Predicare

01 Overview

Replacing a 20-year-old case-management system used by nurses at a US health insurer.

02 Problem

Nurses spent most of their day copy-pasting between tabs, not on patient care.

There was no real way for the team to work on a case together, and the case-management process was held together with manual steps. Nurses were spending most of their day on the non-clinical bits — admin, lookups, copy-paste — instead of actual patient work.

03 My role

I owned personas, workflows, wireframes, and the prototype — reviewed by a senior lead.

• — Built the personas out of the contextual interviews — back-office nurse, case manager, claims reviewer — with each one's goals, frustrations, and the actual tasks they did inside the legacy app.
• — Mapped the current workflow to make the pain visible, then drafted what the new one should look like — data pulling itself together, the system surfacing care gaps before the nurse had to look for them, and the team able to work on a case together.
• — Wireframes across the core screens, four rounds, reviewed with the client's clinical experts, the business analyst, and the system architect before anything got locked.
• — Built the clickable Figma prototype with real data so the client could walk the whole workflow before engineering opened a ticket.

Working under a strong lead, what you actually get is someone who'll tell you when your reasoning is the part that needs more work.

04 Design process

Twelve weeks. Five phases.

Twelve weeks on the clock. The business analyst and I split it into five phases — interviews, task analysis, ideation workshops, reviews with the clinical experts, and the hi-fi build.

• 01 User interviews Structured interviews with nurses, case managers, and claims reviewers
• 02 Task analysis Detailed walkthroughs of the existing case management workflow
• 03 Ideation workshops Joint sessions with the BA, SMEs, architect, and developers
• 04 SME reviews Wireframe rounds reviewed with partners before sign-off
• 05 Hi-fi design Visual system established, then extended across the remaining screens

05 User research

HIPAA blocked me from touching the app, so I sat next to nurses instead.

Analysis

Key insights

• — The system worked. It just made everything slow. Nurses weren't fighting bugs; they were fighting click counts.
• — The interface looked twenty years old, and people felt it. The thing they kept asking for was a product that looked like it belonged in this decade.
• — Most of the daily work was busywork — copying data from one tab to another to answer one question.

06 Define

We rewrote the workflow before drawing a single screen.

07 Ideation

Wireframes were where the argument happened.

Once the workflows were locked, I built a clickable wireframe with real data so the client could feel the experience instead of imagining it. The wireframes did most of the heavy lifting in the reviews — by round four the major calls had been made, and we moved into the high-fidelity build.

For the visual side, I looked at a handful of well-designed enterprise products to find a starting language. Two of the busiest screens became the place where the design system was set — and the rest of the screens built on that base as we went.

08 Key solutions

Auto-pulled patient data. Shared case surface. Predictive care plans.

Pulled patient data from the different external systems automatically, reconciled it, and rolled it up into one master report — so the nurse wasn't doing the assembly by hand any more.

Made case assignment and patient information a shared surface, so anyone reviewing the case could see what their colleagues had already done and pick up the thread quickly.

Used predictive analysis on the patient's health record (EHR) to flag care gaps and suggest a tailored care plan, instead of waiting for the nurse to spot the pattern.