15 March 2019
Dear Interested Readers,
An Introduction To Deep Medicine
Rapid advances in the information sciences, coupled with the political commitment to broad extensions of health care, promise to bring about basic changes in the structure of medical practice. Computing science will probably exert its major effects by augmenting and, in some cases, largely replacing the intellectual functions of the physician…
Change takes time. Ideas hang around for decades seeking to be fully appreciated and effectively integrated into the ways we think and work. The quote that began this letter was taken from the abstract of William Schwartz’s 1970 paper in the New England Journal of Medicine, Medicine and the Computer — The Promise and Problems of Change, You can tell it is not a contemporary quote. It is either from the distant past or a projection into a desirable future because of the phrase in the first line, “coupled with the political commitment to broad extensions of health care…” That is certainly not a picture of today’s political climate. In 1970 we did seem to have some sort of “political commitment to broad extensions of healthcare.” The consensus that created Medicare, Medicaid, and launched the HMO movement fell apart not long after those words were written.
Back in the seventies there was also excitement about the potential of augmenting our diagnostic skills and therapeutic techniques with our rapidly expanding IT capabilities, but then our computers were co opted for finance functions and doctors became the inputers of data that was more often used to justify a bill than to solve a clinical problem. When Schwartz wrote those words in the early seventies Steve Jobs and Bill Gates were both fifteen year old adolescents playing with computers and programing, but Octo Barrnet had already created an electronic medical record and it was in use at the recently opened Harvard Community Health Plan. In one paragraph Schwartz gives us notice of the possibilities that computers offer and then immediately advises us of the potential barriers to the use of computers in medicine. Schwartz continues:
As the “intellectual” use of the computer influences in a fundamental fashion the problems of both physician manpower and quality of medical care, it will also inevitably exact important social costs — psychologic, organizational, legal, economic and technical. Only through consideration of such potential costs will it be possible to introduce the new technology in an effective and acceptable manner. To accomplish this goal will require new interactions among medicine, the information sciences and the management sciences, and the development of new skills and attitudes on the part of policy-makers in the health-care system.
Did Schwartz foresee millions of clinicians spending 50% of their professional time demonstrating their poor “keyboard skills”to frustrated patients who were just hoping that they would turn around and give them some attention?
I am into a new book about the future of medicine, Deep Medicine: How Artifical Intelligence Can Make Medicine Human Again, by Dr. Eric Topol. It was Dr. Topol who reminded me of Schwartz’s paper and the conversations it generated so long ago before so many clinicians learned to hate their computers. The book has gotten significant attention and Dr. Topol has been making the rounds of interviews with newspapers and talk shows. I would recommend that you read the article “How Artificial Intelligence Could Transform Medicine” by Anahad O’Connor in the New York Times, and a listen to a recent “On Point” interview with Dr. Topol on NPR.
This is the third book that Dr. Topol has produced describing the problems and possibilities of practice with positive suggestions and predictions for the future. Dr. Topol is an interventional cardiologist who first gained a national reputation while he was at the Cleveland Clinics. A few years ago he moved to the Scripps Clinic in La Jolla where he is the founder and director of the Scripps Research Translational Institute.
As I began to read, it occurred to me that every book, and many of the policy papers that I have read over the last decade, follows a format that begins like a Lean exercise. The first step is a statement of the reason for action. Our healthcare is fragmented, expensive, available as a function of socioeconomic status, and even for the rich, is hard to access, and lacks quality and safety. Unfortunately, we have become numb to that description, or if concerned, we shake our heads in frustration because we realize that as individuals we are at the mercy of forces beyond our control. Some of us are depressed, and some are fighting mad like the physician who called in from New Orleans near the end of the “On Point” interview with Dr.Topol.
The better books about what might improve our future, like John Toussaint’s On the Mend: Revolutionizing Healthcare to Save Lives and Transform the Industry, Elisabeth Rosenthal’s An American Sickness: How Healthcare Became Big Business and How You Can Take It Back, or Robert Pearl’s Mistreated: Why We Think We’re Getting Good Health Careand Why We’re Usually Wrong, continue with an attempt to describe in detail the current state, and move on to their image or outline of an improved or ideal state. The best books discuss the barriers that must be overcome for real change to occur, but almost all of them leave us feeling vulnerable to the process of implementation. At the core of their common dilemma is the fact that the vested economic interests of those who profit from the status quo coupled with the fear of change that most health consumers feel, even as they complain about the cost and inadequacies of their care, creates a powerful coalition that is quite capable of overwhelming challenging ideas and innovations. If patients are apprehensive about changes to improve care that will change their relationships, their concerns are vague compared to the feelings of many clinicians who feel their world has been changed enough and making things “better” doesn’t necessarily mean making things better for them. Some say, “If I become more efficient, then ‘they’ will just ask me to do more.”
