When older people are in hospital, there can be delays in discharging them back to their homes. This is because an assessment needs to take place to check that they are able to look after themselves. This is a major problem for the health and social care sector. Delays to hospital discharges cost the NHS £820m annually. Longer stays in hospital are associated with increased risk of infection, low mood, and reduced motivation. Difficulty putting a package of support in place is the cause of nearly a third of all delays. We want to use our conversation-based AI-technology to support this process. Our project will introduce non-conventional methods of collecting and analysing data (gait, voice and conversation) which provides health insights, in addition to established clinical data sources (like blood pressure and heart rate) to help assess overall wellbeing of patients in the home environment following a discharge. We already have an AI-led product that supports older people. This project will develop this to support what is known as Discharge to Assess (D2A). This will reduce delays, improve patient care, and save money. By developing MiiCare's existing product we will be able to: * Automate and schedule patient care assessments * Use conversations between the patient and Monica, our voice assistant, to evaluate mental health status * Step in if symptoms of delirium/hallucinations/delusions are detected * Analyse changes in voice patterns to identify possible symptoms of RTIs * Apply clinical gait analysis on the patient's footstep recordings to evaluate physical fitness, gait abnormality, and risk of fall. * Build a comprehensive digital health record depicting the overall patient wellbeing from conversations, gait and vitals data captured in their home. * Reduce number of patients needing to be discharged to a 24-hour residential care setting * Provide 24/7 interactive virtual support to users that can help reassure them and their families. Our approach will put users at the heart of the design process. We will use their knowledge, experience, and expertise to build something that really works for them. We will also work with families, care providers, and doctors to make sure that what we design meets their needs too. By the end of the project we will have designed, built and tested a product that is ready for a wider scale rollout in community care settings. This will increase the number of patients being discharged to their own home and save money for the NHS and Local Authorities.

Diabetes affects 4.7 million people in the UK, costing the NHS at least £10 billion a year, equivalent to 10% of its entire budget (Diabetes UK, 2020). 37% of all diabetes patients aged over 70 (1.7m people). There are lots of digital apps that can be used to help prevent and manage the condition, which many older people find helpful. However, these apps can only be used on a smartphone. We know from research that 60% of those over 65 do not use a smartphone. Therefore, we need to find ways to put the appropriate risk stratification measures in place to help both prevent diabetes in the first place and manage it if it does develop. That is the focus of our project. MiiCare, the lead partner for this project, already has a voice-controlled intelligent personal assistant (VIPA), a little like Alexa or Google (without the background recording), which currently provides support to older people living with dementia. By placing one of our devices in the home, it learns about individuals' activities of daily living and overall health using Artificial Intelligence and using the appropriate health therapeutics care plan and nudges, helps older people to stay well. Users don't need a smartphone, or to know about technology. Our device comes preconfigured. It just needs to be plugged in and it can start to help. For our project we are teaming up with DDM, a leading provider of digital health apps. DDM already has platforms that specifically help prevent and manage diabetes. We plan to work together to adapt one of DDM's existing, clinically proven programmes, and incorporate it into our voice-based assistant. And because people from some communities are up to 6x more likely to develop diabetes, in the longer term we also plan to make it available in community languages such as Panjabi, and Urdu. What this means is that older people will be able to access support tailored to them in the comfort of their home via a friendly, intuitive, voice-based assistant. This can help them maintain a healthy lifestyle and prevent the development of diabetes. This benefits them and their families, and it also benefits the NHS -- reducing costs of treatment and pressure on GPs.

Diabetes affects 4.7 million people in the UK, costing the NHS at least £10 billion a year, equivalent to 10% of its entire budget (Diabetes UK, 2020). 37% of all diabetes patients aged over 70 (1.7m people). There are lots of digital apps that can be used to help prevent and manage the condition, which many older people find helpful. However, these apps can only be used on a smartphone. We know from research that 60% of those over 65 do not use a smartphone. Therefore, we need to find ways to put the appropriate risk stratification measures in place to help both prevent diabetes in the first place and manage it if it does develop. That is the focus of our project. MiiCare, the lead partner for this project, already has a voice-controlled intelligent personal assistant (VIPA), a little like Alexa or Google (without the background recording), which currently provides support to older people living with dementia. By placing one of our devices in the home, it learns about individuals' activities of daily living and overall health using Artificial Intelligence and using the appropriate health therapeutics care plan and nudges, helps older people to stay well. Users don't need a smartphone, or to know about technology. Our device comes preconfigured. It just needs to be plugged in and it can start to help. For our project we are teaming up with DDM, a leading provider of digital health apps. DDM already has platforms that specifically help prevent and manage diabetes. We plan to work together to adapt one of DDM's existing, clinically proven programmes, and incorporate it into our voice-based assistant. And because people from some communities are up to 6x more likely to develop diabetes, in the longer term we also plan to make it available in community languages such as Panjabi, and Urdu. What this means is that older people will be able to access support tailored to them in the comfort of their home via a friendly, intuitive, voice-based assistant. This can help them maintain a healthy lifestyle and prevent the development of diabetes. This benefits them and their families, and it also benefits the NHS -- reducing costs of treatment and pressure on GPs.

miiCARE Ltd aims to develop a non-wearable assisted living healthcare solution through the ADAPTIVE project, incorporating novel AI-based machine learning capability to learn about the acoustic characteristics of elderly subjects' footsteps in order to predict the likelihood of falls, the changes in postures relating to other health conditions or the progression of cognitive issues such as Dementia. The solution is specifically developed to address an unmet need among People Living With Dementia (PLWD). Existing technologies for ambient assisted living are not designed specifically for dementia and require users with dementia to adapt to the technology by wearing devices that indicate falls (after they have happened). ADAPTIVE extends miiCARE's innovative IoT solution (miiCUBE) and uses emerging techniques in acoustic events detection to predict fall risks among PLWD, enabling preventative measures to be taken early and prevent escalation. This is the first time such technology is being used to improve the quality of life of PLWD within their own homes. miiCUBE's feasibility has been proven through trials with elderly households (not with dementia) across Kent, demonstrating several successful use cases. The ADAPTIVE project will give miiCUBE audio monitoring capabilities by developing novel AI to specifically assess fall risk and monitor other distressful behaviours in PLWD that would help improve the delivery of dementia care as a whole. Initial trials will be conducted on a mix of 50 elderly people with and without dementia in the Bristol communities. Following initial trials and iterations, ADAPTIVE will be deployed in the Dover Harmonia Village, a controlled environment with 30 residents with dementia, which aims to make use of telecare solutions for residents' benefit. East Kent Hospital NHS Trust will manage the trial in a multi person environment at the Harmonia Village and provide clinical input, yielding further data to train our AI algorithms. The qualitative and quantitative data analysis will be undertaken by the University of Kent to determine clinical and functional outcomes of the project. ADAPTIVE will enable people with dementia to remain in the comfortable, safe environment of their own homes, whilst giving peace of mind to carers and families and improving safety in care environments by reducing unnecessary alarm calls and supporting a new care model respectively. It aims to reduce the £26.3bn cost of Dementia to the UK economy (£8.8bn of this is combined state social and health care costs), as well as cost of dementia-related falls to the NHS.