For most of us if we’re healthy, falling over can lead to a skint knee, and if it happens in public, maybe a bit of embarrassment.

However, for some people, the lasting effects can be much more serious, leading to hospitalisation and often the beginning of a cycle of poor health or even a loss of their independence at home. Injuries caused by falls are a leading cause of hospital admission and death for those aged over 75. On average, there are over 4,000 hospital admissions in Glasgow each year due to a fall.

Glasgow City Health and Social Care Partnership (HSCP) recognises the devastating impact falls can have in its Falls Prevention and Management Strategy, which states that “injuries caused by falls are a leading cause of hospital admission and death for those aged over 75. Fear of falling can result in inactivity, deconditioning, loss of confidence and increased risk of falls in older people. It can also result in reduced social interactions leading to isolation or loneliness.”

The falls strategy highlights the importance of early intervention and prevention, along with data collection, reporting and analysis in reducing the number of falls.

Now, a research project has identified how improved data analysis could help with the prediction and prevention of falls.

The initiative, part of the strategic partnership between Glasgow City HSCP and the University of Strathclyde, working together with Tunstall as telecare providers and the Digital Health and Care Innovation Centre, has identified how improved data analysis could potentially help with the prediction and prevention of falls.

Early intervention and prevention are key, along with effective data collection, reporting and analysis.

The research recommends how better data analytics within telecare equipment – home alarm systems that are typically used by older and vulnerable people to help keep them safe at home and social care – could help prevent and reduce the impact of falls.

The research also identifies that Artificial Intelligence (AI) can help to build more personalised, predictive and proactive models for allocating health resources more efficiently and effectively, at the right time in the right place.

The key challenge was to investigate how to routinely use the vast amount of data collected across the health and social care system to identify or predict people who are at risk of falling, hospitalisation or needing other specific telecare or social care services.

Data from more than 28,000 Glasgow residents who use the telecare system was analysed to understand what is currently collected on whom and about what and what questions need to be answered about telecare users and their usage.

Other questions were what, if any, data is missing and how can it be better captured to be ‘analytics ready’ and how data access, management and sharing for future research and innovation projects can be facilitated.

Telecare devices gather and electronically communicate information to health and social care providers using both ‘passive’ technology such as sensors and wearable devices including pendants and wrist straps and ‘active’ technology where data is purposefully entered into the device by the user.

Marilyn Lennon, Professor of Digital Health and Care at the University of Strathclyde, said: “Telecare devices, systems and users produce vast amounts of data, and we needed to carry out detailed analysis to work out how it can be categorised and used in very pragmatic ways to predict people who are at risk of falling, so that ultimately, preventative steps can be put in place.”

The research also flagged that in some cases the way legacy data was organised made searching and retrieving data challenging and time consuming. Systems and equipment that did not communicate well with each other were also identified as an issue.

Potential inequalities were identified because of IT literacy, and also because internet services and devices were less accessible to some communities.

Another finding was that there were almost half a million reasons recorded for falls in the database because of variations in the way data was manually entered and encoded.

The team recommended better data analysis could help predict service users’ needs and deliver a more proactive service. Integrating systems could also improve the reliability of data and make it easier to update and access, while standardising data organisation and automating tasks could reduce the manual workload. The researchers also recommended steps to improve how data could be used in a more preventative way.

The research findings also encouraged a less risk averse approach to data sharing across organisations, to identify and anticipate who is at risk of falling.

Professor Lennon added: “It is not straightforward to share data but when we do, we get great results. We have the opportunity to share innovative machine learning for the greater good.

“This work has the potential to really make a difference for the better, resulting in timely and early interventions that can ultimately prevent falls.”

Glenda Cook, Planning Manager for the HSCP added: “We knew this was a complex data picture with multiple forms of data, and this study highlights the need to address more efficient data entry, control and storage.

“Improvements can enable a better platform for data analysis, which is necessary to identify alternative approaches to service provision within a time pressured changing technological environment operating within increasingly tight budget parameters.”

Lucille Whitehead from Tunstall, who provides the telecare equipment, said: “These early insights on the data collected from Glasgow City HSCP, and the early analysis by the University of Strathclyde, may help to target care where and when it’s needed most.”
Strategic Partnership

More information is available on the Strategic Partnership between Glasgow City HSCP and the University of Strathclyde, which seeks to drive innovation in health and social care.