How might we 01
Plan an intelligent visit tour
Order the day by patient priority, geolocation and an estimate of how long each visit takes — then support navigation between homes and the preparation of the medical material to bring along.
The context
Home hospitalisation (ospedalizzazione a domicilio) is an alternative to a hospital bed for patients in the acute phase of an illness: doctors and nurses go to the patient's home instead. In Piedmont the service has been running since 1985 at the AO CSST, for patients with geriatric and metabolic bone diseases.
It works, and it runs on paper. Every morning the staff has to decide who visits whom, in what order, with which materials — while the patients' conditions keep changing overnight.
The problem
Interviews and task analysis in the OAD units of Molinette in Turin and Cresla in Novara surfaced three problems that fed each other:
- Visit planning was manual and fragile. Nurses updated the day's visits by hand, doctors had to check and approve them, and changes travelled by phone depending on each patient's status. Daily alignment meetings ate into the morning.
- Information was verbal and hand-written. Notes lived in a paper notebook and in a pad carried on visits, then had to be copied over. Sharing, storing and finding anything was slow, and what existed was scattered.
- Caregivers were left out. The only channel with the patient's family was the telephone, and nobody told them when the staff would actually arrive.
The team's estimates put this at the order of one and a half to three hours a day, per unit, spent planning the visit tour and rewriting paper records — time taken directly from clinical work.
How we framed it
Four questions carried the project. They're worth reading as they were written, because each one is a constraint disguised as an opportunity.
How might we 02
Absorb the unexpected
A patient's complexity changes overnight, and with it the priority of the visit. The plan has to be re-made in minutes, not re-negotiated in a meeting.
How might we 03
Make information reachable
Patient information has to be accessible and up to date for the whole OAD staff — both on a visit and back in the office.
How might we 04
Fit the shift, not the ideal day
Visit tours have to be organised around the real shifts of doctors and nurses, including who is specialised in what.
Process
Understand
Interviews and task analysis
UX researchers ran interviews and task analysis inside the OAD units, producing insights, user requirements and a task analysis. In parallel, best practices from comparable services and platforms were reviewed.
Model
Opportunities and service model
Pain points and opportunities were mapped, then turned into a concept: task flows, user journeys, a detailed service blueprint and a system map.
Design
From wireframes to prototype
Information architecture, hi-fi wireframes, a UI design system, and a click-through prototype iterated on expert feedback.
Two nurses, two days
The journeys were built around two real shifts: Giovanni, the nurse on mornings, and Elena, on afternoons. Mapping their days hour by hour is what turned a generic "coordination problem" into a list of specific moments where a tool could help.
8:00 — arrival at the unit and preparation of the material, checking which events overnight have changed the plan: a death, a return home, an ambulance call, a fall. 8:30 — the tour is defined for the 9-to-12 window, weighing clinical needs, priority, patient location, who is on shift and continuity of care. 9:00 — departure with the paper records, exam results and materials. 9–12 — during the visits, needs and problems emerge, written on a pad. 12:30 — back at the unit, the pad is copied into the visit notebook. 13:00 — handover to the afternoon nurse, by voice.
Read like that, the opportunities are obvious: the handover is a shared record, the pad is a diary updated from the patient's home, the 8:30 negotiation is an algorithm plus a screen you can argue with.
What we designed
01
The tour, calculated
An algorithm proposes the best possible visit tour by crossing staff availability and shifts with patient data: priority, clinical complexity, location and estimated visit duration.
02
A board you can argue with
The proposal is not the decision. On an interactive whiteboard or display, office staff re-arrange teams and visits by drag and drop — together, in the room where the call gets made.
03
One record, many hands
Staff update and share patient information simultaneously — from the office and from a patient's home — with a slice of it exposed to the caregiver, including when the staff is actually arriving.
Decisions worth naming
Decision 01
Propose, don't decide
The algorithm could have produced the schedule and been done with it. It doesn't: it proposes, and the staff re-arranges. In a clinical service the person carrying the responsibility has to keep the last word, and a tool that removes it gets worked around instead of used.
The trade-off: less automation, and a screen that has to make the variables legible enough to argue with.
Decision 02
Three surfaces, not one app
Doctors and nurses in the office work on a desktop and an interactive whiteboard; the same people on a visit work on tablet and mobile; caregivers get their own mobile view. Same data, three different situations, three different designs.
The trade-off: more to design and maintain, against a single responsive interface that would have fitted none of the three properly.
Decision 03
Give the caregiver a window, not a login to everything
The family's problem was not lack of clinical data, it was not knowing when someone would ring the doorbell. The caregiver's view was scoped to arrival, schedule and communication.
The trade-off: it doesn't answer every question a family has — but it answers the one they were phoning about.
Where it ended
The work closed on a click-through prototype with a UI design system, iterated on expert feedback, and delivered with the service blueprint, system map and information architecture behind it. It was presented as the WP4 output of the CANP project in April 2020 — a validated design, not a shipped product, and this page doesn't pretend otherwise.
What I took from it
- In a service that already works, the design problem is rarely the interface. It's the handover, the phone call, the notebook copied twice — the seams between people.
- Mapping two shifts hour by hour was worth more than any feature workshop: the opportunities wrote themselves once the day was on the wall.
- Automation in a clinical setting has a ceiling, and it isn't technical. It's who is accountable when the plan is wrong.
[DA COMPLETARE: una riga sul tuo ruolo specifico dentro il team di tre designer — su cosa hai lavorato tu in particolare. Il resto della pagina è scritto al plurale, che è corretto, ma per un colloquio serve la tua parte.]