AI and hospital operations
Bed management and patient flow automation: what to automate first
Patient flow automation means using live data from hospital systems to keep bed state, discharge tasks and delays visible and moving, so staff spend less time chasing. NHS England's model discharge pathway asks hospitals to predict and manage bed demand daily, start discharge tasks early and track them, record a reason and owner for every delay, and it says AI-enabled tools can help prepare discharge documentation where clinically appropriate and governed.
What NHS England asks hospitals to do
The model discharge pathway (July 2026) sets out, among other things:
- Predict and actively manage bed demand and capacity every day, with a consistent number of planned discharges, and share the data with partners.
- Work in parallel, not sequentially, across medical, nursing, therapy, pharmacy, discharge co-ordination, social care and community partners.
- Identify and track pharmacy (TTOs), therapy, equipment, documentation and transport needs early, and plan known dependencies such as tests in advance.
- Make criteria-led discharge the default, with daily consultant-led board and ward rounds seven days a week.
- When a discharge doesn't happen, record the specific reason, a named owner and actions with timescales, and review the delay data at senior level.
- Record the discharge ready date, pathway and reasons for delay consistently in the electronic patient record.
The numbers NHS England tracks
For now, the pathway asks trusts to use two operational metrics, visible from ward to board: the number of discharges required and delivered each day by pathway (0 to 3), and the number of patients who no longer need a hospital or community bed and are ready for discharge, by pathway. NHS England's UEC acute patient flow guide adds measures such as the proportion of people discharged on their discharge ready date and the average length of delay, and reports that delayed patients in January 2025 waited 6.1 days on average, an estimated 324,000 bed days lost.
Where automation helps
- Live bed state from the ADT feed (admissions, transfers, discharges and their cancellations), so the board doesn't depend on phone calls.
- Bed turnaround: when a patient leaves, request cleaning, then offer the bed, with escalation if a step stalls.
- Discharge task tracking: TTOs, transport, equipment and documents as tasks with owners and due times, started early.
- Drafting: discharge summaries and letters drafted from the record for a clinician to check and sign.
- Delay capture: prompt for a reason and owner when a planned discharge slips, and roll it up for the daily review.
What stays human: the clinical decision that someone is ready to go, the discharge destination agreed with the patient and family, and anything that needs judgement about risk.
What it does not mean
- Automation doesn't create capacity. It makes delays and tasks visible sooner; the beds, staff and community services still have to exist.
- A discharge event isn't a free bed. The bed still has to be cleaned and allocated.
- A dashboard isn't the fix. NHS England asks for named owners and actions on every delay, not just reporting.
Where Thalix fits
The Thalix demo is built around this. HL7 v2 ADT from a simulated hospital fills Ward 4's board; the Discharge workflow runs the TTO (doctor), transport (ward clerk) and a drafted summary in parallel, a doctor signs the summary, and when the nurse confirms the patient has left the bed goes to cleaning. That starts Bed turnaround: housekeeping confirms the bed is clean, which frees it, and after 20 minutes the task moves to the nurses as urgent. Workflows run only tier-1 actions (bed states, tasks) by themselves. Workflows act in their units without ever seeing names, and every step has a history. Boards show bed maps, tables and timelines resolved as the viewer, so a porter and a nurse see the same ward differently. It runs on synthetic patients today.
Thalix is at pilot stage. Demos run on synthetic data, and a hospital deployment needs its own clinical safety case, information governance approval and security review. We don't claim any certification or assessment result the product doesn't have.
Questions people ask
What metrics does NHS England ask trusts to use for discharge?
The model discharge pathway asks trusts to use two operational metrics: discharges required and delivered each day by pathway 0 to 3, and the number of patients ready for discharge who no longer need a bedded setting, by pathway.
Can AI help with hospital discharge?
NHS England's model discharge pathway says digital solutions, including automation and AI-enabled tools, can support timely preparation of discharge documentation and reduce administrative burden, where clinically appropriate and appropriately governed.
How long are delayed discharges in England?
NHS England's UEC acute patient flow guide reports an average delay of 6.1 days for delayed patients in January 2025, an estimated 324,000 bed days lost.
Sources
Each source below is an official page, read and checked in October 2026. Where a page and this guide differ, the official page wins.
- Model discharge pathway NHS England · checked October 2026
- Urgent and emergency care acute patient flow NHS England · checked October 2026
- HL7 table 0003: event type HL7 Terminology · checked October 2026
General information only, written for hospital digital, IT, operations and clinical safety teams. It isn't legal, regulatory, information governance or clinical safety advice, and it doesn't replace your own organisation's processes or the official guidance linked above.