Two statisticians from the School of Mathematical Sciences have created a new approach to monitoring activity changes in older adults to improve their health. The problem of health and care of people is being revolutionised. Disease prevention and health improvement from home are important components of that revolution. Dr Israel Martinez-Hernandez and Professor Rebecca Killick have been working on this complex issue using statistical methods to detect changes within the daily profiles of human activities. There are many different activities to consider within and across a day for an individual. By tracking changes and noticing them, this technology can see where something has changed for that person as changes can be indicators of potential health problems or recovery following acute incidents. However, due to a person’s daily pattern, changes will be observed throughout each day, with, for example, an increase of activity around meal times and lower activity during the night, which makes it challenging to quantify the changes that are not part of a routine.

Their approach combines two statistical methods in a new way to observe data continuously over time, looking at each day and its pattern as a whole. This makes it easier to identify changes and differences in a way that has not been achieved before. It can then identify changes in comparison to the average for that person’s routine. For an older adult who lives alone, this could alert a family member to potential issues in the home. To do this in the past would have taken a huge amount of computing power and be very slow, but the new method reduces compute time and gives a quicker and more accurate result.

The new method proposed can scale up from small to large datasets. This allows monitoring not just of a single person, but of a household and communities. Such research brings a multitude of benefits. To bring these benefits to life, the researchers are working with Howz, a company focusing on technology enabled care. The proposed model is trained with data collected using smart plugs, movement and door sensors. Howz have the capability to trial and utilise these methods in their technological pipeline to show the benefits of this approach at scale. It allows for visualizing and interpreting the results, changes, and trends over time, both at an individual and group level, allowing the detection of potential health decline. It has the potential to make a difference to individuals very quickly.

Dr Israel Martinez-Hernandez said: “I love interdisciplinary work, and my research is real-problem based. At this specific project, it was great working with Professor Rebecca Killick and representatives of Howz. My vision is to develop new statistical models that can model large datasets with a low computational cost using a new paradigm of data analysis. If colleagues have large and complex datasets, we will be happy to collaborate on analysing the data or developing new methods/projects.”

Jonathan Burr, founder and CTO of Howz said “Howz is delighted to deepen our collaborative relationship with the University of Lancaster, bringing new change detection methods to benefit older adults. By integrating these advanced analytics into our technology, we’re seeing significant benefits for users, including reductions in hospital admissions, lower rates of care home placements and faster assessments”.

This research is funded from a grant by the Engineering and Physical Sciences Research Council titled “Quantum Imaging for Monitoring of Wellbeing & Disease in Communities”.

To see the full paper, click here