AI prescribed to boost population health and support clinicians
Artificial intelligence and machine learning are transforming all sectors of the economy, and promising enormous human impact — nowhere more than in health care. McKinsey & Co research suggests that harnessing AI could deliver 44% greater value to health care than more traditional analytics techniques — and that globally the impact of AI on health care is approaching US$400 billion.
In November this year AIMed will hold its first summit in Australia — only the fourth such event anywhere in the world — bringing together 400 delegates and 30 speakers, alongside important opportunities to network, take part in workshops, hear from leading-edge practitioners and attend a special ‘Shark Tank’ session to learn about groundbreaking local innovation.
Speaking ahead of the summit at an AIMed briefing in Sydney, AIMed chairman and founder Dr Anthony Chang stressed the need for AI-enhanced medicine. He said that the best results would arise when machine intelligence combined with clinical human intelligence into a new brand of medical intelligence that would in the future be embedded in all manner of medical services.
AIMed was founded in 2014 by Dr Chang — a practising paediatric cardiologist and the chief intelligence and innovation officer at the Children’s Hospital of Orange County. AIMed’s intent is to bring together clinicians, physicians, C-suite executives and technology experts to accelerate progress towards AI-infused health care, and particularly precision medicine and population health.
AI triggers transformation and disruption
Application of AI to health care will be limited only by the imagination — but early use cases have concentrated largely on disease diagnosis and improved care. Successful solutions weave together rich datasets incorporating image and video from diagnostic tests, overlaid by specially developed machine learning algorithms that are able to interpret the data and provide support to consumers and clinicians. At the AIMed Sydney briefing in August, hosted by Best Case Scenario and PwC, delegates were reminded by PwC technology ethics specialist Richard Kelly that just as trust was a key element of the Hippocratic Oath, so trust needed to be instilled into any application of AI in health care.
“The adoption of AI into Australia’s healthcare system requires the shedding of legacy thinking,” Kelly said. “Moving from a reactive healthcare system that treats patients at the point of crisis to one that is proactive, preventative and focused on ongoing wellness requires the coordination of infrastructure, platforms, devices, data and personnel.”
PwC is optimistic about the impact that AI will have across the global economy and has forecast that global GDP will be up to 14% higher in 2030 due to the accelerating development and take-up of AI. In health care it believes that the impact will be both transformative and disruptive. In a recent report it noted that AI has the potential to both improve healthcare service delivery and lower the per capita cost of managing a growing and ageing population.
The company also warned: “Those incorporating AI should never lose sight of the fact that technology best supports human capabilities, rather than replaces them, and that compassion, intuition and emotional intelligence will remain pillars of true health care.”
Holistic approach required for AI success
The pace at which AI is transforming health care is accelerating as computing power becomes more affordable, health-related data becomes more accessible and machine learning algorithms proliferate. Professor Enrico Coiera, director of the Centre for Health Informatics and lead of the Australian Alliance for AI in Healthcare, is trained in medicine and has a PhD in AI. During his presentation at the AIMed briefing Prof Coiera noted the rapid rise of FDA approvals for AI-infused healthcare solutions — one in 2014, four in 2016, six in 2017 and 23 last year.
“There is a rapid growth in certification and an explosion in commercialisation,” claimed Prof Coiera, who said that the US investment in AI for health care is tipped to reach US$6.6 billion by 2021.
In his view the greatest benefit will come not from even more technology breakthroughs, but from the judicious application of existing AI capability. That, he said, needs careful consideration. Prof Coiera offered the example of a well-regarded research paper which revealed that an artificially intelligent diagnosis system had proven equal or better than humans in identifying thyroid cancer. Rather than racing to replace clinicians with a diagnostic platform, however, Prof Coiera said practitioners needed to consider other important factors; although the rate of thyroid cancer diagnosis has risen dramatically, the death rate has remained static.
The question that needed to be addressed, he said, is not ‘Should we use AI to diagnose thyroid cancer?’, but rather ‘Do we need to treat all cases?’. AI could perhaps help determine the answer to that more relevant question.
Prof Coiera said that a better use of AI was to find a way to embed it into the ‘last mile’ of health care where it would deliver the greatest benefits to both the patient and the healthcare system. But he lamented the relatively slow progress of AI in health care in Australia. While the UK’s National Health System has earmarked a further £250 million to establish a national AI Laboratory, on top of £1 billion already committed to the area, Australia had nothing in the pipeline, he said. Australia also faces unique challenges associated with the healthcare responsibility divide between Commonwealth and states.
