A health technology client came to us with a familiar brief: improve the dashboard. Usage data showed high login frequency but low interaction with critical alerts. The assumption was visual design — hierarchy, colour, density.
Research told a different story. Clinicians were not ignoring alerts because they could not see them. They were ignoring alerts because the system had cried wolf for eighteen months. Every notification looked urgent. None felt trustworthy.
Research in clinical context
We shadowed sessions across two sites — not lab tests. We counted interruptions per hour, documented alert response sequences, and interviewed nurses about workarounds they built outside the product.
The workaround pattern was consistent: teams maintained parallel paper lists for what actually mattered. The official dashboard had become administrative theatre while real decisions happened elsewhere.
When users build a shadow system, your product has stopped being infrastructure and started being obstacle.
Strategy: tier urgency, not volume
We separated informational updates from actionable alerts from escalation events — each with distinct visual language, delivery channel, and acknowledgement requirements. We reduced total notifications by forty-one percent in pilot while increasing response to critical alerts.
Design changes alone would have failed without operational agreement on what qualifies as critical. The case study lesson is familiar: experience design is policy design when behaviour is the outcome.
Outcome and durability
Clinicians reported spending less time dismissing noise and more time on patient-facing work. Support tickets about missed alerts dropped in the first quarter post-rollout. The product began earning back the role it was purchased to fill.
This is what grounded in research, built to last looks like in practice — not a prettier dashboard, but a system people choose to trust again.