Autonomous fall risk assessment in Australian Residential Aged Care Facilities using passive sensors: A feasibility study
David Silvera-Tawil, Jane Li, Liesel Higgins, Deepa Prabhu, Jennifer Hewitt, Katie Packer, Wei Lu, Maggie Haertsch, Marlien Varnfield · Experimental Gerontology · 2025
Falls in residential aged care home (RAC) remain a critical issue in Australia, contributing to diminished quality of life and increased morbidity among older adults. This study investigates the feasibility of passive sensor technologies to proactively identify behavioural changes, such as reduced mobility and sleep disturbances, that may signal elevated fall risk. It also explores resident and staff acceptance of the technology. An open-label, non-randomized feasibility trial using a single-group, post-test mixed methods design was conducted with 24 residents at a RAC in Sydney, Australia. Ambient and wearable sensor data and clinical records were collected, alongside interviews with residents and staff. Data were analysed using quantitative and qualitative techniques to assess feasibility and user experience. Sensor data revealed diverse resident routines and rapid staff responses to alerts. Predictive analytics showed promise for identifying elevated fall risk, though further validation is needed. Qualitative feedback from 10 residents indicated residents found the system mostly unobtrusive but raised concerns around privacy and false alerts that triggered staff interventions. Despite this, many residents acknowledged its value, especially for individuals with higher vulnerability. Interviews with eight staff members echoed the system's potential to enhance monitoring and safety, but noted technical and training challenges. The study demonstrates that sensor-based monitoring in RACs is technically feasible and generally acceptable. The findings support its integration into aged care as a proactive, person-centred approach to falls management, provided that implementation is supported by thoughtful design, clear communication, and staff training. • Sensor-based monitoring in residential care is feasible and broadly accepted by staff and residents. • Passive sensors can be used to detect diverse resident routines and fall-related behaviours. • Predictive machine learning models show promise for identifying fall risk from movement data. • Residents found sensors unobtrusive but raised privacy and alert concerns. • Staff valued alerts but noted technical issues and training needs.