01 the business
Teaching hospital
800 beds, 30 inpatient wards
Patients who are medically ready wait for medicines, transport, equipment or care packages. Each task belongs to a different team, and nobody sees the whole picture. Beds stay full while the emergency department queues.
02 before
The problem
Patients who are medically ready wait in beds for medicines, transport, equipment or a care package. The tasks sit with different teams, and nobody has one view of what is blocking each discharge. Beds stay full while the emergency department queues.
- Delays are found at the bedside, late in the day.
- Ward staff phone other teams to chase tasks.
- Managers lack reliable data on delay reasons.
03 the build
What the AI does
- Step 1: Reads expected discharge dates and ward round notes from the record each morning.
- Step 2: Builds a checklist per patient: discharge summary, medicines to take home, transport, equipment, care package and follow-up.
- Step 3: Checks each item's status in pharmacy, transport and social care systems where they connect, and asks owners in Teams where they do not.
- Step 4: Shows a ward board of who can go today and what is holding up the rest.
- Step 5: Reports the most common delay reasons each week for the patient flow team.
paused: waiting for a person to approve
How it was put together
- Read expected discharge dates and ward round notes each morning.
- Built a checklist per patient: summary, medicines, transport, equipment, care package and follow-up.
- Checked status in pharmacy and transport systems, and asked owners in Teams where systems did not connect.
- Showed a ward board and a weekly report of delay reasons.
04 human in the loop
Where a person stays in charge
The clinical team decides who is ready to go home. The tracker only shows and chases outstanding tasks; it never changes a discharge decision.
- Clinicians decide who is ready to go home.
- The tracker only shows and chases tasks; it never changes a discharge decision.
- Access follows existing ward permissions.
05 rollout
From kickoff to live in 8–10 weeks
- Weeks 1–3Agree checklist items with ward managers and the patient flow team.
- Weeks 4–6Pilot on four wards.
- Weeks 7–10Roll out across inpatient wards.
06 after
What changes
Patients who are ready go home earlier in the day, and the ward sees blockers in time to act on them. Managers see where delays really come from.
Assume 120 discharges a day across the hospital, and each leaving 3 hours earlier on average.
120 × 3 hours = 360 bed-hours a day, the equivalent of about 15 beds freed every day.
What to measure
- Share of discharges before midday.
- Hours between medically ready and leaving.
- Delay reasons by category.
- Emergency department waits for a bed.
07 evidence
What the research says
Around 1 in 10 patients is harmed in health care, and more than half of that harm is preventable.
World Health Organization, Patient safety fact sheet (opens in a new tab) · 2023
The global nursing shortage was 5.8 million in 2023 and is projected to be 4.1 million in 2030.
World Health Organization, State of the World's Nursing 2025 (opens in a new tab) · 2025
08 lessons
Mistakes to avoid
- Another board nobody updates. Pull status from systems automatically.
- Blaming teams. Use the data to fix processes.
- Ignoring social care. Many delays sit outside the hospital.
09 questions
Common questions
- Does it decide who can go home?
- No. Clinicians decide; the tracker only shows and chases what is still outstanding.
- Can it include community and social care?
- Yes, where those partners share status, otherwise as tasks owned by the discharge team.