Project summary
HSC Technology Group in partnership with The Frank Whiddon Group and the CSIRO eHealth Research Centre has been awarded an ARIIA grant for their ‘Building Excellence Sensor Technology (BEST): Clinical Assessment Response Experience (CARE) project’.
This feasibility study will utilise an IoT sensor platform (Talius) to improve falls prevention of elderly Australians living in a residential aged care facility (RACF). Using a combination of sensors with settings adjusted to individual risk assessments, pre-fall activities are identified quickly through the integration of sensor data on the Talius platform.
The suite of sensors located throughout the RACF will provide prompts (pre-emptive reminders and checks) and alerts (reactive notifications) sent through mobile devices to the direct care staff and the clinical dashboard. Staff are provided with a sensor technology enabled process that aims to identify the early indicators that lead to pre-fall risks based on the validated Falls Risk Assessment Tool (FRAT).
This study will explore the acceptability of the aggregation of autonomous sensor technology for residents, their families and direct care staff as well as assessing organisational impact and identifying any barriers to its adoption. The sensors, used throughout a resident’s room and common areas, will be used in conjunction with observational data and clinical assessments to identify changes in individual health status, mobility, behavioural patterns and pre-fall activities to facilitate timely support and staff intervention.
Project outcomes
Background and Aims
The BEST CARE project explored how innovative sensor technology could proactively monitor and reduce falls in residential aged care. Using Talius' advanced Smart Care™ platform, the study integrated ambient and wearable sensors to track resident activities and identify early signs of increased fall risk, improving safety and responsiveness in care delivery.
What We Did
A feasibility study was conducted at Whiddon’s Arthur Webb Court, involving 24 residents and care staff. Sensors collected detailed data on residents’ activities, sleep patterns, and location. This was complemented by qualitative insights from interviews with residents and staff to assess the technology's acceptability and effectiveness.
Outcomes
A total of 10 residents were interviewed out of 24 total participants in the study. Participants were invited based on their cognitive impairment assessment records at AWC, history of falls, and willingness to be interviewed. The exclusion criteria included residents with cognitive impairment that would limit meaningful participation in an interview. Nine of the residents that participated did not have cognitive impairment, one of the residents with minor cognitive impairment was included as their initial responses to interview consent conversations were reasonably clear. The resident interviews were conducted by two CSIRO researchers (one face-to-face and one remotely via Microsoft Teams) who both had experience in conducting qualitative research and clinical expertise. Interviews were audio recorded using a digital audio recorder and were professionally transcribed. Interview sessions lasted between 10 and 30 minutes, with most of them taking less than 15 minutes.
Most of the residents (n=9) were aware that sensors were installed, although their knowledge of the sensor locations varied. Many were aware of the sensors installed on the ceiling of their room, some were aware of the bathroom sensors, but the majority were not aware of the sensors under the leg of their bed. Majority of the residents had limited understanding of the sensors’ function beyond their physical presence, and reported they didn’t fully understand what the sensors were or what they were recording. Some residents (n=4) had knowledge of the function of the sensors and were curious about what they were recording. These residents had variable levels of understanding. Some reported understanding that the sensors reacted to movement; were activated without them touching anything; were used to track/monitor their movements; were there in case they had a fall, hurt themselves, or needed help; could set off an alarm to nurses; and were aware of the inclusion of an “Artificial Intelligence (AI)” component:
- “Yes there’s one up there and there’s in the bathroom they’ve put it in behind the toilet.”
- “There’s one facing the toilet and occasionally I have noticed that when I walk in the bathroom it will... it flickers without any touching, it just knows that there is something there.”
- “It’s AI, it’s connected to everyone’s rooms, and the space out there, everywhere inside the building, we’re monitored, not in person, but our movements.”
Overall, the residents’ perceptions of the care they received from staff at AWC was very positive. Many of the residents (n=3) went out of their way to comment on how good the care staff in AWC are, and how happy they were with the care that they received. While most residents (n=4) indicated that care staff usually respond promptly to a call bell, some (n=2) mentioned that at times, when the care staff were busy or attending to other residents, that they have had to wait for a response. Most of the residents (n=7) did not observe any changes in nurse behaviour or response times before and after the installation of the sensors and introduction of wearables.
