Visibility saves lives

A nurse at 3 a.m. is not your average user. She's tired, she's running three patients, and the screen in front of her is making a decision about insulin she'll have to defend if the patient codes. The interface isn't a productivity tool. It's safety equipment.

I work on insulin management software. Insulin is one of the three medications in any hospital most likely to cause an adverse event — give too much and a patient goes hypoglycemic; give too little and their blood sugar climbs into the territory where things start to fail. The right dose is a moving target that depends on weight, kidney function, what they ate, how stressed their body is from infection. No human keeps all of that in their head. They shouldn't have to.

So design steps in. Not "design" as in colour palettes. Design as in: which number do you put first on the screen?

What "visibility" actually means in clinical software

In consumer UX we talk about visibility as part of Nielsen's heuristics — show users where they are in the flow. In healthcare it's a different verb. Visibility means: can the clinician see, in the half-second they have, the thing that matters most?

• The current blood glucose, and where it's headed.
• The last dose, and how long ago it landed.
• The next decision the system is recommending — and why.
• The patient's trajectory over the last 12 hours, not just the last reading.

Get those four right and you reduce errors. Get them wrong — bury them in a tab, hide them behind a modal, demand a click — and you've added cognitive load to a person who already has none to spare.

"The design of a clinical interface is the design of how a decision feels. Confident or hesitant. Clear or hedged. That feeling is the product."

The two failure modes

I see clinical software fail in two predictable ways.

The first is hiding the work. A system shows a recommended dose but doesn't show the reasoning. The clinician now has two choices: trust the black box, or open a different system to verify. Most will verify, which means the software didn't save time — it added a verification step.

The second is dumping the work. The opposite reflex: show every input, every variable, every chart. The clinician has to do the synthesis the software was supposed to do. Cognitive load goes up. Trust goes down.

The job sits in the middle: show enough reasoning to be trusted, hide enough complexity to be usable. That balance is craft. It's not a design system token. It's not an AI prompt. It's the work.

The standard is different when decisions carry risk

The standard is different when decisions carry risk

I have designed dashboards. I like dashboards. There is real value in helping teams understand performance, spot patterns, and make better business decisions.

But clinical software has changed the way I judge visibility.

In many products, a confusing interface creates friction. In healthcare, confusion can create risk. The margin for ambiguity is smaller. A buried value, an unclear recommendation, a poorly timed alert, or a missing explanation can change how confidently someone acts.

That does not mean every screen needs to be heavy or overly cautious. It means the design has to respect the environment it lives in.

Healthcare is full of constraints: legacy systems, compliance requirements, fragmented workflows, exhausted teams, and decisions that need to be made across shifts. It is not always the most glamorous design space. But it is one of the clearest.

You learn quickly whether the work is helping.

That is why clinical visibility matters. Not because it makes the interface more impressive, but because it makes the decision easier to understand, easier to trust, and harder to get wrong.

— Victor