The lack of a national government-led AI in healthcare initiative had prompted the establishment of the Australian Alliance for AI in Health, according to Prof Coiera. The Alliance’s four flagship areas of focus are:
- Ethics, safety and quality
- Workforce reskilling
- Consumer health care
- Precision health care
It also runs a special interest group exploring issues associated with cybersecurity in AI-infused health care. The Alliance is an official partner for the AIMed Summit.
AIMed 2019 Shark Tank — the innovation challenge
One highly anticipated feature of the AIMed Summit will be a Shark Tank-themed event where innovators are invited to pitch their solutions to a panel of experts. Three challenges were identified at the AIMed briefing in August, which innovators are invited to respond to.
Chronic disease management
At present only one in three people with a chronic disease get effective care, often due to the inherent complexity associated with managing these conditions. For example, a 64-year-old woman who presents at hospital with an asthma attack may be given steroids and also have her underlying diabetes uncovered. Both are treated, but when the asthma is controlled and the steroid dose reduced, the team treating her diabetes is not informed and the high insulin dose is maintained, leading to further health issues that could have been avoided if the information gap had been closed.
The challenge medical hospitalist Haroon Kasim posed to innovators was: “Is there an AI solution? The potential impact is huge. In aged care one in two people are sent home with the wrong medications, there are 230,000 adverse effects and 2000 deaths — costing the health system $1 billion a year.”
He outlined the challenge at the AIMed briefing, noting that the problem lies around the complexity of the case and the lack of standardisation. While he noted that hospitals do have clinical emergency response systems, these have limits.
Personalised pregnancy support
Around 300,000 babies are born each year in Australia — and about 1% are stillborn or die in the first month. The challenge that Dr M Talat Uppal, obstetrician and gynaecologist, VMO Northern Beaches Hospital and Hornsby Ku-Ring-Gai Hospital, posed for innovators attending AIMed Summit is: could AI solutions be developed that would help reduce the incidence of deaths and pregnancy complications by analysing risk factors and providing a more connected communications framework centred on the woman, but including clinicians and allied health providers to reduce risk and improve outcomes?
“Perhaps use AI to bring together diagnostic information — can we quantify an individual pregnancy risk?” she asked.
Could that translate into something as simple as a wearable for a pregnant woman, or a unique code that would allow all the data about the pregnancy to be more easily accessed and assessed? Dr Uppal explained that continuity of care has already been shown to have positive impact on outcomes — improved connections and communications, including language translation of clinician advice may have further significant impact.
Aged-care intelligence
There is clear evidence that older Australians will use technology if there is a perceived benefit. AI could help Australians navigate the aged-care sector or lift the quality of care, according to healthcare specialist Dr Michael Costello. My Aged Care is the front door service to aged care in Australia; however, it spans three government departments. It typically receives 1.5 million calls a year and there are 246,000 new registrations each year. Australia currently has 1824 aged-care providers, in an industry generating revenues of $22.2 billion currently growing at 4.5% a year.
According to Dr Costello: “When people interact with My Aged Care they don’t know what they don’t know. Is there the opportunity to use AI to help people navigate the site?”
One of the revelations from the Royal Commission into Aged Care has been the quality of care provided by institutions. Would AI be able to provide an alert to aged-care providers when a resident requires additional health support? Early intervention could reduce the risk of patients being admitted to hospital later. At present there are 750,000 preventable hospital admissions each year in Australia — very many of them aged 65 and over.
The AIMed Summit will provide an opportunity for innovators to respond to these healthcare challenges, and also serve as a forum to help accelerate the progress of AI in health care. Part of the only global movement that is focused on AI in medicine that is designed by clinicians, the event will take place from 12–13 November 2019. Meet and hear from global and Australian speakers as they present a myriad of topics including machine learning, artificial intelligence, augmented and virtual reality, blockchain and cybersecurity, cloud AI, and internet of things and everything (IoT and IoE).
To learn more and register, visit the AIMed website. Early-bird rates expire on 12 September.
To listen to podcast interviews and recorded presentations from the August briefing, click here.
Originally published here.
Why more needs to be done to support home-grown innovations
Commercialising new medical devices or drugs is highly risky, extremely expensive and returns can...
Opinion: Securing the backbone of health care
Unified, reliable databases provide healthcare organisations with immediate access to...
In conversation with AEHRC's Dr David Hansen
Of all the innovations across CSIRO's digital health portfolio, there is one which stands out...