustrations
  • Every extra setup step risks failure
  • No clear way to undo a wrong choice
  • Abandons onboarding when unsure

Every added setup step was a risk, so the reminder-configuration flow had to forgive a wrong first attempt.

Before MediSync, a 65-year-old with low vision would open the old app, fail to pinch-zoom the calendar, give up on the booking, and simply call the clinic instead. So the app did not save her any time. It just added one more failed step in front of the same phone call it was supposed to replace. I mapped this current journey to find exactly where things were breaking, and the drop at the calendar is where most of the bookings were dying.

1Open MediSync

Margaret needs a check-up and opens the app.

AnimationLarge type loads first
2Pick a date

A large tap-target calendar replaces the pinch-to-zoom picker.

AnimationCalendar grid fades in
3Confirm visit

One clear confirmation, the “done” before the details.

AnimationToast slides up
4Reminder, her time

She sets the reminder to her own dose schedule, not a fixed 8am.

AnimationSaved state
Margaret, Patient, 68

Wants to book a visit and stay on top of her medication without calling for help, but the old app fought her at every step. This current-state map is what the redesign set out to fix.

Open the app
Book a visit
Fall back to the phone
Miss a dose
  • Opens the old app to book a visit
  • Text and tap targets too small to read
  • Reaches the calendar at step 4
  • Can’t pinch-zoom to pick a date
  • Abandons the booking
  • Calls the clinic instead
  • Fixed 8am reminder fires
  • Doesn’t match when she doses
Satisfaction
“I can barely read this.”
“I can’t select the day, I give up.”
“Easier to just ring them.”
“It reminds me at the wrong time, so I ignore it.”

For me the personas here were not just slides. They were basically my recruitment criteria, because if I had tested with the wrong group, the exact problems I was trying to find would have stayed hidden.

What made the cut, and what didn't

The clinical team wanted more than three months could safely deliver, so I had to treat scope as a triage exercise instead of a wish list. I anchored on the three flows with the highest drop-off, records access, scheduling and medication reminders, and made everything else earn its place against these.

Messaging, family access and health tracking were all cut from v1 on purpose. Each one was useful, but each one was also competing for attention with a flow that patients were already failing to finish.

Shipped in v1

  • Health records access, offline-capable
  • Appointment scheduling, rebuilt end-to-end
  • Medication reminders tied to the patient’s own schedule

Deliberately cut

  • In-app messaging: High effort, and it pulled focus from the flows patients were abandoning today.
  • Family / caregiver access: Raised consent and privacy scope the timeline couldn’t responsibly cover in v1.
  • Health tracking: A retention feature for users we hadn’t yet earned, premature before the core flows worked.

Content was sequenced to the moment of need: a plain-language medication label at the point of dosing, a confirmation the instant an appointment was booked, and nothing extra in between. The clinical team reviewed every health-related string for accuracy before it shipped.

Cutting three real features was the most useful scope decision I made. Protecting the flows that were already failing was more important than adding new flows that nobody had validated yet.

How the product was organised

Since the users had low confidence with smartphones, the structure had to be shallow and predictable, so that a patient never had to remember where something was kept. I divided the product into three clear areas, each one mapping directly to one of the three jobs the app does.

The booking flow is where the structure did the heaviest work. The abandonment was clustering at step 4 of a nine-step path, which was the calendar picker, and not at the final step. So the fix had to be structural and not just cosmetic.

  • MediSync
    • Onboarding
      • Sign up
      • Verify identity
      • Set reminders
    • Records
      • Summary
      • Documents
        • View
        • Download (offline)
      • Appointment history
    • Scheduling
      • Upcoming
      • Book appointment
        • Pick date
        • Confirm
      • Reschedule / cancel
    • Medication
      • Today’s doses
      • Reminder schedule
        • Add reminder
        • Edit times
      • History
    • Account
      • Profile
      • Accessibility
      • Help

Book an appointment, 9 steps down to 4

I removed the multi-touch calendar drum that motor-impaired patients couldn’t operate and replaced it with a single tap-target grid. The path now branches only once, book new versus reschedule, and every decision point is a full-screen, single choice.

