
That distinction is the whole premise behind the Ethos Design System. GE HealthCare's software portfolio spans more than 200 applications: imaging suites, patient monitors, diagnostic tools, cloud platforms, built over years by teams that rarely spoke to each other. Each had its own logic, its own visual grammar, its own way of asking a nurse to confirm a dose. None of that fragmentation showed up on a spec sheet. It showed up as a half-second of hesitation, repeated across thousands of clinicians, thousands of times a day.
EDS was built to close that gap. Not as a style guide layered on top of existing products, but as the shared visual foundation an entire portfolio could finally stand on. One recognizable language, flexible enough to work everywhere it needed to, familiar enough that nobody has to relearn it under pressure.

Defined the typography, iconography, motion, and color foundations now running across 100+ healthcare products in Imaging, Ultrasound, Patient Monitoring, and Pharmaceutical Diagnostics.
Built the animation and interaction behavior across every component and interaction in the system, defining how transitions, feedback, and state changes move, not just how they look.
Contributed directly to the EDS component library, from spec through to shipped, tested across screen sizes from bedside monitors to desktop dashboards.
Wrote usage rules and documentation that let other teams build correctly with the system without needing to ask.
Partnered with engineering, product, and research to turn principles into production-ready components, so what shipped matched what was designed.
The complexity compounds because no clinician lives inside a single interface. Over one shift, they move across bedside monitors, tablets, desktop systems, and shared displays, and everything changes with them. Screen size changes. Input changes, from touch to mouse to keyboard, sometimes through gloves. Lighting shifts from a dim ICU to a bright operating room. The obvious fix, one rigid interface everywhere, was never realistic across that much hardware and that many use cases. The harder, more honest goal was building one recognizable visual language that could flex across every context without asking anyone to relearn it.






The second insight came from watching how people actually behave under stress. They don't explore, they rely on habit. A clinician mid-shift expects a button to work the way it worked the last three times. That expectation became the argument for shared pattern libraries, not as a nice-to-have, but as a direct lever on error rates. Once interaction states, transitions, and behaviors stop being open to interpretation, the interface adapts to the user's expectations instead of asking the user to adapt to the interface.

The third thread tied it all together: continuity between the physical and the digital. Clinicians don't think in terms of platforms or systems. They experience one uninterrupted workflow that happens to touch a dozen different screens. EDS was built to erase the seams in that workflow, aligning tokens, interaction states, icons, and motion so a bedside device and a cloud application speak the same visual language without translation.

Every color tied to severity or state carries specific meaning, so urgency reads instantly rather than being inferred.
Layouts were structured so priority is visually obvious. Without it, everything on screen competes for equal attention, which is its own kind of risk.
Buttons, feedback, and transitions behave identically everywhere, turning consistency into a direct tool against hesitation.
A subtle animation confirming an action removes doubt in a way static feedback can't. These are small decisions with an outsized effect on confidence.
Color, spacing, and typography lived as tokens rather than fixed values, so intent moves from design to code without ambiguity, and themes like dark mode adapt without a redesign.
Every component was built to handle touch, mouse, keyboard, and gloved input, because clinical hardware varies wildly by department and by hour.
Typography was tested against the conditions that actually break it: a dim ICU at 2 AM, a fluorescent-lit or at noon, a glance from three feet away mid-procedure. Every type scale and contrast ratio holds up against WCAG 2.1 AA, so accessibility isn't a compliance checkbox bolted on at the end. It's load-bearing.







That continuity is the whole point. The system was never meant to be noticed. It was meant to disappear into the workflow, doing its work quietly in the moments when a clinician has no attention to spare on figuring out an interface. When it works, nobody thinks about the design at all. They just move faster, with less doubt.
For the teams building on top of EDS, the tax of reinventing typography, contrast, and interaction from zero disappeared. What they inherited instead was a system already tested against real clinical behavior, freeing their time for the specific, harder problems each product actually needed to solve.


But the number worth sitting with isn't a dollar figure. It's the shift in what design and engineering stopped arguing about. When a component in Storybook is the same component in Figma, teams stop negotiating over interpretation gaps and start contributing back to the system itself. EDS moved from a top-down initiative to something closer to shared infrastructure people had a stake in improving.
What this project proved, more than any metric, is that the "life-saving pixel" isn't a metaphor. It's the accumulation of thousands of small, intentional decisions: a color, a spacing value, a consistent transition, each one reducing friction at a moment when friction has a real cost. Done well, that kind of design becomes almost invisible. But it shows up anyway, in faster decisions, in fewer errors, and ultimately, in better care.
