Meet Mr. Ouch. He's a patient, but he's also your parent, your grandparent, or just someone tired and hurting who wants to get discharged home.
A discharge decision has real money and real health riding on it. Send Mr. Ouch home too soon and he can land back in the hospital, sicker and now more expensive to treat. Keeping him too long carries its own risk: hospitals need the bed back, and a longer stay means more exposure to complications he would not have faced at home. It takes a team of well orchestrated professionals — doctors, nurses, therapists, and care coordinators — to run these checks, and naviHealth's software was supposed to help them do it well. When it slowed them down instead, decisions got made late or without enough information behind them, and ultimately people like Mr. Ouch suffered.
The original ask was a "lift and shift," a facelift with no structural change underneath — the kind of ask that assumes UX means visual treatment and nothing more. This was my first major engagement at naviHealth, and I used it to set the tone for how I'd work with their engineering team going forward: UX needed a real say in how the system worked, not just how it looked. The legacy screen backed that argument up. It was still carrying fields and steps left over from contract requirements that had expired years earlier — dead weight nobody had gone back and cleared out — and it forced very different jobs through the same screen, a field nurse and a centralized data-entry specialist staring at identical layouts to do work that had almost nothing in common. A facelift would have just repainted both problems.
"After years of piecemeal additions driven by shifting business priorities, Coordinate had become a Swiss Army knife — a tool that tried to do everything, and did none of it particularly well."— on the legacy interface
The legacy layout forced an inefficient, non-linear scan just to complete basic tasks. I mapped the data users needed most against the F-Pattern — the natural eye-tracking path people take through dense interfaces — and used it to put the right data in the right place and cut cognitive load.
Low-fidelity wireframes kept users focused on layout and interaction, not visual polish still being developed in parallel — letting people give honest feedback on structure without the distraction of aesthetics. Testing surfaced real efficiencies, like re-sequencing tasks to cut extraneous clicks — observations from the people doing the work, not designer opinions.
Patient demographics surfaced at a glance in the header. Care settings and clinical assessments were prioritized by hierarchy. Authorizations, tasks, and documentation were reordered to match real clinical workflow rather than the underlying data model.
That argument won, and it set the terms for the rest of the engagement. A user-centered approach increased trust in UX from both Product and Engineering, and by the time I left, the UX team was helping define product direction, not called in afterward to supply visual polish. Mr. Ouch still had a long road between the hospital bed and home — but the people helping him get there finally had software that worked with them, not against them.
Continue to the next case study.
→ View HCA HealthcareSquatchbot
Mitch's (slightly hairy) assistant