The majority of residents could not recall any instances where nurses had arrived without being called. Two of the residents indicated that they had noticed the care staff were carrying a device/phone that made a noise and alerted them to attend to residents:
- “All the same, they’re always coming in to see – or come and have a chat.”
- “They’ve got a little device that it makes a funny noise if it goes off...It’s a little phone...they have it with them all the time and that I believe is what alerts them and they come.”
Some of the residents (n=3) had noticed changes in the nurses’ behaviours as a result of the sensors, having noticed that nurses were sometimes spontaneously alerted to a change in their activities (without them having to press anything) resulting in nurses coming to check on them and ask them if they were okay. These residents reported noticing responses to various activities including sitting on the edge of the bed, restlessness or moving a lot in bed, going to the bathroom, or spending a longer period in the bathroom. Residents were sometimes unsure whether the sensors had automatically alerted care staff or if they had accidentally pressed the call bell on their wearable. In most of the cases where residents recalled care staff responding spontaneously, the residents felt that the care staff responded when assistance was not required (also see the Section of “Concerns and Suggestions”):
- "It did go off apparently one time and I was moving around and I said how did you know? And they said you must have touched it without me knowing because it was a night that I was restless and it went off, so I don’t know.”
At the time of writing this report, staff interviews are taking place. Formal analysis is yet to take place; however, the response thus far has been overwhelmingly positive regarding towards the potential impact of the technology on the nursing workforce. Preliminary results indicate that staff find the call bell system, as sparked by the sensor system, particularly helpful in preventing falls as staff can respond to calls and alerts more quickly than they could without the call bell. Staff report that this has been particularly beneficial for higher risk residents (that is, those who are identified as a falls risk and need some level of supervision for transfers). There are further interviews to conduct, and formal analysis yet to take place but the commentary thus far is positive from the for impact on the aged care workforce’s perspective and the research team are keen to understand more of these reported themes.
Inferential statistics and machine learning models are being explored to predict fall risk in aged care residents based on ambient and wearable sensor data. Ongoing analyses aims to identify the most influential factors contributing to falls, such as increased walking and room exits, time spent in specific areas of the aged care facility, or disrupted sleep patterns.
Early findings suggest that sensors can be effectively used to monitor pre-fall activities and behaviour patterns. Analysis of bed, door, humidity, vibration, location, and sleep sensors has revealed key indicators that may predict falls. Specifically, variations in call bell activity were observed in the hours leading up to a fall, offering a potential early warning signal. Sleep sensor data also suggest that changes in deep and light sleep patterns may play a role in predicting fall risk.
Finally, preliminary analysis of clinical data a moderate association between post-fall pulse rates and the occurrence of falls. Future analysis is planned to explore and identify additional attributes in the data that are relevant for fall risk prediction, and to confirm these early findings. Future research should also be considered to validate these results in a longitudinal study using a larger participant cohort.
Impact on Aged Care and Workforce
The uptake of sensors and the use of voice AI through handsets was viewed very positively by care staff and helped them prioritise who to attend whilst providing care to another resident. It allowed improved coordination and timely attendance knowing exactly their location (for those wearing their pendants). The subsequent outcome is a more satisfied workforce.
Staff wearing staff duress beacons was optional. More time was needed to assist the staff to understand more fully the features and support for them in the system and will require further investigation.
Resources Developed
This will become available on our website once the findings from the analysis are available.
Next Steps
- Expansion of the technology across additional residential aged care facilities.
- Further research to refine predictive analytics for fall prevention.
- Ongoing advocacy and education to highlight the value of digital care models.
- The BEST CARE project demonstrates that sensor-based monitoring is both practical and valuable, providing a strong foundation for enhancing safety and care quality in aged care facilities.
Key contact for further information:
Mr Graham Russell, Managing Director