banner
Home / Blog / Could watches and fitness trackers spot disease and pandemic threats? AI scientists hope so
Blog

Could watches and fitness trackers spot disease and pandemic threats? AI scientists hope so

Apr 26, 2023Apr 26, 2023

Imagine having a little device on your wrist feeding your body's signals back to a robot that's been trained especially to tell you whether you've been infected by a disease.

It sounds like an Orwellian-like reach for a desperate society facing a hidden enemy like the cordyceps in The Last of Us.

But this type of technology is more fact than fiction for some Australian scientists, who have revealed they're testing the possibility of such a solution with hardware that you can buy right now.

Researchers at the University of South Australia will use consumer-grade technology like smart watches to see if they can detect the earliest signs of infection by measuring subtle changes in things like:

They're hoping that by measuring these vital signs, they'll be able to teach an artificial intelligence algorithm to pinpoint when someone is getting sick.

They'll measure the most intimate details of those who sign up to the study, probing the highly private parts of someone's life in the name of research. But they'll stop short of asking questions about things like sex.

It represents another privacy trade-off in a world where technology is becoming more and more intrusive. But defence scientists say it could have big benefits, in both peacetime and wartime (more on that in a second).

One potential benefit: The researchers hope it could lead to a monitoring system that could slow or stop the next pandemic.

Axel Bender, who leads a team of scientists focused on innovation at the Australian government's Defence Science and Technology Group, said it could even prevent another lockdown.

He said health authorities would be able to respond to a pandemic threat more quickly, quarantining people while infection rates were still low.

"You would not need to lock down anymore because you would just break infection pathways so early," Dr Bender said.

The research, which is an Australian first, is part of a growing body of work leveraging the use of the human body as a sensor to detect biological or chemical threats. It is formally known in the Australian defence community as the Human Integrated Sensor System.

The project is being funded by the Defence Science and Technology Group, which is an arm of Australia's Department of Defence.

It hopes the research will also help soldiers on the battlefield and in disaster zones facing biological warfare threats.

Biological warfare refers to pathogens like viruses, bacteria and fungi being used against soldiers and spread through infection.

Dr Bender said this type of technology would enable the military to monitor soldiers to see if they've possibly been infected by a pathogen.

Under the plan, sick soldiers wearing devices could be isolated from their units sooner and then treated, preventing the spread of infection.

"We know when it comes to …. chemical or biological warfare agents … not everyone abides by conventions that prohibit those weapons," Dr Bender said.

The promise of the technology has also prompted some scientists, including Dr Bender, to personally call for Australia to invest in this type of research as part of a national effort.

There's a fair bit of research required first.

Through millions of years of evolution, the human body has become sensitive to things that cause it stress, and generally reacts to them immediately.

The plan from scientists is to use the body as an alarm to detect diseases or biological threats, including chemical warfare.

They'll do that by giving 100 study participants three devices — an Oura ring which measures sleep, a Garmin smartwatch which can track activity, and an EmbracePlus smartwatch which collects health data.

Those devices will measure a participant's vital signs before and after they receive a COVID or flu vaccine, which will challenge their immune system in a similar way to contracting an infection.

The data that is collected from study participants can be punched into an artificial intelligence algorithm, which will track the subtle changes that occur when there is a threat to the immune system.

"We know that there's going to be a certain sort of physiological response," said the lead researcher, Professor Siobhan Banks from the University of South Australia.

"From this data we hope to be able to develop algorithms to … see those early signs of an immune response.

"If people knew that they were sick or they had a high risk of being sick and that they were possibly going to be contagious … this would really mean that people would make right choices about whether to go out into social situations, whether to go to work or stay home."

Dr Bender said that knowing someone is infected as soon as possible means they can be treated much quicker with medication, long before their body is loaded with a pathogen.

That would also mean outbreaks could be better managed at a population level through targeted quarantine, he said, and that could end the need for lockdowns.

"You can actually isolate infected people from the rest of the population at a time when the infected person is not infectious, which means you will break that infection pathway much earlier," he said.

Dr Bender also believes that the technology around a Human Integrated Sensor System will eventually be able to identify new and emerging infections in the same way it's detecting old ones — by noticing tiny changes in the body.

That information will then be combined with current technology to quickly produce algorithms to detect these newer viruses.

Some public health experts are pretty cautious about the proposed benefits.

Professor Nancy Baxter, who leads the Melbourne School of Population and Global Health at the University of Melbourne, said that the plan was "probably feasible" but would be dependent on the type of virus or bacteria that was the source of infection.

"I think when they're talking about trying to prevent lockdowns, I think that that's kind of blue sky thinking and and probably over-promising," Professor Baxter said.

"I think that there are more useful, more realistic ways to think about wearables and human health."

Professor Tony Blakely, an epidemiologist also from the University of Melbourne, said the research might encourage people to get tested or isolate sooner.

"You could think of an Orwellian world where we're all required to wear this device and it's monitored by somebody centrally and as soon as your temperature goes up or your heart rate, you're directed to go and get a test," Professor Blakely said.

"Under all those circumstances it would help a lot, but I don't think society would accept that level of draconian type of implementation.

"It's no panacea but it might help a little bit alongside other measures like having masks ready to go."

Dr Zygmunt Szpak, whose company Insight Via AI is building the algorithm, said the project was a significant move for Australia, which has "immense artificial intelligence capability and talent".

"But we are not very good at commercialising it," Dr Szpak said.

"And that's kind of a tragedy and I think it will be great to really scale this capability up so that we can tap into our own excellent graduates."

Dr Bender hopes this kind of technology will be available by the end of the decade.