18 Nov Artificial Intelligence for Fall Detection and Proactive Care
Machine learning, a component of Artificial Intelligence (AI) is now capable of providing proactive care and enhanced monitoring capabilities tailored to meet the healthcare demands of hospital and aged care groups, without intrusive devices or complex deployments. This type of emerging innovation is supporting the organisations workforce, which is currently suffering from low staff-patient ratios and funding models all which are currently under review as part of Australia’s royal commission into Aged Care (more information available at: https://agedcare.royalcommission.gov.au)
In this post, we review emerging AI based technology and the features available with the intent on providing a surface level understanding of AI, but also highlight the commercial benefits of deploying such innovation with healthcare settings.
What is Machine Learning?
So let’s start with machine learning, which is defined as the ability to teach computers how to perform tasks by learning from data being delivered, rather than being explicitly programmed. In general, the more data the computer provided, the more they can learn which then enables more decision making capabilities. An example of this would be the ability for a computer to watch video streams
to learn when a human is detected within the video frame. When the computers are first shown the video of humans, it would not know what it is seeing until some input is provided, this is known as supervised learning. Data models are built and overtime the computers can start making accurate recognition of humans within a video with a high degree of accuracy.
It is important to note here that machine learning is a sub-component of Artificial Intelligence, which is the broader concept of allowing computers to achieve human like decision making abilities whilst being able to carry out tasks. Machine learning is one possible model that contributes to this overall goal, as is deep learning (which is the next-generation of AI).
As our understanding of how the human mind works progresses, our concept of what constitutes AI is changing. Rather than performing increasingly complex calculations, work in the field of AI concentrated on mimicking our decision making processes and carrying out tasks in ever more human ways is evolving.
Emerging Aged Care and Hospital Innovation
With this basic understanding of artificial intelligence, we can now explore how AI is being used within healthcare. SYNO Global is offering innovative solutions built on AI, with one particular solution (namely the Digital Angel suite) leveraging that of machine learning models to make informed decisions around incidents that have occurred within a particular setting. Now patients and residents of hospitals and aged care facilities alike can utilise this technology to provide improved care for fall and incident analysis, whilst contributing to workforce optimisation to ensure care is provided where and when it is needed most.
Through proactive monitoring and alerting capabilities that are based on machine decision making capabilities, residents and patients can be monitored within personalised settings for a range of incidents that frequently occur that have tangible negative effects on patient / resident well-being, recovery and overall care. Such incidents can include mandatory fall alerting, wandering, circadian rhythm deviation, ensuite timers, room entries and then predictive analysis such as potential bed exit, sit-up in bed preparing for exit and lack of movement. AI can also analyse sound to make decisions around potential incidents occurring within the rooms that could contribute to an incident occurring, or about to occur without the need for intrusive devices and wearables.
Proactive Care – Not Reactive
The problem with traditional sensor technologies, such as that of smart floor devices is the lack of proactive monitoring to prevent potential incidents. With the use of AI – in particular machine learning, patterns of behaviour can be analysed (over a period of time) and the machines can make decisions before the incident occurs. Traditional sensors will alert once the incident has occurred, eg a fall on the ground, but in most cases the mental and physical trauma has already occurred if the patient or resident is not responded to in a satisfactory amount of time. How can a sensor designed to detect a fall on the ground predict the fall is going to happen if it is only designed to respond based on a criteria?
New innovation is now allowing not only reactive incident alerting (once the incident has occurred such as fall), but also proactive incident alerting through machine learning capabilities, an example of this would be a patient or resident who is at high-risk for falling because they struggle to exit their bed. AI can analyse the actions of the person attempting to exit the bed and raise an alert when such is occurring, allowing staff to act and provide immediate support before the fall occurs. This example is then carried across to routing deviation, that is learning the behaviours of the individual and then raising an alarm should something seem out the norm.
AI in Commercial Settings
The challenge with deploying AI within commercial healthcare environments is the requirement for individualisation and customised data – as in the ability to provide customised care settings on a per resident or patient basis. With individualisation comes different forms of incidents awareness and different data sets to provide proactive care, however a software platform that has been designed for scalability and feature roll-outs will enable personalised care plans based on each resident or patient. An example would be a low fall risk patient vs. a high-risk patient where a preventative incident response to a bed exit were to be applied to both individuals (without customisation), staff would be responding to both incidents vs. a single response which results in directed care to the individual who needs it most. With individualisation, residents or patients can have proactive care provided based on their individualised care plans, which has an immediate positive impact on staff workflow.
Some interesting statistics from Aged Care organisations adopting AI proactive care and incident response see’s a 50% decrease in patient falls within the 6 months of operation, 65% reduction in emergency medical services and hospital visits and 71% less intrusive observations though proactive 24/7 fall and incident monitoring. Another major benefit is the notable change in staff workflow, this type of innovation allows for more focused care to provide support where and when it’s needed most whilst reducing intrusive behaviours.