The algorithm will see you now

Call to let AI shake up Kiwi healthcare

Here’s something to make you feel better, and even tell you if you’re getting sick: artificial intelligence (AI). Healthcare is rich in data which, if harnessed, can speed patients through the health system, monitor patients’ health and eventually reduce health sector costs, Orion Health chief executive Brad Porter says.

Problem is, primary care, secondary care and personal data are silos that don’t easily talk to each other, Porter says.

Yet they can. Orion Health is already using AI to improve healthcare in Ontario, a Canadian province of 15 million people.

Orion Health hasn’t been able to do the same in its homeland, he says.

“It is frustrating for a 100% New Zealand-owned and operated company,” Porter says. “There's parts of the world embracing us more than what we are able to do in New Zealand.”

In September Porter delivered a keynote address to the Third Digital Health Symposium at the United Nations General Assembly 78 in New York. He told audiences that if AI is the key to unlocking insights and the patient-centric healthcare of the future, it first needs to harness the right data, at the right time, from the right places.

University of Iowa has found medical knowledge will have doubled at least five times in the past year. At the same time hospitals are creating a sky-scraping 50 petabytes (a petabyte is a million gigabytes) of health-related data – sometimes joined up, sometimes not.

And in New Zealand, mostly not. “We are disconnected, and probably the biggest disconnection is between primary care and secondary care, which is what needs to be glued together to the point of care on both sides of the equation,” he says.

Porter uses his 96-year-old grandfather in Canterbury as an example of the problem for patients.

When he caught Covid and went to the GP, he was told to go home. That night he had to go to the emergency department and tell his whole story again.

“They said, you actually have to go to the hospital. So that was eight hours at ED. He went to the hospital and again had to repeat his whole story – at 96. That’s unacceptable, we've got to think more about the efficiencies of connecting up things.”

In his perfect world, a patient would feed information into a computer app or website, which would advise a GP visit, or the emergency department.

“It would do that right away and give you the closest one to your house because, through your digital engagement, I know exactly where you are. So there’s a high level of sophistication. This is the best part. I turn up at the emergency department. I don't have to repeat myself. They know everything about me.”

The key is the best possible data. Generative AI can sort, study and analyse data, so can highlight problems, even looking ahead to where your health is headed.

“Think of a world where you’ve got a health version of ChatGPT which studies your complete longitudinal patient record, every interaction you’ve had, your healthcare journey, all as a data source.”

AI can then compare against this exploding knowledge of medical content through clinical trials, textbooks to determine what’s possibly wrong with you, based on your symptoms. It can predict steps you should take to avoid future illness, something that just wasn’t possible pre-AI. But the data is a siloed mess, Porter says.

“Healthcare has got a massive data problem. It’s disorganised, first and foremost. It makes up approximately a third of the world's data, and it's growing faster than any other data type in the world. Medical knowledge is expanding at a rate that has never been seen before, it now doubles essentially, every month.”

Individuals hold some data. Smart watches such as Fitbits and Apple Watch collect a wide range of health information. Porter sees the day where that information, the GP notes and secondary care data merge into your digital health footprint.

“I’m worried for New Zealand, we’re taking too long to get to that ambition of sharing records across healthcare, across the whole health system, when we really need to be moving fast. We’ve got a lot of innovative New Zealand companies that are banging at the door to get into Te Whatu Ora and Orion Health would love to be a part of it. We do things around the world that I’d love to bring back here. Innovation we’re doing in the Middle East, in Canada, in parts of the US in the UK’s NHS, we haven't had the opportunity to bring that back.”

The flies in the AI ointment are privacy issues. Essentially health data belongs to the patient. So does smart watch data, so sharing it poses privacy issues.

The Office of the Privacy Commissioner, which seeks to develop and promote a culture in which personal information is protected and respected, has concerns.

AI presents new privacy risks for collecting, using, and disclosing personal information, a spokesperson for the commissioner says. These risks are heightened when it comes to sensitive personal information such as health information.

“Collecting health information from personal devices may be a useful way to support improved health outcomes. But

“Healthcare has got a massive data problem. It’s disorganised, first and foremost. It makes up approximately a third of the world’s data, and it’s growing faster than any other data type in the world. Medical knowledge is expanding at a rate that has never been seen before, it now doubles essentially, every month.” Brad Porter, right

CEO Orion Health

it could come at the expense of individual privacy – these personal devices collect a lot of personal information, often more than people realise,” the spokesperson says. “Agencies must ensure that they only collect what they need. Openness and transparency are a central pillar of the Privacy Act, and we encourage all agencies to be open and transparent about any AI tools they use.”

There are further issues around accuracy and transparency. “Sometimes people do not have a full understanding of what they are signing up for. It is important that businesses are very clear, in an easy-to-understand way, what personal information is going to be collected, why, and how, so they can be fully informed about the choices they are making.”

Orion has unified records for more than 70 health systems across the world covering a population of 150 million – data can range from hospitals, primary care practice, specialists, and community care clinics.

AI insights can identify earlier treatments, reducing how many people end up in hospital and prevent tying up valuable GP resources.

“The WHO have forecasted the world will be 10 million healthcare workers short by 2030. We can't train that many people. So we’re going to have something different,” says Porter.

“It is a slam dunk. We’ve got data with existing vendors across the healthcare ecosystem that are ready to go in sharing this type of data and we should be moving quick. We can’t afford to wait.”