“Hospitals as we know them today will eventually be extinct.” – Eric Topol
You’d be forgiven for thinking that Topol is referring to a distant future where lifelike robot doctors tend to our every medical need, but he’s not. The extinction he refers to is much more imminent, in fact, it is already well underway.
I came to this conclusion after attending a breakfast meeting last Thursday at the H2 Health Hub in Vasastan titled “Artificial Intelligence in Healthcare”. It was organised by Johan Gomez de la Torre from Stockholm Science City and chaired by Henrik Ahlén from Alfa Bravo with support from Stockholm Innovation and Growth. After a mere 2 hours of talks, I was forced to completely rethink how I view the future of hospital care. Put it this way, have you ever carried out surgery in complete darkness? I mean full-on, blackout curtain and eye mask for good measure darkness. I’ve got a feeling that that is exactly how the state of current medicine will be remembered.
Artificial intelligence, or AI as it is commonly known, might seem like a relatively new concept but it started back in the 50s. The reason it has become so popular now is data, explained panelist Daniel Gillblad from RISE SICS. Now we have machine learning technology coupled with vast amounts of data. We also have powerful computers in everyone’s pockets connected to even more powerful computers in the cloud. It’s a perfect storm that has lead to AI being used and woven into the fabric of almost everything we do.
To give an example of how ubiquitous AI is, if you’ve ever used an online tool to connect with your friends on a site called “Facebook” or “Twitter” you’ve used AI. If you’ve ever used a program for searching the web called “Google” you’ve used AI. What does that have to do with health you might ask? AI today means machines searching through vast oceans of data and finding meaning. This is a huge leap forward because it is something that humans simply cannot do.
Imagine a 5 by 5 spreadsheet filled with data, could you spot a trend? How about 10 by 10? Surely not 100 by 100?
With AI, computers can spot trends in spreadsheets with billions of rows. When we apply this to human health, the benefits are huge, especially when you consider new devices that are coming on the market to provide the computers with data on our health, generated by our own bodies.
Examples of AI being applied today
Take, for example, Coala. Philip Siberg and Titti Lundgren represented the company that makes an AI heart monitor. They presented a rather alarming statistic that while we can predict heart disease in a large number of cases based on epidemiological risk factors, up to a third of at risk people cannot be identified by current methods. Their solution to this is a non-invasive device sits on your chest and monitors your heart while sending data to the cloud which can be analysed by AI. The results of the AI analysis are then returned to your phone, and your doctor if something is wrong. Coala say that the idea is to identify patients before they have a heart attack and intervene. This product is available to buy right now from their website.
Another example is the company Aifloo presented by Arun Thorsteinsdottir. Aifloo make armbands that monitor elderly people in their home. Using an array of sensors, it can tell if users have fallen, or even if they are walking more unstably and are more likely to fall. All of this data is monitored and analysed in real time and sent to care staff who are on hand to help. For Aifloo, this is just the tip of the iceberg. With the data they are continually collecting from users of their armband, they believe they could find markers for other diseases. For example, if using the data from the armband, they are able to predict if users are missing meals, this may be indicative of memory problems from an underlying neurological disorder.
The funding body Vinnova are very enthusiastic about these developments. They are allocating some of their capital to digital health projects, including AI and have an online quiz to test if your project fits their funding goals explained Linda Swirtun. AI-enabled diagnosis is the future of medicine and Vinnova are encouraging more companies and research groups to take part.
Do we even need doctors?
So if AI can do everything, do we even need doctors? Well, as Daniel Gillblad says, AI cannot do everything. It can spot trends and learn to do impressive things, like drive cars. But, it cannot explain why. In medicine that is a big deal. We need to know the causes of diseases in order to treat them properly and discover new cures. So there is still a role for the human doctor after the hospital extinction, however, they will have access to oceans of data that are currently unavailable today. This is why I believe we are doing surgery in the dark. Imagine walking into the doctor’s office 100 years from now. They know everything about you. To start with, your genome sequence along with any hardwired disease risks and treatment options. Then they know what you have been eating, if you have been smoking, how well you have been sleeping. They know if any of your internal organs have been under stress or if your blood contains any strange biomarkers. They knew all of this, before you walked through the door. Is this going to change medicine in the future? I would say so.