Pulse Health
- —UX Research
- —Service Design
- —UI Design
- —Design System
- —Usability Testing
A digital health platform with 200,000 users and a patient portal that hadn't been touched since 2019. Four years of accumulated debt, and users who had learned to work around its failures. We started from their workarounds.
Healthcare UX fails in a predictable way: it prioritises clinical accuracy over human comprehension. Pulse had a portal that was technically correct and practically unusable. Users were calling their GP to do things the portal was supposed to handle. We had to make the designed version easier than calling a human.
START WITH THE WORKAROUNDS
Before touching any design, we spent three weeks in research. Forty-two user interviews. Three rounds of usability testing on the existing portal. We mapped every workaround users had developed — these are the true design requirements.
The most common workaround: users photographed their own appointment details because they couldn't trust the portal to show them at the right time. That became our first design problem to solve.
CONSISTENCY AS CARE
We built a design system from scratch — 94 components, 12 pattern templates, a full token structure. In healthcare, inconsistency doesn't just frustrate users. It erodes trust. Every component had to be accessible to WCAG AA as a baseline.
— Felix Romero, Frontend DevA design system is only as good as the decisions it makes for you. This one makes the right call by default and makes the wrong call impossible.
NUMBERS THAT MATTER
Task completion rate went from 31% to 87% in usability testing. Average task time dropped by 64%. GP call volume for portal-related queries dropped by 40% in the first month post-launch.
The App Store rating moved from 2.1 to 4.7 stars in six weeks. The most common review category changed from "broken" to "finally works."
RESEARCH
Three weeks of field research, user interviews, and competitive analysis. No assumptions. The brief was wrong in four places — research found them all.
SYSTEM
Design system before screens. We built the token structure, component library, and pattern library before designing a single page. Six weeks of infrastructure work that saved twelve weeks of screen-level rework.
VALIDATE
Three rounds of usability testing with real Pulse users — including elderly users, users with low digital literacy, and users with visual impairments. Iterated after each round.