As aged care providers seek smarter solutions to improve efficiency, Regis Aged Care has given sector and glimpse into what the future after revealing their implementation of a purpose-built, AI-powered technology to streamline the laborious process of progress note management for their registered nurses (RNs).
Speaking at the RLDatix Connected Health & Care Summit Asia-Pacific 2025 this week, Chief Information Officer (CIO) of Regis Aged Care, Dr Imtiaz Bhayat, detailed the technical aspects of the system that will be the envy of other providers within the sector who continue to grapple increasing workloads, complex care needs, and the growing demand for high-quality documentation.
One of the significant challenges for aged care clinicians is the sheer volume of progress notes they must review daily. At a Regis aged care facility, an RN beginning a shift may need to read long and complex notes, a task that typically takes between two to four hours.
Given the frequency of interruptions in a clinical setting, this process can be both time-consuming and overwhelming in any health care and aged care setting, leading to potential delays in care decisions and inconsistencies in service delivery.
Aged care generates vast amounts of data, extending beyond standard health metrics such as blood pressure and oxygen levels. Observations on residents’ behaviours, moods, and other subtle changes are also recorded, providing valuable insights into their well-being.
AI technology now enables Regis Aged Care to synthesise this wealth of information into concise, actionable summaries, allowing clinicians to focus more on providing care rather than sifting through excessive documentation.
Instead of adopting a generic AI tool, Regis Aged Care developed a structured system tailored to clinical workflows. This tool categorises progress notes based on predefined health concerns such as agitation, medication administration, refusals of care, pain presence, and infection risks.
By organising notes into structured summaries, the AI enables RNs to quickly identify residents requiring urgent attention.
To ensure the tool’s effectiveness, the AI does not merely generate general summaries but uses carefully designed prompts to extract clinically relevant information. This approach ensures that summaries remain accurate and useful, enhancing the efficiency of care delivery.
A key challenge in implementing AI in a clinical setting is ensuring accuracy and consistency. Left unstructured, AI models can generate varied responses depending on how information is presented. To mitigate this, Regis employed advanced “prompt engineering” techniques to refine the AI’s performance.
For instance, the AI must recognise signs of agitation even when the term itself is not explicitly mentioned. An example describing a resident as “calling out loudly,” “agitated at staff,” and “wandering the hallways” clearly indicates agitation.
However, another case may describe a resident refusing personal hygiene care despite multiple interventions. While the term “agitation” is absent, the AI detects patterns of distress, ensuring that clinicians are alerted to potential concerns and can help ensure resident care and wellbeing support is provided.
The system also aids in identifying infection risks. While a straightforward case may explicitly state symptoms such as a sore throat and a cough, other cases may be more ambiguous. If a resident is experiencing vomiting without a clear diagnosis, the AI analyses patterns across multiple progress notes to flag potential health issues.
Another important function of the AI is tracking the use of non-regular medications. Given the extensive medication regimens in aged care, it is vital to monitor the impact of short-term prescriptions. The AI highlights these instances, ensuring that clinicians can assess their effectiveness and determine whether further intervention is needed.
Beyond summarising progress notes, the AI at Regis is being developed to identify predictive patterns based on data collected over 30 days. Fall prevention is a primary area of focus, with the AI analysing behavioural changes, mobility issues, and medication adjustments to predict which residents may be at higher risk of falling.
This proactive approach allows staff to intervene before incidents occur, improving resident safety.
Aged care settings differ from acute care environments, where immediate interventions are often required. Instead, recognising subtle patterns over time is crucial. By continuously analysing data, the AI enables staff to detect emerging health concerns before they escalate.
The proof of concept of AI-driven progress notes at Regis Aged Care has already demonstrated significant benefits. RNs in the future may be able to access summarised information rather than raw data, enabling faster and more informed decision-making. This enhances care quality while alleviating some of the administrative burden on staff.
With the AI system in place, RNs can begin their shifts with a clear overview of residents who require immediate attention. Additionally, AI-generated summaries standardise documentation, reducing the risk of human error and inconsistencies.
However, it was noted that RNs will still need to review the progress notes in their entirety throughout the day. Despite this, AI still helps important care to be provided early.
Given the increasing regulatory scrutiny in aged care, particularly following the Royal Commission’s findings, this level of accuracy and reliability is essential.
Challenges and Financial Considerations
The Regis AI tool has demonstrated impressive accuracy in identifying key clinical concerns within progress notes, outperforming general AI systems like Microsoft’s Copilot in most categories.
The system was rigorously tested against registered nurses’ assessments, achieving approximately 80% accuracy for immediate clinical concerns and strong performance in detecting signs of pain, though nurses still excel in more nuanced cases. More testing needs to occur to ensure a repeatable result is provided by the AI and that ‘response quality drift’ doesn’t occur or is managed.
However, the financial sustainability of the system remains a critical factor. Running the AI model for 120 residents over a few weeks incurred cost that make it difficult to implement, unless the model is modified to be more efficient. As Regis operates 70 homes, the expense of daily system use would be substantial, necessitating further refinement to ensure long-term viability.
Despite these challenges, the potential for the AI tool to transform clinical workflows, significantly reducing the time required for staff to process and analyse resident data, is impressive.
Tasks that may traditionally take hours for an RN to complete manually, will in the future be streamlined into near-instantaneous results, enhancing efficiency while maintaining high standards of care.
However, a key consideration remains ensuring that RNs continue to review notes for compliance and safety, as required by legislation. Balancing these requirements with cost efficiency will be essential in optimising AI-driven progress note systems for aged care.
Dr Bhayat concluded his presentation by discussing the broader potential of AI in aged care, noting that similar technologies could be expanded to areas such as predictive analytics for fall prevention and personalised care planning.
As AI capabilities continue to advance, the sector is likely to see further innovations that enhance efficiency and improve resident outcomes.
The model of AI-driven progress notes at Regis Aged Care marks a significant step forward in addressing administrative challenges. By leveraging AI to streamline documentation, organisations is not only improving workflow efficiency but also enhancing the overall quality of care for its residents.
With the aged care sector set to expand in the coming years, innovations such as these will be crucial in ensuring sustainable, high-quality service delivery.
Sounds good! If managed appropriately AI could enable the critical people skills required by RNs and carers to be used /applied. The outcome wouid be that service consumers feel cared for and enjoy their lives and health problems are well managed . .
Alternately, the workloads of RNs could be reduced by employing more of them, so they don’t have such a huge number of residents that they are responsible for?
This is why handover with consistency of staff helps.
They are able to see these patterns without AI, its called a skill.
No harm in trying though.