Data Analytics for Health-Care Profiling using Smart Meters

Through the completion of this research, we will demonstrate that a small simple change, using existing infrastructure technologies, can have a large impact with significant benefits for society and academia. As such, the premise of this research is to investigate whether data analysis of smart meter electricity readings can be used to support social care that meets a person's individual needs, maximises independence and promotes a sense of security for those living alone.

By the end of 2020 it is expected that 55% of global electricity meters will be smart meters. Within the UK, Energy suppliers and the government are funding the cost of the smart meter roll out and ongoing maintenance. We envision that by investigating advanced machine learning and load disaggregation techniques of this highly accurate sensing network, detailed habits of an individual's interactions with electrical devices can be mathematically modelled.

This research is needed to support and enable a larger number of people to remain independent whilst living with long-term health conditions, such as Alzheimer's. For example, in the UK, around one in five adults are registered disabled and more than one million of those currently live alone. These conditions place significant demands on healthcare services globally.

Existing monitoring services (such as motion sensors, cameras, fall detectors and communication hubs, wearable body networks) are intrusive, expensive and are met with patient resistance. Additionally, current technical solutions are tailored to a specific application and do not meet the ongoing changing requirements of a patient; whereas our approach would require minimal installation, and builds on the smart meter infrastructure, without the need for user interaction.

Analytics are tailored to an individual's health condition for monitoring, early intervention, detection and prediction of self-limiting conditions. If abnormal behaviour is detected, an alert could then be sent to a carer or family member. Specifically, the research will allow us to devise a system that can detect when an Alzheimer's patient has left an oven on or remained awake at night.

The technology creates a personalised profile of the user's behaviour at home. Our system is a disruptive technological solution within tele-health/tele-medicine. Uniquely, there is no requirement for the deployment of sensors around the home. We are employing an existing highly advanced sensor system, which is readily deployed, to provide peace of mind and remote patient care, compared to current technologies available on the market today.

This research complements other recently funded EPSRC projects conducted on smart meter analytics. However, this research is unique in that it is the first project to propose using the smart grid for health analytics, as other projects are concerned primarily with load balancing and energy reduction practices.

The successful completion of the proposed project will involve research in the areas of computer science, specifically big data analytics, and healthcare. Our collaborators from Mersey Care NHS Trust are supporting the research by providing medical advice on Alzheimer's disease profiling and providing patient trials.

Who will benefit: This is a disruptive research proposal, offering significant benefits to patients, carers and the NHS through the remote monitoring of medical conditions and social care management. Our initial work in this area has generated external interest from the public and the commercial sector. For example, the Westminster Sustainable Business Forum invited us to present our work on the use of smart meters in healthcare provisioning in July 2016. We also received £35,000 funding from the Innovation to Commercialisation of University Research programme (by Innovate UK) to undertake a market validation. This confirmed our initial ideas relating to the beneficiaries and the impact the technology could achieve. As such, we envision that this research project will directly impact local NHS trusts, housing associations, councils, local authorities and policy makers. Most importantly, it will directly benefit the end-user, their family, friends and carers, providing peace of mind in their own home.
How they will benefit: We envision that healthcare and social care providers will rely heavily on our patient monitoring system to ensure patient safety and welfare in the future, particularly for individuals living alone with Alzheimer's disease. This research may also directly benefit various policymakers, who are setting guidelines for the prioritisation of the smart meter rollout. In particular, we will highlight the importance of a rapid smart meter deployment for vulnerable people in our society. These groups must be given priority and seamless access to smart meters to ensure that state-of-the-art assistive healthcare monitoring services are implemented quickly. Therefore, appropriate policies and guidelines must be established between energy regulators, energy providers, charities and health and social care providers. This is important to ensure a seamless framework is developed between the various parties.
Planned Actions: In addition to the conference attendance and dissemination plans, as outlined in the case for support and pathways to impact, to distribute our results we plan a number of complementary actions. The aim of these actions is to disseminate the work and show proof of concept. Specifically we will exploit existing links (for example with Mersey Care NHS Foundation Trust and Liverpool Council) to attend and host seminar days/sessions to engage with wider industry targets and government organisations.
Track record: The PI is a member of the LJMU PROTECT research centre and the Applied Computing research group, which are comprised of multidisciplinary teams that develop novel approaches and applications in the areas of data analytic and data science, knowledge mining and machine learning. Both groups focus on the application of computing to real-world problems through research and industrial projects. The groups have an excellent track record in developing software and hardware solutions for a variety of domains, from commercial and industrial solutions to experimental evaluations. Generally, their projects have the aim of facilitating personalisation and interpretation of data for specific user needs and to minimise the cognitive burden of information overload. The group attracts on-going research funding from various sources in the UK and the European Union, including from industry, research councils and charitable foundations. For example, their most recent projects have included a one-million pound Innovate UK project on the use of machine learning to investigate to predict the onset of nocturnal enuresis; a (£) 52k project (funded by Al-Khawarizmi International College) on Medical Data Image Compression for Mobile Devices and a Knowledge Transfer Partnership grant on Data Analytics in the Tourist Industry to the value of (£) 60k.