How AI can help Doctors ‘care’ again
Dr. Eric Topol’s 2019 book “Deep Medicine” is remarkable for many reasons. This is the 1st of a series of articles devoted to discuss the wonderful book. In this first episode, I’ll simply give you 4 reasons why I love “Deep Medicine”.
Its core theme is humanistic
Talking about cutting edge tech is fine, but amidst all the sophistication, what are we trying to accomplish here? The ultimate objective of any digital health implementation, including AI, should be to improve the Health of your consumer. That’s what Dr.Eric Topol, a leading digital health researcher himself, highlights from the word go. In fact, the tagline of the book is “How AI can make Health Care human again”. It’s a strong claim, and at a glance, may seem contradictory to some. But the author builds up a strong argument about how Artificial Intelligence can help restore the time-honored human to human bond in health care and why we don’t have to worry about machines replacing humans, but should rather dream and strive for a future where both work hand-in-hand.
It gives a startling account of “Shallow Medicine”
The author gives a specified plan on how to restore the “care” in health care and calls that framework “Deep Medicine”. We’ll get into that in a moment, but the opposite of that is what he calls “shallow medicine”. He cites evidence from other researchers to explain this: in the USA, the Health care has grown to be the biggest business in the country; per person expenditure per year has grown from USD 550 to USD 11,000. But during the same period,the amount of time a doctor spends with a patient has reduced from 60 minutes to a startling 12!(in a first visit); and it’s not just a USA problem, the author highlights that many other developed countries face the same scenario. Patients are not the only victims; about half of the doctors suffer from burn-out, and the suicide rates among doctors are in an all-time high.
But the most harrowing of it all is the author’s own experience described in painstaking detail. Despite being an “insider” in the medical field and a much-respected figure for being a leading medical researcher in the USA, Dr.Topol had to face the dire consequences of Shallow Medicine first-hand. He was misdiagnosed, mismanaged, suffered and almost incapacitated. I will leave you to read the book without spoiling it all, but this shocking account of how the lack of individual attention and caring for the sufferer, not the suffering, could lead to dangerous and potentially irreversible consequences, brings home the importance of human-to-human bond; the ‘care’ in healthcare. It will not just make your patient happy, it will help you avoid misdiagnosis and mismanagement, and provide the best possible care for the individual in front of you.
So is this all just an anti-capitalist rant? It’d be that if not for the solution the author suggests in great detail in the next chapters. Let’s see what it’s all about.
The realistic promise of Deep Medicine
The author’s solution to the problem of Shallow Medicine is the promise of Deep Medicine. It is a vision, a framework of three interconnected parts.
First one is “Deep Phenotyping”, it involves, in the author’s own words, “Digitizing the medical essence of a human being”. In simpler terms, it’s about having the data about an individual from her days in the womb to present, as much as possible. Data captured should be as long as possible: covering much of the life span, and it should be as “thick” as possible: having a variety of data from each stage of your life. This data can then be used to make timely and accurate recommendations.
Second component of the Deep Medicine framework is “Deep Learning”, which is a newer type of Machine Learning(ML). In ML, a machine can learn and improve with experience and make better decisions based on patterns, without explicit instructions being given. Deep Learning differs from other traditional ML methods because it is designed to mimic human neural connections in the brain; it can identify more complex patterns and sometimes make more accurate recommendations. With exponential improvements in computing power, its accuracy of interpreting data has improved significantly over the years, racing past human threshold. But all this will depend on the quality of data a deep learning model is “trained” with.
Third part of the Deep Medicine framework is “Deep Empathy”; it is the important human touch we’ve been talking about. You need a caring, attentive physician to tailor the evidence to the individual in front of you, and make the decision with the patient, not for her.
So you’d have realized by now, that these three are interconnected. Imagine a situation where you are faced with a rare condition without a proper initial diagnosis and your treatments are not working. But you have “deeply phenotyped” data at your disposal(including genomic, medical, social and family history), you feed that data to a Machine with AI capabilities, and it runs through the data in seconds and suggests you the possible diagnosis and the best treatment; now you, a savvy and caring physician with a good understanding about the individual in front of you, decide on the best possible course of action, and your treatment works!
This is the solution the author proposes, and before you think it’s some futuristic nonsense, read the remarkable anecdote involving a child admitted to Rady’s Child Hospital in San Diego; a child with repeated fitting attacks resistant to the usual drugs, recovers in minutes with a simple dietary modification thanks to the beautiful execution of “Deep Medicine”. It will take some effort to make this a reality world-over, it will take not only time, but effort and intention; as the author himself writes, it will be a “marathon without a finishing line”.
It’s has an awesome collection of cutting edge research in Digital Health
Deep Medicine is remarkable for its vision, but the book also serves as a collection of the most cutting edge and remarkable AI research in health in recent times. The book was published in March 2019 and discusses AI tools and products mainly developed through Deep Learning methods. The areas range from Radiology, Dermatology to Mental Health and Nutrition. Products range from wearables that can predict and warn about your heart rhythm abnormalities to tools that can recommend suitable meal plans by analyzing your stools. While many of them remain to be clinically validated, they hold the promise of a brighter future for Healthcare.
I promise you to discuss these awesome inventions and how they fit into our Deep Medicine framework in future articles. Until then, I will share with you the link to purchase the book here. (This is not an affiliated link, I just shared with you the place where I came across the book.)
Have great day then :)
