Cardiovascular medicine as a data science

Attendees at the 73rd Irish Cardiac Society (ICS) Annual Scientific Meeting and AGM 2022 heard the prestigious Stokes Lecture 2022, which was delivered by Dr Robert Harrington, Professor of Medicine and Chairman of the Department of Medicine at Stanford University, US. The meeting took place in the Radisson Blu Hotel, Little Island, Co Cork, between 6 and 8 October. 

The Stokes Lecture is always one of the most highly anticipated highlights of the conference, and this year Dr Harrington delivered a talk titled ‘Cardiovascular medicine as a data science’. 

Dr Harrington presented an overview of the development of computerised data from the 1960s onwards and provided some examples of initiatives in Stanford University to use data to complement care, and to serve as another tool in a doctor’s armament. He discussed artificial intelligence (AI) machine learning, neural networks and other technological data advances that can be used by doctors, and how these are being incorporated into different types of research, from large outcomes-based projects, to randomised clinical trials where unique data sources are used to answer clinical questions. 

“One of the things I want to stress is the importance for the Fellows to think about the years to come,” said Dr Harrington. “That includes to think about cardiovascular medicine not just as a biologic science. It is a biologic science – there is anatomy, physiology, pathophysiology, and so on. But where I think the great discovery is going to come in the years ahead is on the data science side of cardiovascular medicine. So I am hoping that I can encourage some of you to pursue some additional training in the data sciences; biostatistics, economics, epidemiology, computer science, math – whatever it is that interests you. But if you can get those two different sides of your brain to appreciate the quantitative side from the biologic side, and you can bring those together, then I guarantee you that you will do good work and you will have many opportunities.” 

Dr Harrington discussed what he described as “the skills needed for success in this new environment” and further defined the term ‘data science’. “It’s not just biostatistics, it’s not just epidemiology or computer science; it actually takes all of that,” he told the conference. “That’s a description that works really well, because you can apply data science to any field of human endeavour, and that brings together three things. First, there is the domain or content knowledge, and in this case that’s cardiovascular medicine. Then there’s computer science and information technology and how one thinks about moving not just vast forms of data, but really different and disparate forms of data, and that’s really what is different now. 

“People talk about ‘big data’ in terms of the amount; it’s the disparate nature of the data – the fact that you can bring clinical data together with imaging data. We used to think of those separately, but now we look at them in a combined way – that’s what is different now. And finally, there is the math, the computation and statistics,” Dr Harrington continued. “This is not new stuff – we have actually had the basic principles behind a lot of this for decades, but what’s different now is the fact that data is ubiquitous. Almost everything has been turned into data, and so there are large amounts of data. It’s very disparate, and yet [there is better] access to it in many ways in real time. 

“But the big advance in the last decade has been cloud computing,” Dr Harrington told the attendees. “Now, we have the ability to handle it, we have the storage capabilities to handle huge amounts of information that we never had before.” 

Dr Harrington elaborated on the relevance of data sciences in medicine and cited research on the topic. He told the attendees: “Only about 1 per cent of the time or less are any of us actually patients in the healthcare system,” he said. “The rest of the time we are living our lives but generating enormous amounts of data all the time” through wearable devices and home sensors or monitors, for example. “Why can’t we use that data to better human health?” he posited. “How do you take structured data, unstructured data, physical activity data, clinical data, and bring it all together in a way to inform human health? 

“At the end of the day, what’s special about medicine is the human element of it,” Dr Harrington concluded. “I like to tell out residents, ‘Don’t forget, medicine is not science – it’s a human art built upon scientific principals,’ and you can’t take the human element out of it…. It’s the notion of how we bring together human and artificial intelligence in a way that’s going to make healthcare and medicine better.”