Set a medication reminder

Instead of a fixed system time, the flow asks the patient for their own schedule, then confirms each reminder against the prescription. It adds one setup step, which I accepted because it fixed the signal-to-noise problem that made the old reminders easy to ignore.

I designed the flows around where people were actually dropping off, and not where the step count simply looked long. The data was pointing at step 4, so that is where I concentrated the redesign.

From lo-fi thinking to hi-fi execution

I stayed in grayscale wireframes far longer than usual. With accessibility as the constraint, I needed to prove layout, hierarchy, and touch-target spacing before colour could paper over a structural problem.

The wireframes existed to answer one question per screen: can this patient find the primary action without help? Everything decorative was deferred until that answer was yes.

The low-fidelity layouts tested spacing above all. I mocked adjacent targets at varying gaps and confirmed that anything under 8px caused mis-taps, the finding that set the spacing rule for the whole system.

Moving to high fidelity, the biggest change was the calendar. The native date picker survived in early prototypes until testing killed it, replaced by a custom grid that cost roughly 3× the build time. The clinical risk of an inaccessible picker made that trade obvious.

I kept the wireframes deliberately plain for as long as I could, because a grayscale layout cannot hide behind good looks if a patient genuinely cannot reach the button.

The look, the system, and the words

At the surface level I treated legibility as an actual design goal and not just a compliance checkbox to tick off. The type scale, contrast and spacing were all tuned for the real users, and the important clinical flows were pushed beyond AA to AAA contrast.

The end result feels calm and unhurried, and for an anxious patient reading a diagnosis, that calmness is itself a feature.

Gallery image 1
Gallery image 2
Gallery image 3
Gallery image 4
Accessible calendar

Large, 8px-plus-spaced tap targets, the reusable pattern that unblocked motor-impaired booking.

Type scale

Sized for low vision first; every step meets AA, clinical content meets AAA.

Touch-target tokens

Minimum size and spacing encoded as tokens so no future screen could reintroduce the mis-tap problem.

Booking confirmation“You’re booked, Tuesday, 10:40 AM”: Leads with the reassurance, then the detail; anxious users need the “done” before the specifics.
Reminder setup“When do you usually take this?”: A question, not a form field, it invited the patient’s real schedule instead of imposing a default.
Offline state“Saved. This will sync when you’re back online.”: Removes the fear that a lost connection lost their data, critical for low-connectivity patients.

The microcopy carried a lot of the emotional weight here. In health software the difference between “Error” and “Saved, we will sync later” is really the difference between panic and trust.

What testing and the numbers said

I tested the prototype with 6 patients from the target population. Five completed reminder setup on the first try; two needed a second attempt, which flagged the onboarding step as still too fragile and sent me back to make it more forgiving.

4Steps to book an appointment (down from 9)
100%WCAG 2.1 AA compliance across all screens
2.3×Improvement in medication reminder engagement
23Accessibility failures resolved

Offline access shipped for records and upcoming appointments but not for the medication schedule, which, for patients in low-connectivity areas, was just as critical. I scoped offline too narrowly, and I’d widen it from the start next time.

I tested with real patients and not with colleagues, because that second-attempt rate on reminder setup is exactly the kind of failure a confident internal tester would have sailed right past.

What this project changed in me

Designing for a 65-year-old with low vision and a complicated medication schedule taught me one thing very clearly. Accessibility is not a limitation on good design. It is good design. Every constraint this project put on me ended up making the product clearer for everyone, and not only for the users it was built for. So now I no longer keep accessibility for a final audit. It is the first question I ask.

In short

MediSync is a patient-first healthcare app for older adults managing records, scheduling, and medication. Designing against a cohort aged 55–72 with real accessibility needs, I recruited matching research participants, cut booking from 9 steps to 4, and tied medication reminders to each patient’s own routine. Every screen meets WCAG 2.1 AA, key clinical flows reach AAA, and reminder engagement rose 2.3×. For me accessibility was not a checklist at the end. It was the brief itself.

Next project

Sentinel Dash