Australia’s aged care system is a lifeline for nearly half a million older Australians. As of June 30, 2023, 456,000 Australians were utilising aged care services, with 193,000 in residential care and 258,000 receiving home care support, the AIHW’s GEN Aged Care Data revealed. This support system, however, faces immense pressure to meet the growing demand for quality care.
Addressing these challenges head-on, Telstra Health and RMIT University along with the Digital Health Cooperative Research Centre have unveiled a groundbreaking AI-powered tool to predict health risks, transforming the way aged care facilities monitor and manage resident well-being.
RMIT University’s project lead and data scientist, Dr Tabinda Sarwar, says the tool is a much-needed advancement for a burdened system.
“The tool is capable of automatically monitoring both structured and free-text electronic patient records for 36 evidence-based indicators of deterioration,” explains Dr Sarwar. “These indicators are further linked to predicting various health risks, providing a comprehensive system to support nursing staff and improving resident care outcomes.”
“In aged care homes, the elderly are provided with care and support, with nursing staff responsible for both their daily needs and health monitoring. This dual responsibility creates a significant workload as staff must oversee multiple residents and their varying health conditions. Given this existing burden, introducing a manual screening tool is not an optimal solution.”
In collaboration with Telstra Health, RMIT developed a data-driven tool designed to not only monitor residents’ health conditions but also predict adverse health events.
This digital health tool is now winner of the Research Australia’s Digital & Data Health Innovation Award for 2024.
For the team, this recognition holds immense significance.
“It represents both recognition and achievement in making a positive difference—enhancing the work of nursing staff and, in turn, improving the quality of life for elderly residents in aged care homes,” Dr Sarwar shares.
How the tool works
The tool relies on daily operational data collected at aged care homes.
“Nursing staff routinely document notes and record health-related details, which formed the foundation for developing this tool,” Dr Tabinda Sarwar adds.
By processing this data with natural language processing (NLP) techniques, the tool identifies early signs of deterioration and generates predictive alerts for various health risks. Geriatric assessments, observation charts, and progress notes are key data inputs, ensuring a comprehensive approach to health monitoring.
“We applied advanced data analysis and machine learning techniques to daily collected information, enabling us to predict signs of deterioration,” Dr Sarwar explains. “This includes risks such as falls, depression, and even mortality, based on evidence from the extracted data.”
Telstra Health’s Clinical Manager system, deployed in over 360 facilities across Australia, provided the essential infrastructure for the project. The collaboration also included input from aged care nursing staff and support from the Digital Health Cooperative Research Centre (CRC).
“Telstra Health provided access to aged care homes and nursing staff, while RMIT contributed researchers and technical experts to develop digital tools and solutions,” Dr Sarwar notes. “This project would not have been possible without the pivotal role played by the Digital Health CRC. By bridging academia and industry, the Digital Health CRC showcased the immense potential of collaborative efforts in leveraging technology to address complex health challenges.”
Overcoming healthcare challenges
Developing a universal solution for diverse aged care facilities was no small feat. Nursing staff across different homes highlighted unique challenges, requiring the tool to be highly adaptable.
“Consequently, designing a solution that could address broader issues and have a meaningful impact on a larger population proved to be the most complex part of the project,” she adds.
To validate the tool’s usability, the team conducted an independent study to ensure that the developed solution is user-friendly and can be easily adopted by nursing staff without difficulty.
“To validate and evaluate its performance and accuracy, we employed statistical and machine learning techniques, which were essential for ensuring the tool’s clinical feasibility. Additionally, the prediction of deterioration relied on machine learning models, highlighting the critical role of data mining and machine learning in the success of the project.”
The tool is currently with Telstra Health, which is in the process of deployment.
“We have been informed that numerous aged care homes have already expressed interest in adopting the tool,” Dr Tabinda Sarwar says.
Telstra Health holds the rights to the tool so any plans to expand its functionality to predict additional health risks depend on their future strategies.
But Dr Sarwar points out: “This tool has the potential to be extended to other healthcare settings, such as hospitals, and could play a key role in transforming the healthcare industry.”
READ ALSO: Aged care facility for South Asian seniors gets off the ground
