27.03.19
Utilising smart meters in assisted living
Source: NHE March/April 2019
Dr Sudip Sikdar, consultant old-age psychiatrist, and Pauline Parker, head of research at Mersey Care NHS Foundation Trust; alongside Dr Carl Chalmers, senior lecturer in computer science, and Dr Paul Fergus, reader in machine learning at Liverpool John Moores University (LJMU), outline the innovative work being carried out by the Mersey Care NHS FT in collaboration with LJMU to utilise smart meters to monitor those living at home with dementia.
In the UK, the number of people living with progressive neurodegenerative disorders, such as dementia, is increasing. Supporting their ongoing needs places significant strain on national health and social care resources. Providing 24-hour monitoring for patients is a significant challenge, which is set to increase further due to the aging population within the UK.
A key philosophy in dementia care is supporting people to live well with dementia by promoting independence and enabling them to continue living at home for as long as possible. Over 80% of patients prefer to stay in their own home in the later years of life, and evidence shows that moving people into residential care can hasten the progression of dementia. However, allowing patients to live at home safely is often quite challenging as there are limitations in current care provision, and the responsibility of caring often falls to informal carers.
The impact of carer burnout is also well documented in literature and, given that community care provision relies on feedback from informal carers to raise concerns, those people who do not have a social support network are at increased risk of adverse events. This often means that opportunities for early intervention in dementia care are missed.
Currently, there are no technological solutions capable of monitoring the progression of dementia 24-hours a day, seven days a week. Though telehealth solutions are available, they are often expensive, intrusive, and increase the cost of overall standard care plans by 10%, with little or no benefit to the patient. Where such technologies are being used, they are incapable of being personalised to the individual. This significantly limits the effectiveness of most current solutions. Consequently, large-scale usage within NHS trusts, councils, and social services in the UK is not feasible.
Until recently, this has remained a significant problem. However, with the introduction of the smart metering infrastructure, new and exciting opportunities for a variety of emerging applications (such as health and social care) are possible. Smart meters can continually monitor household electricity consumption, and our technology utilises this infrastructure to capture the detailed habits of an individual’s interactions with electrical devices, identifying anomalies or changes in a person’s routine to facilitate early intervention practice (EIP) by monitoring deviations in behaviour that correlates with disease progression.
This is achieved using artificial intelligence, techniques that model patterns in electricity usage, and a person’s day-to-day routines at home. The proposed solution is able to identify when individual appliances are used in the home and model both normal and abnormal behaviour. The current system can identify kettle, toaster, microwave, cooker, and washing machine usage. Interactions with these devices are used to help detect significant variations in activities of daily living, and can be used to safeguard the patients.
An initial six-month clinical trial has been completed in a partnership between LJMU and Mersey Care NHS FT. Energy readings were monitored every 10 seconds and used to identify interactions with appliances in multiple homes. By detecting when appliances are used (how often and when), routine behaviour was established. This allowed us to identify any abnormal behaviour when the interaction pattern changed and can be used to raise alerts when required. The results demonstrate that the system can monitor and support patients in an unobtrusive and personalised manner.
The research team is now planning to evaluate the technology using 50 patients with mild to moderate dementia, living in their own homes over a 2.5-year longitudinal study. If the study proves successful, a much larger case control clinical trial will be undertaken to determine to effectiveness of the technology as a clinical decision support system.
Enjoying NHE? Subscribe here to receive our weekly news updates or click here to receive a copy of the magazine!