Every day in England, 14,000 patients remain in hospital despite being medically ready to leave, mainly due to delays in arranging onward care. These delays impact hospital flow, admissions, and patient outcomes. By predicting care needs earlier, AI can help discharge teams act sooner and more effectively.
To ensure the model fits real-world needs, the project team is running five co-design workshops with clinicians and operational staff across Wessex.
These sessions explore:
• Variations in discharge planning across hospitals
• Where AI predictions could best support the process
• How to visualise and communicate the model's output.
The model, originally trained on data at University Hospital Southampton NHS Foundation Trust, is now being validated with data from Portsmouth Hospitals University NHS Trust and enhanced with post-admission information.
A working group, including Health Innovation Wessex (HIW), is guiding the regional roadmap for implementation. This collaboration strengthens partnerships and supports knowledge sharing across NHS trusts.
A high-impact paper will share findings from the clinical sessions, highlighting how AI can improve patient outcomes and system efficiency.
In parallel, HIW and the PROCED-DST team will prepare a major funding application to the NIHR Health and Social Care Delivery Research programme, targeting submission by January 2026.
This initiative aims to deliver scalable, evidence-based AI tools that address real clinical challenges, improving care delivery, equity, and patient experience across the Wessex region.
