Smart meters and associated technologies are being used as part of a ‘ground-breaking’ new trial to alert carers to possible health-related incidents faced by elderly and vulnerable people living alone in their homes.

The University of Edinburgh’s School of Informatics, Blackwood Homes and Care and government innovation agency The Data Lab have teamed up to create the Smart Meters for Independent Living (SMILE) project.

The project will see the consortium develop and test machine learning and artificial intelligence (AI) methods to analyse energy usage data from consenting residents’ smart meters, creating a view of their daily routines and spotting unusual changes in behaviour which could cause concern.

Blackwood Homes and Care provides a range of accessible housing for people with disabilities and older people. The trial began in November 2019 and, despite some delays caused by the Covid-19 pandemic, is currently analysing energy usage data in several homes across Scotland. It is hoped the system will be tested in up to five homes.

Individuals and their loved ones or carers can set specific ‘rules’ for the system, telling it which changes in routine are a cause for concern, such as the duration of a shower being longer than usual, or a change to normal cooking schedules, which could indicate that an incident has occurred.

Machine learning algorithms use energy usage patterns to identify the timing of people’s relevant activities in the home, looking for changes that should be flagged up. The system will then alert the individual, their loved one or carer, enabling a decision on the best course of action to take.

A spokesperson from The Data Lab said: “Smart meters record whole house energy data, this is what you see on your electricity bill. There is a readily available device called a consumer access device that looks just like the smart meter displays that the energy companies provide the customer.  The smart meter sends data to the device and then the device sends energy usage data to the University.

“The team at the University then take the whole house energy data and apply a computer algorithm to disaggregate the whole house energy usage allowing for data to be collected on the energy consumption of several domestic appliances such as an electric shower, a kettle and microwave.

“Through this, we can monitor the use of those devices and identify any unusual disparities according to specific rules set by the user, signalling that that they might need help.”

The ambition is that the new predictive digital technology will provide an additional service to complement the traditional proactive push button personal alarm worn by residents – particularly aiding people with dementia and those who may be confused, may forget or be unable to activate their current alarm.

The technology also has the potential to be used as a decision support tool, meaning that if it detects a resident getting up frequently during the night, health and care professionals can review whether they need changes in their support.

The spokesperson added: “This technology is ideal for a person who cannot raise an alert themselves as the initial set up of the system can be carried out by a carer. Once the system is set up, which is very easy to do, then if an incident occurs the impacted user does not need to interact with the system at all, and an alert will automatically be generated and sent to the carer or call centre.

“The system is designed for anyone who lives on their own but is of particular value for users who may have dementia, or for someone who is potentially concussed or confused following a fall, exceeding current systems in place at the moment – which require individuals to be able to press a button to send an alert for help.”

Findings of the trial are expected to be published in autumn 2021. The project is also supported by CareBuilder, Hildebrand, Mydex CIC & Smart Energy GB.

Gillian Docherty, CEO of The Data Lab, said: “This project has the potential to shape the way we view machine learning and AI in social care settings, by empowering individuals to go about their daily routines without worry and only receive carer intervention when necessary.

“Scotland has an aging population, and in the next few decades we need to find new ways to deliver the best possible social care against a backdrop of stretched resources and falling carer numbers. Machine learning and AI can be a non-invasive way to do this and will also encourage greater personalisation of care based on an individuals’ data.

“The SMILE project is funded as part of The Data Lab Collaborative Innovation programme and further strengthens the relationship between Blackwood Homes and The Data Lab, cementing the relationship for further support in terms of skills, network access and external funding support in the years ahead.

“We’re proud to be involved in such a forward-thinking project and look forward to receiving the initial findings soon. It is another fantastic example of data being used as a force for good.”

Dr Lynda Webb from the School of Informatics at The University of Edinburgh said: “It is very exciting to be working collaboratively with Blackwood Homes and the industry partners on this project. It provides an opportunity to apply the machine learning outputs from our previous EPSRC (Engineering and Physical Sciences Research Council) research project, IDEAL, in a new real world setting for social good. The fact that we are also co-designing the service with Blackwood customers means we can take forward the research in a way that is adapted to people’s true needs.”

Colin Foskett, Head of Innovation at Blackwood Homes, said: “At Blackwood we are always looking for ways of enabling our customers to live more independently. The UK smart meter rollout programme presents an opportunity to use energy usage data for good. If we can prove the principle of the technology with this project, then we have an opportunity to provide a safety net for vulnerable people, to identify patterns of decline and provide early intervention, potentially saving lives and reducing hospital admissions.”

Jane Wilson, Chief Operating Officer at Hildebrand, said: “This project is a powerful demonstration of the potential value of smart meter data in helping to support the vulnerable; in my own family we used a clamp-based solution 15 years ago to monitor my father – this solution is scalable and has significant potential to provide valuable, actionable information to remote carers.”

David Alexander, CEO at Mydex CIC, said: “This is another example of the power of personal data being managed and controlled by the citizen to feed into new approaches to health and care management and drive innovation in remote monitoring and support.”

According to The Data Lab, the project at this stage demonstrates the feasibility of this methodology, and a small co-creation with users has identified initial areas to develop. With the support of further funding, the project can then be rolled out on a larger trial with the final vision being for the service to be scaled up and available to users across the UK.