D2A Smart Incubator – Version 1

  • Research type

    Research Study

  • Full title

    Discharge-to-Assess Smart Incubator: Leveraging AI-based Algorithms for Optimising Assessments at Home - Version 1

  • IRAS ID

    338269

  • Contact name

    Richard Wong

  • Contact email

    richard.wong3@nhs.net

  • Sponsor organisation

    Willows Health

  • Clinicaltrials.gov Identifier

    10034156, UKRI-Innovate UK application number

  • Duration of Study in the UK

    0 years, 5 months, 8 days

  • Research summary

    This project uses Artificial Intelligence (AI) algorithms to support frail older patients being discharged home from hospital with carers. An AI-augmented approach ontop of existing care offers the potential to detect problems earlier; it might facilitate timely intervention, reducing complications and readmissions. Information from AI may also allow care optimisation, allowing gaps to be addressed but also reducing unnecessary overprovision.
    Delayed discharges cost the NHS £820m annually. Longer hospital stays are associated with infection risks, delirium, cognitive decline, low mood, and loss of physical function; these increase risk of hospital readmission and mortality. The biggest reason for delays (32%) is assessing and establishing a package of support in the home. Hence the ethos has shifted to discharging patients with an initial care package and finalising assessments and care post-discharge (Discharge-to-Assess – D2A). These D2A processes are time-consuming for local authorities and healthcare staff and subject to inter-assessor variability.
    The project explores supporting D2A processes with a voice-based virtual assistant, 'Monica', an existing technology already finding commercial adoption. In developing this further with bespoke AI systems, based on user-centred design processes, the aim is to:
    1. Automate and schedule care assessments through conversations with ‘Monica’ (MiiCare’s voice-enabled conversational AI).
    2. Use conversations between the patient and ‘Monica’ to screen for mental health status, including delirium.
    3. Apply acoustic gait analysis on the patient's footsteps, screening for changes in activity and falls risk.
    4. Build a comprehensive digital health record depicting overall wellbeing from conversations, activity and gait, sleep and night activity, and vital signs, all from data captured ‘in home’ (directly from ‘Monica’ or linked accessories).
    5. Reduce need for patients to be discharged to a 24-hour care setting, by providing 24 hour monitoring at home.
    6. Provide 24/7 interactive virtual support to users that can help reassure them and their families.

  • REC name

    London - Brent Research Ethics Committee

  • REC reference

    24/LO/0074

  • Date of REC Opinion

    17 Apr 2024

  • REC opinion

    Further Information Favourable Opinion