Perth Pawsey supercomputer predicts ICU patient outcomes before they worsen

From mapping the stars to predicting weather patterns, Perth's Pawsey supercomputing centre has been used by the brightest minds to solve some of the most complex problems.

Key points: An algorithm is being used to track patient data and predict brain bleeds

About 40,000 hours of patient data has been used to create the algorithm

It could give medical staff up to 20 minutes' notice for patients with traumatic brain injuries

Now researchers hope a "world first" algorithm that predicts serious brain bleeds before they happen could save lives across Australia.

Royal Perth Hospital head of data science, Shiv Meka, said more than 40,000 hours of patient data had been collected from 200 patients from Royal Perth Hospital, Alfred Hospital, and Royal Melbourne Hospital to develop the algorithm.

After working on 20 different data models, Mr Meka and his team created one that could predict if a brain bleed was imminent — giving medical staff up to 20 minutes before the bleed to intervene.

Shiv Meka and Robert McNamara are trying to improve patient outcomes. ( Supplied: Pawsey Supercomputing Research Centre )

"So you make sense of this data, and then you train the model to predict an event," he said.

"And based off that, the doctor makes a decision whether this is a high-risk patient and if they have to intervene.

"The only way that this could have happened is because of Pawsey, if we used the cloud [storage], we would have burned a big hole in our budget."

The Pawsey Centre's Setonix supercomputer was recently ranked the 15th most powerful research supercomputer in the world, making it Australia's most powerful system used for research.

Royal Perth Hospital intensive care specialist Robert McNamara said the project was a breakthrough in the treatment of traumatic brain injuries.

"Across Australia, there are about 2,000 severe traumatic brain injuries a year. In Royal Perth [Hospital], we'll see about 60 to 80 of these types of patients," Dr McNamara said.

"About half of those patients will have a bad outcome, they will either die or be severely disabled because of their injury, which led to a re-evaluation of how we actually managed brain injury in the intensive care unit."

He said it was common for medical staff to be caught off-guard when severe brain swelling or bleeding occurred in a patient.

"You can be sitting in intensive care looking at some of the [brain] pressure that's in the safe zone and with very little warning, or sometimes no warning at all, suddenly that patient's pressures are in unsafe levels," he said.

He said even a few minutes could be the difference between life and death.

"Like any bruise or broken bone, swelling to the brain comes in about 12 to 24 hours after the injury has occurred," Dr McNamara said.

"But unlike your limbs, the brain sits within the skull that acts as a closed box.

"Once you reach the limits of this box, the pressure ramps up quite significantly."

A graphics processing unit from inside the supercomputer. ( Supplied: Pawsey Supercomputing Research Centre )

Intricate software allows researchers to monitor every "heartbeat, vital sign, and pressure wave" from the patient.

"Suddenly we're now generating much higher frequency data for analysis," Dr McNamara said.

"So we had the data, and fortunately, we met Shiv, who gave it a go and actually succeeded where at least 40 or 45 other groups failed."

He said they had created the world's first operational traumatic intracranial hypertension prediction algorithm.

"For the first time we're actually going to be able to intervene before it happens," he said.

The Setonix supercomputer has the power of tens of thousands of laptops working in unison. ( Supplied: Pawsey Supercomputing Research Centre )

Pawsey Supercomputing Research Centre executive director Mark Stickells said the algorithm was an illustration of how advances in data science and computing capabilities, honed through analysing astronomical data, could help people and save lives.

Mark Stickells says advances in computing are helping inform medical science. ( Supplied: Pawsey Supercomputing Research Centre )

"We're like a laboratory that uses a computer as its main tool," he said.

"We do amazing science at the level of the universe with astronomers using advanced telescopes, developing insights into our universe — but this is right down to human scale.

"Increasingly, medical science has been informed by advances in computing, machine learning and artificial intelligence."