TYPE OF SUPPORT
Research Background
Challenges affecting Australia's data infrastructure hinders large-scale research efforts for
chronic diseases, which have significant long-term effects, comorbidities, and economic burdens.
To address this gap, the National Infrastructure for Federated Learning in Digital Health (NINA) project will implement a federated learning (FL) based data infrastructure in Australia.
This innovative approach allows researchers to analyse health data across organizations and states without integrating datasets, bringing the analysis to the data which remains in-situ, providing an ethical pathway to access comprehensive data.
The project aims to:
Pioneer new ways to connect and interrogate separated data sets.
Demonstrate the use of FL to accelerate research and clinical care using AI and analytics.
Focus on chronic disease exemplars: diabetes, rheumatoid arthritis, osteoarthritis, and cancer.
By implementing FL in Australia's healthcare research landscape, NINA will enable researchers and industry professionals to access large-scale comprehensive health data, accelerating research and improving outcomes for individuals with chronic diseases.
QCIF Role
QCIF is providing the technical expertise to design, test, and implement the technical
aspects of federated learning.
This includes evaluating federated learning tools and frameworks, building test bed federated learning resources for researchers, provision of secure cloud services, provision of data infrastructure, coordinating deployment with stakeholders, provision of secure data analytics.
Research Outcome & Impact
NINA is building a federated learning-driven digital health infrastructure to tackle chronic diseases like cancer, diabetes, rheumatoid arthritis, and osteoarthritis.
By bringing machine learning to siloed data, rather than centralising it, NINA preserves data sovereignty while accelerating research.
This approach enables near real-time AI/ML analysis, supports privacy across jurisdictions, and empowers healthcare organisations to use federated learning. It lays the foundation for fast, efficient digital health research that improves clinical practice and patient outcomes.
This is an exciting project in an emerging technology that will improve access to
siloed health data across Australia. This will facilitate research that would otherwise be difficult to undertake for a range of chronic conditions affecting Australians.
Peter Marendy, Head of Data & Software, QCIF
