SMARTT Critical Care Pathways - (Safe, Machine Assisted, Real Time Transfer) An artificial intelligence based decision support tool to enable safer and more timely critical care transfer

Background: With an approximate cost of £2000 per bed per day for ICU patients and no evidence of consistent guidance between or within sites for when a patient is ready to be discharged back to the ward, there is an urgent need for better use of routine data to support discharge decisions.

We have been working on this problem for several years and are ready to design and integrate a smart dashboard into the ICU.

Aims and Objectives: To refine the discharge algorithm and develop a product, tested across multiple NHS sites, with data on clinical safety, efficacy and with a strategy for commercialisation.

Project Plan and methods used: There are 5 work packages including, governance, core algorithm development, back end development and infrastructure, user-centric design, and testing and evaluation.

Timeline for delivery: 36 months Anticipated impact and dissemination: The use of a dashboard with AI enabled decision support to improve the co-ordination and delivery of timely discharge information will Improve system efficiency by ensuring resource use in the critical care pathway is optimised.

Patients that are suitable for discharge to the ward will leave sooner freeing up more ICU beds for patients that need them.

Improve the safety of ICU patients by ensuring only those ready to be discharged are considered, reducing ICU readmission rates.

Reduce cognitive stress on staff within the ICU and free up time for direct patient care rather than wasting time coordinating and administrating multiple steps in the discharge pathway.

We are uniquely placed to scale this solution across an entire health system with 3 acute hospitals working very closely at a clinical, technical and research level to deliver optimal ICU pathways. From here we will have a strong case to scale further across our regional network and throughout the NHS.