1 Aug, 2019

How Machine Learning & AI can Contribute to the Future of Healthcare

Last week on our Genetics Podcast, we had the honour of hosting Eric Topol, author of ‘The Patient Will See You Now’ and ‘Deep Medicine’.

Eric has had an incredible career which has been largely focused on researching cardiovascular disease and heart attacks, both of which he worked on in the Cleveland Clinic and at the Scripps Institute.

In the podcast episode, we discussed wireless medicine and the role of artificial intelligence and machine learning in healthcare.

How will medicine look in a decade from now?

We’ve just walked into a world where there is an immense amount of data for each of us. This is totally new in medicine. Now we’re able to measure genetic data, environmental data and even ‘biology’ layers of data like immunome data.

Because of this massive influx of new information, we can’t rely on traditional record keeping anymore. Therefore, a massive change for the future is going to have to be in data management. The ways we store, use and protect our information is transcending human capabilities and Eric thinks that this means a future where artificial intelligence and machine learning is central to healthcare.

Ideally, this could help us create personalised medicine for everyone. We could use machines to distill all of this data so that doctors can prescribe personalised treatments based on each individual’s needs. This will make healthcare faster, more accurate and more efficient, especially since they’re already likely to know which treatments will be more effective for them, because of all of the information provided by the machines. Obviously, this is incredibly different to the system we have in place today, where doctors have to sift through written records and prescribe the same treatments to most patients, with a sort-of trial and error approach.

How will machines affect doctors?

In the best scenario, machines will be used not to replace doctors but to empower them. The capabilities of AI in regards to data handling far surpasses that of any person but healthcare still requires a human touch. This is roughly the topic for Eric’s newest book ‘Deep Medicine’. In it, he talks a lot about how artificial intelligence can transform medicine by giving the gifts of time, accuracy, speed and efficiency while allowing doctors to reestablish the trust, empathy and communication that seems to have been lost over the years. Eric says that it is of the utmost importance that healthcare has a human factor and unfortunately, the way we do medicine at the moment just doesn’t seem to. Using AI means that doctors can spend less time typing on a computer and more on person-to-person interaction, putting the care back into healthcare.

He goes on to stress that It’s crucial we handle this opportunity in the right way. If we try to put more pressure on doctors because machines effectively provide them with more time, we could mess this up entirely. It’s going to remove all the benefits if we start expecting that doctors will suddenly be able to do ten times more work, ten times quicker, because machines are sorting information for them. Giving doctors more time will give them the chance to provide personalised care with a human touch, which is so important in healthcare.

At this moment in time, it’s going to be pivotal how we use this opportunity; do we seize it to bring back that crucial human bond that’s largely been lost or do we waste it and put business over people?

What will be the first steps to mark this transition?

We’re already seeing medical scans being pre-read by machines, which is a massive step in the right direction because in the past, over 30% of these scans have had a false-negative or missed finding by doctors. Using machines means this rate can be dropped to much lower, so it’s already been getting a lot of uptake in hospitals around the world.

In the same regard, Eric says that he thinks we’ll see much more of this in other parts of the healthcare sector like dermatology, ophthalmology or pathology, because of the accuracy and speed it provides. Even more recently, machines have been used in colonoscopies to pick up small, possibly precancerous, polyps at an extraordinary rate. These polyps are often missed by doctors and leaving them to go untreated could lead to serious consequences.

So it looks like machines are already making their way into our healthcare systems, but at different rates around the world. Estonia is leading in terms of digitization and data ownership, China is implementing AI into many areas of their hospitals and the UK is paving the way for voice to text instead of doctors typing. These are all super important movements for the future, China particularly because their hospital environments are very dissimilar to ours in the UK or US. They can often care for 10-20 thousand people and so really need to rely on AI and machine learning in order to provide care for all of these patients. On the other hand, what the UK is doing in regards to talk-to-typing is groundbreaking because it’s really bringing back that human factor which is vital to providing great healthcare. Eric goes as far as to say ‘everyone talks about going paperless but we need to go keyboardless. Keyboards are the first problem when it comes to getting eye-to-eye contact and good communication.’

Upcoming changes to be excited about

Right now, the UK is leading the way in rare disease diagnosis. They’ve made a big investment in the UK biobank and that’s given research a big step up. We’re now learning so much, so fast because of this incredible resource.

There is also a division of the National Health Service (NHS) that supports Education England. This includes training medical professionals about genomics and that really helps get it into the healthcare system. Eric completed a review on the NHS and says that he got a really deep sense of commitment for genomics - which will be significant for the future.

Right now, we have so many more layers of data than ever before, which means we’ll soon be able to provide personalised healthcare based on everyone’s unique information. Whether that’s their genetics, their environment or even other factors, it’s going to be a game changer for medicine.

In addition to this, Eric says that Polygenic Risk Scores are something that is very ready to be used in common practice. They provide complementary, independent risk scores based on genetics that can help people make active changes to their lifestyle to prevent or prepare for certain conditions they may be more likely to develop. They’re particularly useful for things like heart disease because heart health can be very influenced by lifestyle changes.

What's standing in the way?

Unfortunately, there’s an unwillingness to use tools like Polygenic Risk Scores because medical professionals are unsure how patients will react to being told about their risk for certain conditions. They don’t know whether it will change their lifestyle or if it will have a negative impact on their mental wellbeing. However, Eric said there has been a Finnish study where a large group of people were told they were a higher risk for certain conditions and overall they made positive lifestyle changes, like stopping smoking or losing weight in order to reduce their risk. That’s why Eric believes we should be making the most out of this research and giving people the opportunity to do something about their health.

Not long ago, medical professionals thought the same thing about at-home pregnancy tests; they didn't believe women could handle being told they were pregnant without a doctor present - this is of course, in most cases, untrue. Information is a lot like medication in that way, on some people it will work really well and some it won't but if the net outcome is positive, then it’s something we should be putting out there.

Another huge thing which Eric feels needs to change is that currently, people don’t own their own data. If everyone had their own data, and no one is more entitled to own their data than themselves, then they could share whatever parts of it they needed to with those they trust. Obviously it’s vital that data is shared to provide high-quality care but we should be letting each individual make that call.

Eric believes that we’ve been suppressed by medical professionals under a ‘doctor knows best’ mentality but we’re all just people in the end and he says that there’s no special reason why doctors should own your data. He goes on to say ‘It may be their note but guess what? It’s my body.” It's really time we changed attitudes around this to move forward.

Stay updated with Dr Eric Topol

If you want to stay updated with Eric you should definitely follow him on Twitter @EricTopol, he posts a goldmine of information and great book suggestions - so don’t miss out!

Follow us @sanogenetics

Discover the world of genetics

Join our community to learn more about your health and contribute to the development of medical research.

Sign Up

Related blogposts