There are few human relationships that are more sacrosanct, and yet have generated more injury and disappointment than the “doctor-patient” relationship. For over a century, the AMA has successfully used the potential fear of innovation damaging care for patients and compromising professional autonomy as justification to resist innovative public financing efforts to achieve universal coverage. When the choice has been between change that might improve and expand care for patients, or resisting change to preserve the investments in business as usual, the doctor-patient relationship has been an effective shield to defend the status quo against the best ideas of change oriented policy makers.
The universal digitalization of medicine is at least twenty years behind telecommunications, finance and engineering. The interoperability of systems that is demanded of telecommunication and finance is non existent in healthcare. Many have pointed out that you can access your bank account from any place in the world that has an Internet connection, or that my iPhone can talk to your Android phone as effectively as it can speak to another iPhone, but if I want to seek care at Tufts Medical Center, The Beth Israel Deaconess Medical Center, or the Boston Medical Center, which are all less than two miles from either the Brigham or the Massachusetts General Hospital, forget it. The hospital systems do not effectively talk to each other even when most of them are on Epic! In the digital transformation of society healthcare is the dunce in the class. Why? Is it our fear of sharing what we know about the patient for fear that they will move their care? Is it that we lack the insight that the story and the data really belong to the patient? Is it that we aren’t so interested in moving into an era where we need new skills that will displace the old ones we have mastered?
Dr. Topol has tried to overcome the resistance to AI in medicine by educating us to what AI really is. He believes that AI is such a powerful force that sooner or later healthcare will be compelled to make it a central asset. He believes that it will touch every part of what we do. He does not write to defend the idea of computers assuming a larger role in healthcare. He writes to show us how AI is being, and will continue to be, woven into the fabric of all that we do. He writes to explain what AI is and is not. He writes to show us what AI will always be able to do better than we can do. He writes to show us what AI will never be able to do, and what we must continue to do, and do better with the aid of AI.
Until I read Dr. Topol’s description of AI, I had a distorted understanding of what it is, and what it can do now, will soon be able to do, and what it will never be able to do as well as a human. I am not an electrical engineer, or a computer scientist, but Topol has given me new information and context by going back to the work of Daniel Kahneman and Amos Tversky and discussing their thoughts on decision making and biases. He connects AI to Kahneman and Tversky’s work on the “fast and slow” decision making processes that are associated with practice. The vision that ultimately emerges is that the fifteen minute appointment that is now thirteen minutes with the computer and two minutes with the patient gets flipped to two minutes with the computer and thirteen minutes with the patient. Topol is writing as much to demonstrate the inadequacies of the intellectual approach to practice that make “failure to diagnose” the most common malpractice complaint as he is to explain or sell AI.
Topol never comes out and explicitly says that the initial attempts of IBM to introduce “Watson” have constituted a setback for AI in medicine, but he does demonstrate many of the mistakes that have been made. He also underlines some of the successes of AI that are expanding now and that will continue to expand even as they meet resistance. Topol believes that sooner or later AI will transform medicine just as it is transforming the way we learn, how we work, how we shop, how we choose our entertainment, and how we don’t drive.
The most positive aspect of Topol’s presentation is his vision of how AI will bring us closer to our patients, and facilitate a deeper understanding of their issues and needs. “Deep Learning” is an AI concept that dates back to 1986 and describes the system’s ability to learn. Topol defines “Deep Learning” In his glossary of AI terms and concepts:
Deep Learning–a type of neural network, the subset of machine learning composed of algorithms that permit software to train itself to perform tasks by processing multilayered networks of data
By the way, a neural network is a “software construction modeled after the way adaptable neurons in the brain were understood to work instead of human guided rigid instructions.”
Topol gives me more information than I can process, and even gets philosophical when he gives us the concept that our lives are already defined by algorithms via our hormones, emotions, and the internal machinery that determines our responses to the external forces that impact our personal space. He pulls a lot of different pathways, examples, experiences, and his own biases and understanding of AI together to offer us a new concept, “Deep Medicine.” Deep Medicine defines the ways in which we can contextualize the benefits of AI for patients. I like to remember the days of the “Six Million Dollar Man” who was a cyborg who demonstrated the benefits a guy could have if he was given a few machine capabilities.
Three Components of Deep Medicine (adapted from the text of Deep Medicine)
1. The ability to deeply define each individual— all of one’s medical, social, behavioral, and family histories, as well as one’s biology: anatomy, physiology, and environment. In other words: “deep phenotyping.” Topol describes it a “thick and long.” it spans many types of data over long periods of time, “pre womb to tomb” or “lust to dust.”
