Comment: The Missing Link

Using health data to predict early fall indicators and prevent hospital admissions

In this article, Adrian Smales, a MBA PhD research student at Edinburgh Napier University, discusses how untapped personal and mandatory health data can be used to predict fall risk factors earlier in frail and elderly individuals

Adrian Smales

According to the World Health Organisation, 28-35% of older people aged 65 and over fall each year globally, and a fall is considered the main cause of serious injury, injury-related disability and death in older people.

Progressive healthcare organisations across the country are looking for transformative technologies that enhance health and care services by encouraging early intervention and less reliance on acute medicine

In the UK, the NHS reported that fall-related expenditure was about £4.6m a day, with the annual cost reaching £2.3billion in 2013.

Given the prevalence of falls in older people, we as a nation are challenged with providing high-quality care to frail and elderly individuals under constrains of limited resources, and if it is not addressed and appropriately managed, these healthcare deficits will only become greater in the future.

Prevention rather than cure

Progressive healthcare organisations across the country are looking for transformative technologies that enhance health and care services by encouraging early intervention and less reliance on acute medicine.

In light of this, and the need to manage the increasing deficit; Edinburgh Napier developed a framework in partnership with Polar and CM2000 to merge different types of data as part of the Advanced Risk Modelling for Early Detection (ARMED) project to predict and prevent falls among frail and elderly people in the future.

The team collected information from three different channels:

  • Level 1: Mandatory public-sector data stores obtained from hospitals, medical devices and/or GPs
  • Level 2: Data collected using high-quality sports performance devices from Polar, which are designed for professional athletes. They can record important metrics such as heart rate, sleep pattern and activity levels
  • Level 3: Objective personal data obtained by asking questions, such as ‘how do you feel’ or ‘what is your pain level today’

We use various devices to look at body composition including grip strength, which is a predictive indicator for adverse outcomes, and hydration levels. Frail and elderly people are particularly susceptible to complications as a result of hydration, and it can lead to other secondary symptoms such as delirium.

We also measure muscle mass; the fire that fuels the human body and something that is an important factor to consider when looking at the wellbeing of an older individual.

The project has revealed a number of previously-overlooked indicators that, when measured, could predict risk of falls in the future at a more-preventative stage

By using the metrics generated by the devices, we are able to promote weight gain, but also to determine how much of it is translated into fat and muscle. It is about being focused around rehabilitation for the individual.

We can also look at activity levels during the day and night; resting states and sleep patterns (restful or restless sleep).

Finally we looked at the activities of daily living, and asked qualitative questions about their overall health and wellbeing in addition to measuring their heart rate.

By collating all of this information we can create a baseline for the individual for a personalised healthcare plan, rather than basing decisions on generalised demographic information.

Let’s look at an example. Mr Jackson has a history of falls and, while his existing care package included regular monitoring and multiple visits per week, they were unable to find probable causes.

We provided him with a Polar loop and heart rate monitor, measured his grip strength, weight and quantitative data collected from delivered care visits.

These metrics helped to determine his overall health state.

The data revealed that Mr Jackson sat in his chair for a large proportion of the day. After generating a profile using the Polar device and CM2000 database, we were able to determine that he had repeated sleepless nights due to regular inactivity. He would sit for 16/18 hours per day, and then proceed to have a fall a few days later.

Being able to monitor activity levels was a good predictor for increased risk and prevalence of falls, and allows health workers to make adjustments to improve his overall wellbeing.

Risk indicators

The project has revealed a number of previously-overlooked indicators that, when measured, could predict risk of falls in the future at a more-preventative stage.

Of the individuals we monitored, we found those who were significantly dehydrated, had abnormal weight loss, increased restlessness at night, reduction in muscle mass and strength grip, or reports of generally feeling unwell by the users themselves were the most at risk.

Patient health data is a largely-untapped, yet-powerful tool that could improve wellbeing and independence through personalised models of care

The possibilities to implement the framework to other patient groups are endless - known fallers, those going through a reablement process, anyone on the periphery with low-level telecare equipment, or individuals with mental health or memory problems.

Patient health data is a largely-untapped, yet-powerful tool that could improve wellbeing and independence through personalised models of care.