Published 16 October 2019
Tim M. Jaeger talks about the future of healthcare, the power of knowledge and why human interaction will remain indispensable in the age of AI.
Software that assists physicians with clinical decisions is becoming more and more important. Why?
Let me start with saying that clinical decision support has always been an integral part of Diagnostics. Every test result, from simple blood sugar measurements to complex genomic analyses, is intended to help doctors decide what to do. What’s changed, though, is the unmanageably huge volume of health data that doctors have to navigate their way through. We need cloud-based software solutions that help doctors use the data intelligently and make confident treatment decisions.
The IBM Watson supercomputer hit the headlines in recent months because its treatment decisions were often seriously wide off the mark or failed to come up with anything new. Is the potential of AI in medicine being overestimated?
The most important thing about AI and machine learning is knowing exactly how and where to apply it. We have a profound understanding of disease, diagnostics and treatment, and we search datasets very specifically for actionable patterns. We can use the potential of AI and machine learning to augment what doctors and researchers do. That combination of artificial and human intelligence, of knowledge and experience, is always going to make the difference. Even in the future there will be cases where doctors rely on their gut instinct.
So there’s no hype?
Certainly there is. Hype typically follows a cycle: huge expectations, thousands of experiments, hundreds of failures, and then the mood shifts among the broader population. But approaches that actually work also emerge, and it’s our job to find these approaches as soon and fast as possible. There is no doubt that AI, machine learning, big data and analytics are set to become a fixture in the research and development of diagnostics as well as in treatment and patient care.
The combination of artificial and human intelligence, of knowledge and experience, is always going to make the difference.”
Let’s look ahead to the future. How will data science transform healthcare?
I think the future is going to be a lot more patient-centred and data-driven, and that we will be rewarded more for the value that we add – whether this be a better, earlier diagnosis or an improved treatment that delivers longer survival or an enhanced quality of life. That’s what we are going to be paid for. And I think this ecosystem will increasingly blur the boundaries between research and development on the one hand, and care delivery and practice on the other. This will also open up the boundaries between pharmaceuticals and diagnostics.
Let’s assume that modern prognostics live up to their promises, and we can predict which diseases we’re likely to get long before we get them. Is that actually desirable?
That kind of decision is very much up to the individual. Society’s task is to give each of us the information we need to understand the consequences of knowing or not knowing something. We can help to make this knowledge available and – via the treating physicians – to highlight the treatment options. After that, it’s up to each individual patient to decide what to do with what we provide.
Speaking for yourself, wouldn’t you want to know if you were going to have an incurable disease such as Alzheimer’s?
Yes, I think I would. But that’s partly because I’m a medical practitioner and have – or could acquire – the knowledge to understand exactly what that means for the life I’m living right now. Our task is to make this knowledge of medicine available to laypeople so that they don’t panic or have interventions prematurely that may ultimately not prove beneficial anyway.
A couple of speculative questions to end with: Do you think that one day we will be able to use apps to help determine our moods and emotions?
No, the expression on someone’s face – or that person’s body language – tells us a lot more than apps that supply whatever information on their emotional state. I don’t think the interpersonal element of human interaction is going to be lost.
Will the day come when robots are more intelligent than humans?
Intelligence is an extremely wide-ranging concept for which there’s no uniform definition. While robots will be able to process information in a different way from humans, they will never really be able to replace us.
Will Homo sapiens still be around in 500 years?
I very much hope so! And I hope we will be more homo – i.e. human – and more sapiens – i.e. know more.
Tim M. Jaeger, M.D. is Global Head of Diagnostics Information Solutions at Roche Diagnostics