ML project to unlock secret data on chronic diseases

A new project led by the University of Queensland seeks to create a national data network that will support the development of new solutions for the management of chronic diseases.

Recently, the National Infrastructure for Federated Learning in Digital Health (NINA) project secured $6 million ($3.9 million) in funding from the federal government’s Medical Research Future Fund. It also received an additional A$7.7 million ($5 million) in contributions from UQ, Monash and Macquarie Universities and the Queensland Cyber ​​Infrastructure Foundation.

what is this about

According to a media release, the five-year project will enable researchers to use machine learning to access secret information on debilitating chronic diseases such as diabetes, rheumatoid arthritis and osteoarthritis. It will prepare and harmonize data with global standards while protecting individual privacy. Analytics will also be generated and shared across health organizations and states.

UQ will collaborate with 23 Australian and global partners to co-design the conceptual framework for NINA and accelerate translation and adoption of the data model at a national level.

why it matters

Researchers across Australia are finding it difficult to access health databases and advance their research on digital health.

“Australia has excellent digital health records, but data remains hidden in health systems, leaving talented researchers with access to millions of records about trends in treatment and serious conditions,” said Associate Professor Claire Sullivan from UQ’s Queensland Digital Health Centre. is being prevented from.”

Chris Benn, Professor of Practice in Digital Health at Monash University’s Faculty of IT, said various privacy and data sharing restrictions hinder the meaningful use of such databases.

“This has ultimately led to a huge divide around healthcare sectors, including lack of data connectivity across primary, secondary and tertiary care,” he said in a separate statement.

The intention of the NINA project is to “put data to work” to find solutions to better manage chronic conditions. “Rather than trying to merge different data sets to enable machine learning centrally, the project will bring machine learning to the data,” Sullivan said.

big trend

Despite having the will to drive digital transformation using data and analytics, most healthcare organizations in Australia and New Zealand lack the ability to share real-time data and integrate disparate systems, according to a recent report. According to Study commissioned by Intersystems. They are analyzing a limited number of data in spite of having a lot of data because multiple datasets are not interoperable.

Still, most providers wanted standardization of data exchange. The Australian Digital Health Agency is helping to accomplish this through its National Healthcare Interoperability Plan, which envisions a more connected Australian health system by 2027.

As part of this effort, the agency recently partnered with Health Level Seven Australia to promote Steady adoption of FHIR standards in the country. It also tied up with CSIRO’s Australian e-Health Research Center to create the National Clinical Terminology Service, which will provide terminology services and tools that will enable connectivity across the health system.