Innovation alone does not improve care.
Australia has built a remarkable capacity to imagine better healthcare, yet a sweeping review by Curtin University researchers reveals that imagination has rarely translated into practice: of more than 7,000 risk prediction tools developed to catch heart disease, diabetes, frailty, and other serious conditions before they take hold, only 21 have found their way into routine clinical use. Published in The Lancet Regional Health - Western Pacific, the finding illuminates a quiet crisis — not of invention, but of delivery — unfolding at precisely the moment an aging population most needs medicine to close the distance between what it knows and what it does. The gap is not a failure of science, but of the systems, funding, and institutional will required to carry science from the research paper to the patient.
- A landmark review of over 7,000 Australian health studies found that just 21 risk prediction tools have been adopted into everyday clinical care — a ratio that exposes a systemic failure to convert innovation into impact.
- Workforce strain, funding shortfalls, and the sheer difficulty of fitting new technology into existing clinical workflows are quietly strangling tools that could catch heart disease, diabetes, and frailty before they become crises.
- The urgency is sharpening as Australia's population ages, with conditions like dementia developing silently over years — meaning every delayed implementation is a window of early intervention permanently closed.
- Researchers from Curtin University and the University of Exeter are calling for implementation science to receive the same investment as discovery science, arguing that building a tool and embedding it into real healthcare are equally demanding and equally necessary tasks.
Australia's hospitals and clinics have grown skilled at inventing tools to catch disease before it strikes — but a team at Curtin University wanted to know how many of those tools were actually being used. The answer was stark: of more than 7,000 studies on risk prediction and screening technologies reviewed, just 21 examples of routine real-world use were found. Published in The Lancet Regional Health - Western Pacific, the research exposes what lead researcher Dr. Jennifer Dunne calls a major gap between what medicine can invent and what it actually delivers to patients.
The tools themselves are not the problem. Australia has developed a growing arsenal of AI and digital systems capable of identifying people at risk of heart disease, falls, frailty, and diabetes complications — and both clinicians and patients tend to respond positively when they encounter them. What stops these tools from reaching patients is a tangle of practical obstacles: funding constraints, overstretched workforces, and the genuine difficulty of integrating new technology into the workflows of busy emergency departments and general practices.
Professor Bronwyn Myers, an implementation science expert and director of Curtin's enAble Institute, frames the challenge plainly: developing a tool is only the beginning. Making it work sustainably inside a functioning, resource-pressured health system is the harder part — and it demands its own investment, training, and institutional design.
The stakes are rising. As Australia's population ages, conditions like dementia develop silently over years before symptoms appear. Earlier knowledge of a person's risk could change the trajectory of their health, help them stay independent longer, and ease future pressure on hospitals and aged care. But only if the tools actually reach patients. As co-author Dr. Leanne Greene of the University of Exeter puts it, innovation alone does not improve care. The researchers are calling for equal attention — and equal resources — to be given to implementation as to invention, before another generation of promising tools quietly gathers dust.
Australia's hospitals and clinics have become quite good at inventing tools to catch disease before it strikes. Researchers at Curtin University set out to find out how many of these tools were actually being used. The answer was sobering: almost none.
The team reviewed more than 7,000 studies on risk prediction and screening tools developed in Australia—the kind of software and systems designed to spot early warning signs of heart disease, falls, frailty, diabetes complications, and other serious conditions. They found 21 examples of these tools being used routinely in real health care settings. Twenty-one out of seven thousand. The research, published in The Lancet Regional Health - Western Pacific, exposes what Dr. Jennifer Dunne, lead researcher at Curtin's Dementia Center of Excellence, calls a major gap between what medicine can invent and what medicine actually delivers to patients.
The problem is not that Australia lacks innovation. The country has developed a growing arsenal of artificial intelligence and digital technologies capable of identifying people at risk of future health problems—tools that could, in theory, give clinicians the chance to intervene before conditions become serious. Health care workers generally support using them. Patients often respond positively when they encounter them. Yet the tools sit largely unused, gathering dust in research papers and pilot programs.
What stops them from reaching patients is a tangle of practical obstacles. Funding constraints squeeze health systems. Workforce pressures mean clinicians are already stretched thin. Integrating new technology into existing systems proves harder than it sounds. A tool that works beautifully in a controlled research environment may not fit naturally into the workflows of a busy emergency department or general practice. Staff need training. Systems need to be redesigned. Money needs to be allocated. These things rarely happen on their own.
Professor Bronwyn Myers, director of Curtin's enAble Institute and an implementation science expert, frames the challenge clearly: developing a tool is only the beginning. Making sure it actually works in the messy reality of a functioning health care system—and stays working—is the harder part. "Health care systems operate under significant funding and resource constraints," Myers explains, "so even effective tools can struggle to gain traction if they don't fit naturally into existing workflows or aren't supported by the right systems, training and funding."
The timing of this finding matters. Australia's population is aging. The number of people over 65 continues to rise. Many of the conditions that lead to lost independence, hospitalization, and dementia develop silently over years before symptoms appear. Knowing a person's risk earlier—before damage is done—could change the trajectory of their health. It could help people stay independent longer. It could reduce the future burden on hospitals and aged care facilities. But only if the tools actually reach patients.
Dr. Leanne Greene, from the University of Exeter and a co-author of the review, puts it plainly: "Australia has no shortage of innovative health technologies, but innovation alone does not improve care." The researchers are calling for a shift in how the health system thinks about new tools. Developing them matters. But so does understanding how to embed them into routine practice. The two require equal investment, equal attention, equal resources. Without that balance, promising innovations will continue to languish on the shelf while patients who could benefit from them never know they exist.
Notable Quotes
Australia has developed tools that can help identify people at risk of future health problems earlier, but most are never adopted widely enough to make a real difference.— Dr. Jennifer Dunne, Curtin University
Developing a tool is only the first step. The real challenge is making sure it can be successfully integrated into busy health care settings and sustained over time.— Professor Bronwyn Myers, Curtin's enAble Institute
The Hearth Conversation Another angle on the story
Why does a tool that works in research fail in a real hospital?
Because a hospital isn't a research lab. In research, you control everything—who uses the tool, when, how. In a real clinic, you have one nurse managing fifteen patients, a computer system from 2015, and no budget for training. The tool has to fit into that chaos, not the other way around.
But if patients respond positively and doctors support it, shouldn't that be enough?
You'd think so. But support and adoption are different things. A doctor can think a tool is great and still not use it if it means staying an extra hour after their shift to enter data, or if it conflicts with how their clinic is already organized.
So this is really about money and time, not science.
It's about both. The science has to be sound, yes. But once it is, the question becomes: who pays for implementation? Who trains the staff? Who maintains the system? Those questions don't have easy answers in a system already stretched thin.
What happens if Australia doesn't close this gap?
As the population ages, you miss an enormous opportunity. You have tools that could catch disease early, prevent hospitalization, help people stay independent longer. If they're not being used, that potential just evaporates. The burden on the health system only grows.
Is this problem unique to Australia?
No. It's a global challenge. But Australia has the chance to be intentional about solving it—to invest as much in implementation as in innovation. Most health systems don't.