IBM’s Watson Health about ‘augmented intelligence for clinicians’

IBM’s Watson Health is at the cutting edge of AI in Global healthcare, chief health officer Dr Kyu Rhee tells Digital Health’s Owen Hughes the increasingly powerful tools becoming available will augment clinicians’ knowledge not replace it. 

Artificial intelligence currently sits at the very top of the technology hype cycle, as industries look to the power of machine learning to make themselves better, faster and more efficient.

In healthcare, the opportunities are myriad. Not only could AI support decision-making and take some of the strain off overstretched staff, but also accelerate the discovery and adoption of new treatments and potentially usher in a new era of person-centred healthcare focused on ‘people’ rather than ‘patients.’

IBM has been a driving force in the medical AI space since 2015, when the company established Watson Health in an effort to bring the benefits of what the firm describes as “cognitive computing” to the medical world.

Digital Health News sat down with Dr Kyu Rhee, vice president and chief health officer at IBM Watson, as well Andreas Haimböck-Tichy, the business’s director of healthcare and life sciences, to discuss the opportunities offered by AI in the UK health and care market.

Augmented not artificial intelligence

“We view AI as an acronym not only for artificial intelligence, but also what we call augmented intelligence,” explains Rhee.

“We don’t think of it replacing the oncologists or the radiologist –  we think about the AI supporting the oncologist, for instance,” he says. “It’s the humans that make the decision.”

Watson Health’s oncology suite, one of its best developed clinical applications, is currently in use with over 200 healthcare organisations worldwide to support decision-making in cancer care.

Watson for Oncology, an AI-powered treatment decision support system for oncologists, is trained by New York’s Memorial Sloan Kettering (MSK), a leading global institution in cancer care.

Rhee, an internal medicine and paediatrics doctor by training, has served as Watson Health’s physician leader since the start of Watson Health in April 2015, having previously held leading roles at the US’s Health Resources and Services Administration and the National Institutes of Health.

He explains that international deployments of the solution can be tailored or “localised” according to the needs of that particular market, while still drawing from the single MSK knowledge base.

“We’re discovering that there’s sometimes discordance between what Watson recommends and what the oncologists or group of oncologists recommend.”

“Some of that discordance is related to the fact that what Watson recommends is not locally available – it’s just not covered by the government or the health plan and, therefore, we have responded to requests to have it localised, where the local recommendation is highlighted as a choice.”

Watson Health – named after former IBM CEO Thomas J. Watson and not Sherlock Holmes’ trusted sidekick – is built on a secure cloud providing access to a shared set of APIs and AI technologies, amongst them sentiment analytics, natural language processing and image recognition. Third parties can use these to create tailored applications that leverage Watson Health’s underlying capabilities

Digital assistants for digital patients

In the UK, these tools have to date been used by Alder Hey Children’s Hospital and Arthritis UK to develop intelligent digital health agents that provide personalised support to patients.

“What we’ve done in the case of Arthritis UK is basically take some of those [Watson Health] components to allow people to interact with the charity at any time of the day,” explains Haimböck-Tichy.

“In the case of Alder Hey, it’s their development, but they’ve used our platform to create a way for children and carers to interact with the hospital.”

Analytics for driving personalised care

The London Borough of Harrow is also using IBM Watson technology, in this instance to provide tailored care plans to citizens.

Implemented in 2016, the IBM Watson Care Manager platform pulls information from local care organisations, so individuals can select the provider most suited to their needs and budget.

As a result, people can select treatments in line with their treatment plans, as prescribed by their social worker or by the healthcare system, explains Haimböck-Tichy.

“They can choose from a catalogue of offerings, rather than from a list of what would usually be somewhere between 30 and 50 providers which their local council has negotiated with,” he adds.

The UK has found itself at the vanguard of artificial intelligence development in recent years, largely thanks to widespread enthusiasm from the UK government, which has rallied around the technology with political incentives that aim to make the country a leader in the field.

Looking at the wider prospects for the UK healthcare industry, Haimböck-Tichy suggests there is a unique opportunity to use the NHS extensive pool of primary care level to drive analytics and –  potentially – new approaches to care.

UK leads world in primary care data set

“While we’re sometimes concerned about the maturity of digitisation at the acute level, at primary care level, the UK has got one of the best data sets in the world,” continues Haimböck-Tichy.

“I think there is great opportunity to see how we create insight out of that data set for the primary care physician.”

Or course, the true value of AI rest on its ability to dip into – and drive insights from – a variety of health, social and environmental data, and not just that which sits within clinical IT systems.

Says Rhee: “There’s typically clinical data, genomic data, social and environmental data and behavioural data – there’s lots of different types of data sets, and the science has told us that the determinants of health are connected to all of these.

“What I would suggest has been challenging as a healthcare professional is that I often would focus on clinical data set and trying to scour it for insights. When in fact, I learned that there are so many other data sets, like the social determinants of health, can play an even bigger role in my ability to have an impact on the health of a patient or a population.”

‘Trust is about understanding what the AI is’

For Rhee and his team, unlocking these additional information resources will present the opportunity to transform healthcare from a reactive to a proactive model. However, he points out that this means first instigating a culture of trust and openness in those who hold the data, as well as trust in ‘the machine’.

“This is part of our challenge – to find ways in which these different data sets can be shared and trusted to get a more holistic view of a person and a population.

“A lot of the trust isn’t just about adding AI and making it better – it’s about having people understand what the AI is.

“Everyone wants to innovate and disrupt, but very few people want to change. This new era of AI will require us all to change in order to leverage it and work towards delivering better health outcomes for patients, populations, and the broader health system.”