Case Study · Care Team Workflow

All Patients — designing the nurse's morning, then handing it off to the rest of the care team

A patient panel built around the question nurses ask first thing every day: who needs me right now, and why? Researched in the unit, designed for continuity across the care team.

My role
Product · UX research · Design lead
Research method
Multi-day contextual inquiry
Primary users
RNs · charge nurses · MA / care team
Stage
Discovery → MLP definition
Summary

The short version

I spent multiple shifts on the unit alongside the nursing team to understand how they actually manage their patients across a day. The single biggest friction wasn't documentation, charting, or alarms — it was deciding who to focus on, and why, in the first ten minutes of every shift. The second, just behind it: making sure that when the shift changed, or a different role stepped in, that thinking didn't get lost. The All Patients view was designed to answer both of those questions on a single screen.

Contextual inquiry Jobs to Be Done Service continuity Usability validation
The challenge

Nurses were starting every shift by rebuilding context from scratch

Existing patient lists ranked by room number or last name — sorting choices that made sense for a printed census but not for clinical attention. Risk signals were scattered across tabs. Notes from the previous shift were buried in flowsheets. Caregivers walked into work, glanced at the list, and then spent the next several minutes asking peers, "anything I should know about?" — a verbal handoff that was reliable only when the right people were in the right place at the right time.

The unit didn't have a "who needs me right now" view. It had a directory. Every nurse, every MA, every covering provider was responsible for assembling that view in their head, every time.
Research approach

Multi-day shadowing on the unit

Rather than start with surveys or interviews, I joined the nursing team across multiple shifts — days, evenings, and the start of a night. The goal was to see real friction in the wild, not the cleaned-up version that gets reported in a meeting.

What I observed

  • Pre-shift huddles where a verbal handoff carried critical context — and where a single absence broke the chain
  • Patient lists sorted by room — and the manual mental re-sorting that nurses did the moment they sat down
  • Risk signals (falls, sepsis triggers, behavioral concerns, isolation status) living in five different places
  • Charge nurses re-prioritizing the unit every two hours based on what they were hearing rather than what the system was showing
  • "Pick-up" moments — a covering RN, an APP, a care manager — each rebuilding context from notes

Methods

  • Multi-day contextual inquiry on the unit
  • Pre- and post-shift JTBD interviews (RNs, charge, APP, MA)
  • Artifact review — printed census sheets, scratch notes, "brain" sheets
  • Synthesis into a daily flow + role-overlap map
  • Concept testing with low-fidelity prototypes between shifts
What we learned

Two frictions stood above the rest

1

"Who do I focus on first — and why?"

The hardest moment of the day was the first ten minutes. Nurses weren't asking for more data — they were asking the system to help them rank attention. Acuity, deterioration risk, time-since-last-touch, pending orders, and social factors all mattered, but no single screen combined them. The mental math fell entirely on the nurse. Newer staff felt this most acutely; experienced staff had built workarounds, but those workarounds didn't transfer.

2

"How do I make sure the next person picks up where I left off?"

The second concern was continuity — both at shift change and across roles. Nurses worried that a covering RN, an APP rotating in, or a care coordinator stepping in wouldn't know why a patient was being watched closely, what had been tried, or what the next decision point was. SBAR existed; what was missing was a persistent, system-level place where that thinking lived alongside the patient — not in a paper "brain" or a side conversation that walked out the door.

Design principles

From research to design intent

Three principles emerged from the synthesis. They became the gatekeepers for every subsequent design and prioritization decision.

Rank attention, not rooms

The default sort is acuity- and risk-aware. Room number is still available, but the system leads with the question nurses are actually asking.

Make context portable

Every attention signal — risk, watch reasons, pending decisions, last-touch — is pinned to the patient row, not buried in tabs. Anyone covering can read it in seconds.

Design for the handoff

The patient row carries a "watch reason" and an open thread of nurse-authored context. When the shift changes or a new role steps in, the thinking transfers with it.

The solution

What "All Patients" became

A single panel that opens with the patients who need attention right now, with the reasoning visible inline, and the thinking persistent across roles. Built as a Minimum Lovable Product so we could validate adoption before broadening scope.

