Richard Musgrove
Dr Richard Musgrove is a zoologist and science journalist. He writes for Cosmos and Double Helix.
CHRISTCHURCH May 29: Improving healthcare in the face of increasing demand means embracing digital twins and other technology, say Kiwi researchers.
“Despite growing demand and abundant data from medical devices and systems, healthcare still lacks effective automation, says University of Canterbury’s (UC) Distinguished Professor Geoff Chase. Effective healthcare automation hinges on accurately predicting how patients will respond to treatment.”
Waking up in an intensive care unit (ICU) or the post-operative ward can be a daunting experience, disorientating and even frightening. Beeping machines, tubes and catheters running into all directions, blood and other fluids in bags, and the sights and sounds of other patients’ distress. And medical staff making it all work in what looks like chaos to the untrained and bleary eye.
Nurses, with cheerily comforting words, check the patient, and numbers, tubes, and catheters, and move on. Monitors recording heart rate, blood pressure, oxygen saturation and body temperature are always the focus. Consultants on the daily round with their student entourage linger longer — each patient questioned, and their case chart and monitor scanned and discussed.
Surprisingly, none of that monitor information is saved. That digital glimpse of the body’s inner workings is ephemeral.
As Chase told Cosmos: “If you go into an ICU, there are always screens with data. You see the heart rate and the blood pressure numbers going by. They’re not saved, not necessarily, and they’re certainly not all saved in a single point. They’re not saved in a way that you can use them in real time.”
Chase and UC colleague, Dr Lui Holder-Pearson, both engineers, plan to tackle major challenges in ICU efficiency by pairing validated “digital twins” with “digital cockpit“ technologies.
What are digital twins?
Digital twins (DTs) – digital models of physical processes in the real world – are collections of data taken from sensors on physical objects, which are then manipulated to model how changes would affect the rest of the system, be it a bridge, a plane, or a human.
Making digital twins for an extremely complex live human is challenging. They do exist but are limited to modelling specific body systems.
Hips, for example, vary from person to person, so replacements must be tailored. Digital twins have made that easier. “Everyone has different sized and differently-shaped bones, says Dr Duncan Bakke, a research bioengineer at Auckland’s Formus Labs.
“If we can go in and take a CT scan and make a 3d model of the bone (a digital twin), then we can figure out ahead of time which size of implant is going to be the right one, rather than having to try it out in surgery or with intuition.”
Bakke is not involved in the UC project.
“We have had a digital twin for glycemia (glucose in the blood) working on a tablet in the Christchurch Public Hospital ICU in one form or another, since August, 2005,” says Chase.
“Good glycaemic control, if given to all patients can reduce mortality in the cohort who require it (~33% of all ICU patients) by 10-25%”.
“The DT receives glucose data, assesses patient-specific insulin sensitivity and response to insulin and nutrition, and recommends (within an agreed protocol framework) the dose for the next period of time.. It has the lowest rate for hypoglycemia and highest nutrition delivery rate worldwide over any comparable cohort.”
Around 7500 patients have been treated using this DT in Christchurch, with similar numbers in Malaysia, Belgium and Hungary, he adds.
“We saved about one life per bed space per year, because the number of bed spaces have been growing and the number of patients, and we save 1-2 million dollars a year in net direct costs. So that’s the real power of what these things can do.”
Data entry for this DT is manual, with no inputs from a paired digital cockpit.
Such a cockpit, as proposed by Chase and his team, would be a centralised workstation, connected directly to devices and sensors monitoring each patient. Data would be gathered and commands sent.
Open a valve, adjust a ventilator or send an alert — whatever is necessary, depending on the care needs of the specific patient as reflected in their digital twin, which models their condition and sits within the cockpit.
The glycaemia DT is a standalone tool used by medical staff. A DT sitting within a digital cockpit could potentially make care decisions.
Case asks, “If I make a recommendation you always follow, who is making the decision?”
“Our adherence to glucose control recommendations varies from 96-98% to virtually 100% in different parts of the world. Technically, this comes down to the human “in the loop” agreeing to recommendations which it then administers, or the human “on the loop” overseeing the treatment and intervening only where necessary.
The idea of full “fire and forget” automation is not there yet. Thus, there is clinical practice culture change that comes with such changes.”
“Our project team includes both regulatory and organisational change specialists to help co-design these tools for best acceptance and regulatory approval.”
Chase also emphasises that DTs are not driven by artificial intelligence (AI) but are personalised to each patient using their data. “They are objective, deterministic and unbiased,” he says.
“There’s 2 pieces to this project,” Chase told Cosmos.
“One is the cockpit. The other is a smart — state-of-the-art, clinically validated, natively integrated — digital twin that sits in it.”
“And the goal there is to provide the connectivity through all that with smart and optimised decision making.”
“Our research programme will develop both the foundational automation platform – our Digital Cockpit – and DT applications that, together, will unlock healthcare automation.”
Chase believes the programme has the capacity to improve patient outcomes, reduce health costs, and elevate New Zealand to the forefront of the emerging MedTech industry.
