Blog

New Digital Health Standards Council targets AI safety gap

A new Australian digital health standards body has launched with the aim of helping health services assess the safety and quality of AI and other technologies that are not regulated as medical devices.

The Digital Health Standards Council of Australasia (DHSCA) began operations this week, seeking to provide services for technology vendors, clinicians, boards, executives and health service organisations.

The clinically led organisation will focus on helping health services strengthen their governance of digital health and AI technologies, as well as independently assessing products that fall below the Software as a Medical Device (SaMD) threshold.

DHSCA said only some digital health and AI tools were covered by existing regulatory arrangements, with the Therapeutic Goods Administration overseeing products that meet the SaMD threshold.

The Australian Commission on Safety and Quality in Health Care also administers voluntary national safety and quality standards for digital mental health services.

DHSCA chief executive Chris Boyd-Skinner said many emerging technologies did not fall within those arrangements.

“The TGA regulates medical devices, and it does that exceptionally well. But most of the AI turning up in Australian health services isn’t classed as a medical device,” he said.

“A chatbot triaging a patient, or an app carrying someone’s mental health information, doesn’t clear the SaMD threshold, so health services are left to work it out themselves, one procurement at a time.”

DHSCA plans to assess tools that fall below the SaMD threshold using GUARDS, a six-domain clinical and technical safety framework for AI-enabled technologies, following external validation and piloting.

It will also work with boards, clinical and technical governance leads and executives to help organisations build their capability to govern healthcare technologies.

Accreditation of services against national safety and quality standards is also planned, subject to relevant approval which DHSCA intends to seek in 2027.

Mr Boyd-Skinner said the organisation was not seeking to duplicate the work of existing regulators or governments.

“We’re not aiming to replicate the work of governments, jurisdictions or regulators in any way,” he said.

“This is about an additional unmet need and supporting organisations that want to build, deploy and scale health technologies safely.”

DHSCA said it was already working with health services, technology developers and peak bodies – including Aboriginal Community Controlled Health Organisations, Primary Care Networks and private hospital groups on AI governance.

The organisation is also establishing an External Advisory Council, with its inaugural chair and membership expected to be announced in the coming months.

Mr Boyd-Skinner said both technology developers and healthcare organisations were looking for a clearer way to demonstrate and assess the safety of emerging technologies.

“Developers have wanted a way to show a product is safe, and when I work in the intensive care unit I hear time and again how buyers want something independent to point to,” he said.

“Neither side has had a straightforward path, so everything we’re working towards over the next twelve months is aimed at closing that gap.”

DHSCA has been established to serve Australasia healthcare services, with initial operations focused in Australia. New Zealand and Pacific Island Nation services would follow.