Health systems risk automating existing problems rather than improving care if artificial intelligence is introduced without the governance, workforce capability and organisational systems needed to use it effectively, according to a leading Australian patient safety researcher.
Australian Institute of Health Innovation founding director Professor Jeffrey Braithwaite has looked into the role of AI in healthcare improvement in a new paper published in BMJ Innovations and has questioned whether the technology is transforming healthcare or providing faster ways of tackling longstanding problems.
Prof Braithwaite said AI could identify risks, predict clinical deterioration and support clinical decision-making, but these capabilities did not automatically translate into improved care.
“There’s no doubt AI can help identify risks, predict deterioration and support clinical decision-making. But prediction is not prevention and technology is not improvement,” Prof Braithwaite said.
“The healthcare organisations that benefit most from AI will be those that build the systems, culture and governance needed to act on what the technology tells them.”
The paper points to a shift from periodic measurement and audits towards near real-time monitoring of patient safety and health system performance, with AI systems able to detect clinical deterioration and medication risks earlier.
However, Prof Braithwaite said the ability to identify problems more quickly did not necessarily mean health systems were equipped to respond.
“Healthcare, supported by technology, has become much better at sensing problems. The harder challenge is responding to those signals in ways that genuinely improve outcomes for patients,” he said.
Prof Braithwaite also warned that AI systems reflected the data they were trained on and the assumptions built into them.
“If we don’t pay close attention to governance, transparency and equity, we risk automating existing problems rather than solving them,” he said.
The paper notes that leadership, workforce capability, clinical judgement, teamwork and organisational learning will remain central to improving healthcare – with AI able to support rather than replace those elements.
Prof Braithwaite said AI could strengthen feedback loops and support decision-making, but people remained at the centre of safer care.
“Healthcare is a complex sociotechnical system. AI is a powerful ingredient, but it isn’t the recipe. If we want AI to improve healthcare at scale, we need to focus less on what the technology can do and more on what our health systems must become.”
Add Comment