VIEW PAGE DETAIL

Detailed View All Patients · Unit 4 East · Day Shift 🔎 Search or filter… Sort: Attention Filter: All + Add to my list Patient Watch reason Risk & status Last touch Action RJ Rivera, J. Rm 412 · 67 M · Day 2 post-op ↑ HR + low UOP × 2 hrs · sepsis screen due Charge flagged at 06:50 · MD notified SEPSIS FALL Code: Full · Iso: Contact 12 min ago Open CO Connor, A. Rm 408 · 72 F · CHF exacerbation Lasix held — recheck K+ at 09:00 Per Dr. Patel · awaiting lab WATCH Code: Full · Iso: None 38 min ago Open SP Singh, P. Rm 405 · 58 F · Pneumonia, day 3 Discharge planning today · teach-back pending Caregiver arriving 11:00 · Rx ready DISCHARGE Code: Full · Iso: None 22 min ago Open DM Diaz, M. Rm 410 · 45 M · Observation Stable · routine reassessment 10:00 No new orders · ambulating STABLE Code: Full · Iso: None 1 hr ago Open Tip · Tap a row to see the watch thread — the rolling note from previous shifts and roles.

Conceptual mockup of the All Patients view. Patients are sorted by attention need; each row carries the watch reason, risk badges, last-touch time, and a one-tap path into the rolling watch thread that travels with the patient across roles and shifts.

Key features

Attention-ranked list

Default sort is composite: acuity score, deterioration triggers (sepsis, stroke, fall risk), time-since-last-touch, and pending decisions. Charge nurses can override the sort but rarely do.

Watch reason inline

Every flagged patient carries a one-line "why we're watching" written by a clinician — not a system code. New eyes get the gist of the case in seconds.

Rolling watch thread

A persistent, role-aware note thread per patient. SBAR-friendly templates, but the thread is the artifact — not a paper brain that walks out at 7 p.m.

Last-touch & ownership

The row shows who touched the patient last, when, and what's pending. Covering RNs and APPs see immediately where the gap is, not just what room a patient is in.

Continuity of care

How we make sure the next person picks up where the last one left off

Continuity wasn't a feature — it was the second-most-important Job to Be Done in the data. We treated it as a first-class design constraint.

Watch thread, not handoff sheet

The watch thread persists with the patient. Day-shift adds context, evening adds context, the night APP adds context. The thread is the handoff. Nothing important lives only in someone's head or on paper.

Role-aware visibility

RNs, charge, APPs, MAs, care coordinators, and case managers all see the same patient row but with role-relevant emphasis — so an APP picking up a patient sees pending decisions while a care coordinator sees discharge readiness.

Open-loop detection

The system flags patients with no touch, no note, or no resolved task in a configurable window — turning the silent failures of handoff into visible exceptions.

One-tap escalation

Any team member can escalate from the row — to charge, to the rapid response team, or to the covering APP — with the watch thread auto-attached so context travels with the request.

Validation & outcomes

What we saw in usability and early use

Tested with moderated sessions across RNs, charge, APP, and care coordinators — including new-grad nurses, who are the population most punished by a directory-style list.

Time to identify "who needs me first": ↓ in every session New-grad confidence in first 10 minutes: meaningfully higher Watch thread used by 100% of test participants Handoff verbal time expected to compress (to validate post-launch) Open-loop alerts surfaced patients no one had touched in >2 hrs

The All Patients view didn't replace clinical judgment — it gave nurses a starting point that respected their judgment, and made sure that judgment didn't evaporate at shift change. That was the explicit ask from the unit, and it became the explicit promise of the design.

Reflection

What I'd carry forward

Spend the days on the unit

I would not have arrived at the "rank attention, not rooms" framing from interviews alone. The friction was visible in body language and in the small workarounds — the printed brain sheets, the verbal updates in the hallway — that nurses had stopped naming as friction because they had normalized it.

Continuity is a product, not a meeting

If the only place handoff happens is a verbal huddle, the system has outsourced the hardest part of patient safety to memory. The watch thread and open-loop detection were the design's way of taking that responsibility back on behalf of the team.