Energising Health: A review of the health and care applications of smart meter data

Smart meters are being rolled out for the benefits they can bring to energy consumers and the energy system in Great Britain, but the data they produce may also be useful in health and care applications. This report examines what research, innovation and commercial activity has been conducted so far in this domain, and what the opportunities and challenges for its development could be. To do this we used a systematised review, and consulted with experts in the field. Our review of the use of smart meter data in health and care applications presents a picture of a field in the early stages of development, but with big ambitions. A small number of research projects have presented evidence of the ability to use digital energy data to recognise activities or usage patterns that could be associated with a variety of health conditions. A number of companies also integrate such data into their health monitoring service offerings alongside other technology. As yet there is no clinical trial evidence of the effectiveness of using digital energy data to improve health outcomes. Potentially recognisable health-relevant features include inactivity (such as through falls), sleep disturbance, memory problems, changes in activity patterns, low activity levels, occupancy and unhealthy living conditions. Much of the small amount of research in this area so far has been applied to Alzheimer’s disease and dementia. Other key targets are a range of mental illnesses (such as depression) and care of people who are vulnerable in some respect. Proposed applications include issuing alerts to carers when unusual activity patterns are recognised, and monitoring of things like the progress of conditions (to inform treatment needs) or of living conditions (such as use of heating or showers). We also raise the possibility of using digital energy data to inform diagnosis and public health, drawing parallels with initiatives in other areas, but there is no evidence at the moment that this will be possible or practical. We identified a number of characteristics of smart meters that may prove particularly beneficial compared to other approaches such as wearable sensors and Internet of Things devices: their near-ubiquity (by 2020), low cost, versatility and provision of historical data. While these benefits lead to high hopes in some quarters to see health applications realised, there are also significant challenges. Considering user acceptance, we do not see many objections specific to smart meters as compared to other approaches. However, assurances around privacy may be important. While Smart Energy GB research shows smart meter privacy concern is low, sharing smart meter data with public services may be more controversial. There are still advances to be made in more accurately recognising specific electrical appliances from dwelling-level data, which would increase the usefulness of smart meter-based systems in identifying activities and reduce false alarms. This could likely be improved if the sampling rate of energy data could be pushed beyond the current 10 second limit. Even if better activity recognition can be achieved, there is much more work to be done in reliably and usefully connecting observable energy use patterns with health conditions. The manner in which smart meter data is stored and shared is tightly governed by regulation. However, using this data in health contexts is likely to involve taking it out of the regulated smart meter infrastructure to share it with third parties. Given the sensitive use to which such data will be put, ensuring good data security and privacy after data has left the currently regulated system should be an important focus for regulators considering its use in health contexts. The level of failure tolerance for health critical uses is also likely to be lower than for standard energy metering applications, with potential implications for how the system is regulated. Questions will also need to be considered about where responsibility lies when systems fail (with potential health consequences)