• ADVERTISEMENT FEATURE Advertiser retains sole responsibility for the content of this article

Take-home lessons on diabetes data

There were no smartphones when Professor Nilmini Wickramasinghe began working on what developed into DiaMonD (Diabetes Monitoring Device), an app for diabetes self-management with a feedback mechanism between patients and clinicians. However, over almost 20 years, the project has evolved into a very modern solution to Australia’s fastest growing chronic condition.

A 2018 study at Austin Health, a major tertiary health service in Melbourne’s north-east, found that 34% of patients had diabetes. Professor Greg Johnson, CEO for Diabetes Australia, suggested then that diabetes now accounts for about one third of all hospital admissions. Type 2 diabetes, a combination of insulin resistance and impaired insulin production, represents 85% of all cases, but is largely preventable. It is strongly associated with high blood pressure, abnormal cholesterol levels, and excess weight, so prevention and self-management measures have the potential to ease a huge burden on hospitals.

Professor Nilmini Wickramasinghe is developing an app incorporating article intelligence to predict a diabetes patient's future blood glucose level (based on prior activity and food intake) and to give advice on food and exercise.© Eamon Gallagher

“Early detection and proactive management of diabetes is essential,” says Wickramasinghe. The DiaMonD app does this by gathering, recording, storing and charting all information relevant to a user’s diabetes. That includes blood sugar level readings collected from a glucometer (a personal glucose meter) and entered either manually or via a Bluetooth connection. This is recorded against the time and content of meals and other relevant information, such as exercise. It also lets patients send details about their blood sugar readings to a nominated care coordinator to receive recommendations for diet, exercise and insulin titration, the amount of insulin a diabetic must self administer.

“You have lots of options,” Wickramasinghe says. “If you don’t want to enter all of that, the app will graph blood glucose levels over time, and all this information is then sent to the nominated healthcare professional.” A feedback system allows a nurse or doctor to send back almost instantaneous responses. These could range from “well done, everything looks fine” to “please come in for a check-up.” What makes the system particularly beneficial is that outcomes can be closely connected to patient behaviours and activities, which is linked to cognitive feedback that can result in changed behaviours.

The results of a recent Australian trial for women with gestational diabetes mellitus (GDM), which is usually a short-term disease that develops during pregnancy that has huge potential health implications for the developing foetus, were reported in 2019 in the journal JMIR Diabetes.

“A wonderful initiative,” reported one patient for the paper. Another user appreciated the privacy and reassurance during her pregnancy, saying that being able to check and record blood glucose levels any time of day was convenient: “Could maintain privacy at work, no need to duck out to have conversations and report [blood glucose level].” And that knowing that someone would contact her if there was a concern was helpful.

“We’ve always been very clear we’re not going to cure diabetes, but we’re going to help those people with diabetes have a higher quality of life, to manage better,” Wickramasinghe says. “There’s also a large and growing prediabetic community, 16% of Australian adults, that hopefully we might be able to help avoid developing diabetes.”

Smart glucometers help track blood-sugar levels.© Eamon Gallagher

Ahead of their time

DiaMonD is a creative acronym from the term ‘DIAbetes MONitoring Device’, but it was also seen as an appropriate name, explains Wickramasinghe, because the disease diabetes, like the gem, is multi-faceted. It incorporates a lot of lifestyle factors, such as activity levels, food intake, and genetic factors, so is very hard to manage.

The idea behind DiaMonD harks back to the late 1990s, when Wickramasinghe was completing her PhD at Case Western Reserve University in Cleveland, in the United States.

“That’s where I started my focus on digital health,” she recalls, explaining that although her PhD was in the business school at Case Western, it was about managing technology in healthcare. Wickramasinghe had moved to the US after completing a string of qualifications at the University of Melbourne - a Bachelor of Science in mathematics and computer sciences, an MBA focused on technology management.

With a background like that, she laughs, I knew “I couldn’t cure cancer”, but she became increasingly aware that her tech management expertise could support superior healthcare delivery. Applying technology to health and medical problems was only just emerging.

Wickramasinghe was at a conference in Texas when she ran into Steve Goldberg, who had recently founded the Canada-based tech company INET International, which provides online data collection services to researchers. As the pair talked, the idea of developing a technology solution for people with diabetes began to take shape. While Goldberg was, and remains, very technology focussed, diabetes was “kind of a hobby subject”, says Wickramasinghe. “And I said I think if you’re really keen on this area you should focus on a mobile solution, not a desktop solution.”

“This was 2001 and mobile solutions were almost unheard of, but I thought that’s the way the industry is going. Steve said he would work on one if I helped him, and we developed one of the first mobile solutions, pre-smartphone, to support patient empowerment and self-management of diabetes.” At first, it was a simple texting and tracking system.

When smartphones became available, the pair repurposed their solution to work with the new technological interface and DiaMonD emerged as a downloadable app for smart phones, available across all platforms.

Earlier versions of DiaMonD have been trialled successfully with Type 2 diabetes and gestational diabetes in China, Germany, Canada, Australia and the United States.

But DiaMonD has never been a project Goldberg saw as a potentially lucrative income stream. “It’s been about using informatics in healthcare and it’s just been interesting to support research in this area,” Goldberg explains. “The reason we chose diabetes was to touch something well documented and widely impactful.” There was also medical consensus that diabetes could benefit from improved self-care and management.

“From an industry standpoint, there are very few business models that are very successful at commercialising mobile technology in healthcare,” says Goldberg, explaining why DiaMonD, despite clinical success, is not yet widely available.

Taking DiaMonD public

How far then is DiaMonD from taking off into the real world? That’s where Swinburne comes in.

To integrate the latest artificial intelligence into the DiaMonD system and raise commercial and clinical support for adoption in an Australian healthcare setting, in 2019 the project was chosen as the first Swinburne Cloud Innovation Centre (CIC) Data for Social Good project, powered by Amazon Web Services (AWS), a subsidiary of the tech giant that provides cloud computing and platform services.

Wickramasinghe was last year appointed as Swinburne’s Professor of Digital Health and Deputy Director of its Iverson Health Innovation Research Institute. Her other key external appointments are also seen as highly valuable to the project, particularly her role as Professor of Health Informatics Management at Epworth HealthCare. Her connections and collaborations have also brought in Northern Health diabetes expertise, led by Professor Peter Brooks and Dr Michael Kirk, joining Swinburne’s Professor Penny Schofield and Dr John Zelcer.

The CIC, the first project of its type in the southern hemisphere, brought together collaborators in a 10-week project to develop the concept for a new prototype app incorporating artificial intelligence able to predict a patient’s future blood glucose levels, based on prior activity and food intake, and to give salient advice on food and exercise. Once developed, the prototype will be tested in a clinical trial at Northern Health in Melbourne. “One of the current barriers, however, is about making clinicians feel comfortable with the informal language and regular nature of feedback via smartphone,” says Wickramasinghe. However, as diabetes incidence grows, so does the motivation to spend significant time and resources preventing hospitalisations.

Goldberg, who is still involved, is anticipating a positive impact from the clinical trial. He believes it will provide the impetus to see this sort of technology embraced.

“I’m perpetually optimistic that it will happen,” he says about the technology reaching patients. “It’s now just a matter of time and I think it will be on a grand scale.”

Read about Swinburne’s cutting-edge research in our new magazine.