2.“Deep Learning,” involving pattern recognition [In practice I could identify many of my patients as easily by their EKG as by their picture.] and machine learning including virtual coaches to better guide consumers to better manage their own issues. In the hospital it will improve safety and quality. It will reduce admissions by enabling programs of” hospital in the home.” We are in the infancy of these applications although we have known that they were possible for fifty years since Schwartz’s paper in the NEJM. Dr. Topol sums up this concept by saying: One can imagine that AI will rescue medicine from all that ails it, including diagnostic inaccuracy and workflow inefficiencies (such as mundane tasks like billing or coding charts), but none of these have been actualized yet. It’s an extraordinary opportunity for entrepreneurs working with clinicians, computer scientists, and researchers in other disciplines (such as behavioral science and bioethics) to help fashion the right integration of AI and healthcare.
3. Deep Empathy, is connection between patients and clinicians and is the most important component of ‘Deep Medicine.” Topol sounds sad and wistful when he writes: In the more than four decades since I started medical school, I’ve watched the steady degradation of the human side of medicine …Over that span of time, healthcare became not just a big business but, by the end of 2017, the biggest. It is now the largest employer in the United States, towering over retail. By every metric, the amount of money spent on healthcare has exploded. Yet, even with all the employment in the sector and all the money expended per person, the time spent between doctors and patients has steadily dwindled, whether for office visits or in the hospital. Doctors are much too busy….Consumed by patient care, physicians were passive while major new changes took hold in the business of healthcare, including electronic health records, managed care, health maintenance organizations, and relative value units. Now, the highest-ever proportion of doctors and nurses are experiencing burnout and depression owing to their inability to provide real care to patients, which was their basis for pursuing a medical career.”
I think Dr. Topol wrote this book because he believes that AI can do more than just improve care. He believes it can save us from the exhaustion and frustration that impairs our ability to connect with patients in a way that brings us the joy of professional fulfillment while preserving or restoring the health of patients. Topol really believes that if used correctly AI can help us reduce the cost of care while improving quality and safety.
Topol says that it will not happen all at once. The greatest opportunity now is with “machine pattern recognition.”
As Topol said in his interview with The New York Times.
…to promote the rapid and accurate reading of medical scans, slides, skin lesions, the pickup of small polyps during colonoscopy, and much more. Another is keyboard liberation, or using natural language processing of speech to synthesize notes and eliminate the ultimate source of distraction and dislike in medical encounters. A.I. can also predict key outcomes for both patients and clinicians to promote prevention…
What I’m most excited about is using the future to bring back the past: To restore the care in health care. By giving both the gift of time to clinicians, who are at peak levels ever recorded for burnout and depression, and empowerment to patients, for those who want it, this will ultimately be possible. But it will require substantial activism of the medical community to stand up for patients and not allow the jump in productivity to further squeeze clinicians, upending the erosion of the doctor-patient relationship.
I think Topol has come back full circle to where Schwartz was in 1970. There are opportunities and obstacles. Clinicians lost the first round of computers in healthcare to business interests that have made a lot of money selling us products that have not made the experience of care better for either the clinician or the patient. There are failures in any process of continuous improvement. Just as it is true with machine learning, progress is made when we use prior experience to facilitate deep learning and change a faulty algorithm. I sure would like to be around to see what happens in the next fifty years.
The End (Of Winter) Is In Sight
Daylight Savings Time has arrived, and has given me a preview of spring! There is enough light in the evening to finish a late walk without a headlamp or reflecting vest. The day is balanced evenly between daylight and dark. Even though spring will eventually arrive there is still plenty of snow on the ground where I live as you can see in today’s header, but the blanket is thinning. We had a dusting of snow on Wednesday night, but today it will be in the mid fifties with rain which should cause the sap to rise and may make it possible for a coccus or two to peak out from under their blanket of snow.
It’s been a great winter with enough snow to satisfy any lover of outdoor frozen delights, but nothing lasts forever and it’s time to think spring thoughts. My friends who continue the traditions of sugaring are anxious to see a little sap in their buckets so they can start “boiling.” Next weekend is the annual open house at all the local sugaring operations. In Boston the annual flower show is in full swing to give us an indoor preview of spring and summer delights. Down in Fort Myers the Red Sox are losing more games than they are winning, but none of them count. The games do begin to count in just thirteen days when the Sox open the season in Seattle on the 28th, and that is a good sign.
We are in transition. This time of year each day is an adventure. Will we have sunshine and warmth, rain and mud, or an encore snow dump? But, I am now living in anticipation of fishing, hiking, biking, and picnicking in a light jacket. It’s about time to put the fleece away for a few months and feel the sunshine on my shoulders. Wherever you are this weekend, I hope that the elements will allow you to take a walk in the sunshine looking for early signs of spring.
Be well, take good care of yourself, let me hear from you often, and don’t let anything keep you from doing the good that you can do every day,
Gene
