Catalyst: Strategic – New Zealand-Singapore Leveraging AI for Healthy Ageing 2025

Dementia prevalence is rising globally and projected to affect 168,000 New Zealanders by 2050, with Alzheimer’s disease the most common cause. “What does the future hold?” is a question people worried about their memory often ask health practitioners. However, the complexity of Alzheimer’s and related diseases makes predicting future dementia a huge challenge.

Brain imaging and blood-based biomarkers can provide information on the earliest changes before dementia onset, but on their own do not account completely for a person’s clinical presentation. Artificial Intelligence (AI) techniques provide a unique opportunity to combine large amounts of brain, blood, cognitive, genetic, clinical and general health data to predict risk for dementia. This project will develop an AI-tool ready for clinical use, which will help clinicians identify people who are most at risk of progressing from mild cognitive difficulties to dementia. Using emerging biomarkers and developing new AI applications, this project will grow biomedical technology and AI sectors in New Zealand and Singapore.

Explainable AI methods will allow us to understand what data contributes most to predicting dementia risk, providing new scientific insights, potential treatment targets and diagnostic tests. The project brings together experts in New Zealand and Singapore, leveraging national longitudinal studies in both countries, on which AI models will be developed, validated, and tested. This will enable assessment of cross-population/ethnicity generalisability, reliability and fairness of AI predictions.

The tool will be ready to deploy into clinical settings across New Zealand and Singapore, where it could help stratify patient management in limited resource settings, for example identify who requires priority referral to a specialist or inform risk-benefit discussions on the use of new disease modifying therapies. To facilitate future ease of use by clinicians, we are partnering with Siemens Healthineers to develop brain-image analysis tools, and we will co-design the dementia risk tool with doctors who will ultimately use it.

The AI tool will predict level of risk for developing dementia and identify who will benefit most from targeted interventions (e.g. lifestyle or pharmaceutical therapies) to delay or prevent dementia onset. This will benefit patient lives, while simultaneously reducing healthcare costs for New Zealand and Singapore.

The project is led by Dr Catherine Morgan and Professor Lynette Tippett (School of Psychology), Dr Katerina Taskova (School of Computer Science) at Waipapa Taumata Rau, University of Auckland and Dr Narun Pat at Ōtākou Whakaihu Waka, University of Otago, working closely with dementia specialists at Te Whatu Ora, and collaborators at the National University of Singapore led by Associate Professor Saima Hilal.

Singapore and New Zealand both use interRAI, a standardised assessment tool that supports the care of older adults. While interRAI is reliable and effective, integrating Artificial Intelligence (AI) presents a transformative opportunity to enhance healthy ageing and support older people to live longer, more independent lives.

Our project brings together clinicians and researchers from the University of Otago, Singapore Management University, University of Canterbury, and University of Auckland. We will identify how to effectively integrate AI into the interRAI assessment, risk prediction, and care planning process to improve efficiency, consistency, and personalisation of care. We will achieve this with a three-pronged approach:

1. AI-assisted Assessments: By partially automating the currently manual interRAI process, we can reduce assessment time by 50% while improving accuracy. We will integrate structured health data and multimedia inputs to generate enriched assessments.
2. AI-enhanced Risk Prediction: We will develop predictive models for outcomes such as fracture risk, cognitive decline, and depression. These models will be embedded into interRAI software to support timely, targeted interventions.
3. AI-driven Personalised Care Plans: We will create dynamic, user-friendly care plans using a knowledge- based AI system enhanced by large language models. These plans will be tailored for patients, families, and clinicians, ensuring clarity and actionable guidance.

With support from New Zealand’s Health NZ and ACC, and Singapore’s Agency for Integrated Care, Kwong Wai Shiu Hospital, NWC Longevity Practice, and 59 Socio-Techno Ventures, this initiative will augment existing systems to deliver scalable, cost-effective improvements to aged-care while growing our respective AI sectors.

Older adults often experience changes in memory, processing speed, speech, and executive function, potentially signaling cognitive decline. Currently, these conditions are detected through time-consuming individual clinical assessments, which can miss subtle daily changes. Existing digital tools only monitor cognitive performance, offering little beyond diagnosis.

This project aims to develop an AI-augmented interactive platform which will collect a range of data from the user and the caregiver using a mobile app. It will integrate speech-based interactive tasks and cognitive games, supporting longitudinal monitoring and cognitive improvement, to be analysed by AI, creating a risk assessment and care plan. The platform will provide a continuous cycle of assessment, intervention, and progress tracking. The mobile app will facilitate remote care - supporting users and caregivers outside traditional clinical settings. By combining AI-powered speech analysis with adaptive game mechanics, the tool engages users in fun, personalised exercises which will assess and strengthen cognitive functions; including reaction time, verbal fluency, memory, and decision-making. This dual-purpose approach will shift the focus from ‘Monitoring Decline’ to ‘Strengthening Cognitive Resilience.’

Additionally, we will deliver personalised cognitive training modules that track progress and provide AI-driven feedback. Clinicians will be provided with app-created summaries, alerts of significant changes, and identified risk indicators. The latter may send alerts to family or medical professionals, enabling earlier intervention. Strategically, the platform will deliver data-driven insights on cognitive training for long-term cognitive health.

The tool will be tested in Singapore and New Zealand with clinical and community-based cohorts. This will ensure effectiveness across diverse populations, whereby collaborators National Neuroscience Institute and Lions Befrienders (Singapore), Elli Cares and Dementia NZ (New Zealand) will introduce our tool to users. We will validate improvements in cognitive function while detecting early risk by comparing AI-based assessments with standard clinical tests.