SMARTT Critical Care Pathways V1

• Research type

  Research Study

• Full title

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

• IRAS ID

  297886

• Contact name

  Christopher P Bourdeaux

• Contact email

  christopher.bourdeaux@uhbw.nhs.uk

• Sponsor organisation

  University Hospitals Bristol & weston NHS Foundation Trust

• Duration of Study in the UK

  3 years, 0 months, 1 days

• Research summary

  Research Summary:
  Intensive care doctors make hundreds of difficult decisions every day. Deciding if someone can leave the intensive care unit (ICU) is one of these. Around a quarter of patients stay in ICU for too long or not long enough. If patients stay too long, they have a poor experience and have a longer length of stay in hospital. Delayed ICU discharge also denies the bed for other patients causing delays to the admission of emergency patients or the cancellation of planned surgery. If patients leave ICU too soon, they are at risk of early deterioration on the ward and readmission to ICU. \n\nMonitors, machines and blood tests produce thousands of pieces of information about each patient and this data is collected digitally as part of routine care. In this research we will develop artificial intelligence technology that takes this information and works out when patients are ready to be discharged. \n\nThe tool will assist doctors and nurses to make safer and more efficient decisions about discharge from ICU. \n\nAlongside the development of the artificial intelligence system, we will develop a user interface by asking clinical staff how they would use such technology as part of their normal working day in a series of workshops. This will ensure that the system we develop is useful and used in routine practice. We will also demonstrate that this technology can be deployed in a second hospital.\n\nWe will not be deploying the tool into clinical practice as part of this research. That will be the subject of a further application in the future when we have developed the tool further.\nThe SMARTT Critical Care Pathways Project 5 work packages.\nWP1) Governance\nWP2) Core algorithm development\nWP3) Back-end development and infrastructure\nWP4) User centric design\nWP5) Evaluation\nThis application covers the work undertaken in (WP2) and(WP4).\n

  Summary of Results:
  We have developed and tested an AI-driven decision support system to optimise ICU patient discharge. We successfully refined an AI algorithm, integrated it into NHS IT infrastructure, and evaluated its safety and efficacy. We have secured follow-on commercialisation funding, from two separate schemes, which allows us to develop SMARTT into 2026, and we are currently engaging with key stakeholders to develop our commercialisation strategy.
  The main output of the project is the SMARTT technology platform. This is a decision support tool in the form of an interactive dashboard. It incorporates machine learning predictions about which patients are clinically ready to be discharged from ICU. Once a patient has been declared ready for discharge, it then assists in the coordination of tasks that need to be completed by the multi-disciplinary team to de-escalate that patient’s care. The front-end of this software has been co-designed with clinical users via an agile software development process, and the entire system has been designed with input from our PPI group. The software is modular and highly adaptable, designed to be deployed on top of different data systems and to integrate new machine learning models

• REC name

  Wales REC 4

• REC reference

  21/WA/0340

• Date of REC Opinion

  28 Oct 2021

• REC opinion

  Further Information Favourable Opinion