Bringing Humanity Back Is the 'Greatest Gift AI Can Give'

; Abraham Verghese, MD


January 02, 2020

This transcript has been edited for clarity.

Eric J. Topol, MD: Hello, this is Eric Topol, I want to welcome you to a special episode of Medicine and the Machine. My cohost, Abraham Verghese, and I held our first live podcast episode at Exponential Medicine in San Diego this past November. Here it is now, available to all our listeners.

We're thrilled to be here for our first live podcast. It's a challenge to start a new podcast with all the podcast mania, but hopefully this topic is one that everyone here can identify with.

Abraham, it is always so much fun to be together. Let's talk a little bit about how we came together for this podcast.

Abraham Verghese, MD: I think people assume that you and I are on opposite ends of the pole—that you are the futurist, which you are, and that I'm somehow this Luddite. Neither of us knew that was true of each other, but when we met, we certainly discovered that it was not true. As much as you are a futurist, you are a very traditionally trained and traditional clinician celebrating human values. And as much as I write and talk about humanistic medicine, I'm also in the crucible of technology at Stanford and Silicon Valley, where I embrace it all. I'm carrying an ultrasound, but I don't use it as well as you do. When we first met, there was this moment of recognition that something is unchanging about medicine, and both of us deeply subscribe to that aspect of it.

Topol: If you're not aware, Abraham received a National Humanities Medal from President Obama at the White House in 2015. I don't think there are too many other physicians who have been noted like that. Not only is he my favorite humanist, but he is also my favorite physician author. Three and a half years ago, I underwent a knee replacement with lots of complication, and I wrote an essay for the Washington Post about the experience and to teach others. Before sending it in, I asked Abraham to give me some ideas about it. What was notable is that he cracked the case, in that not only was it a physical therapist who helped rescue me several weeks after I was in such turmoil, but that she actually cared. He brought up the famous Francis Peabody quote.

Verghese: "[T]he secret of the care of the patient is in caring for the patient."

Topol: Right. The point here is that we don't believe that artificial intelligence (AI) and medicine are at all competitive. That is why the podcast is called "Medicine and the Machine." We need the caring, we need all the aspects of the humanism in medicine, but we also would like to get help. And that is not really in the construct of what a lot of people think with respect to AI, where you keep hearing about how radiologists will be replaced and how all sorts of disruption will occur. In fact, our thoughts are very different where we see the leaning on machines and outsourcing of many tasks to patients who are eager to take on more responsibility. There has been a lot of erosion of the human aspects of medicine over decades. What is your perception about that?

Erosion of Human Aspects of Medicine

Verghese: We're extraordinarily lucky to be living in the day and age that we are. I was attending this morning's session and was just stunned by what is there and coming down the pike, and also recognizing that each of these will present its own unique ethical and unanticipated challenges, as people have pointed out. The one that you and I relate to the most is the electronic health record (EHR). On one level, it seemed like the most natural and brilliant thing to do was to have EHRs, and I certainly don't want to go back. But the one or two big systems that are in use are really epic failures, if you pardon the pun. We have the situation now where we are the highest-paid clerical workers in the hospital. We're spending 1 hour on the computer for every 1 hour cumulatively that we spend with the patient. That single invention, and the fact that we are chained to that machine, has been responsible for more distress than almost any other thing that has happened in modern medicine, quite apart from the pace of medicine and the data. So like almost everything that comes down the pike, it requires reflection and anticipation. Parallel to the challenge of developing new technology, it's a challenge to address the human aspects of what the technology means for the patient and for the workforce.

Topol: The EHR has been a big part of the burnout story. Of course, there is also clinical depression, and even the highest rate of suicides throughout the medical profession in history. We cannot attribute that all to EHRs, but part of it is the inability to provide care. Not only are we saddled to this keyboard and data clerk function, but there is very little time with patients that is true presence. You built a whole program about presence at Stanford, so maybe you can talk about that.

Patient Presence

Verghese: I was struck by the fact that when you talk to patients, the one thing they complain about is that we are not present with them. We have one eye on the keyboard and are distracted. Ironically, when I talked to physician groups, they complain about not being allowed to be present. We began this center at Stanford called Presence, and one of our projects has been to try and hone in on that first moment when a patient meets a physician or physician meets a patient. There are thousands and millions of first meetings between two strangers in this context happening every day. One thing we wanted to figure out was, how can we best make that first moment happen well enough so that everything following happens well too. Milan Kundera, the novelist, says that the first moment between a man and a woman predicates everything that is going to happen. I don't know if that is necessarily true, but I do think it's true with the patient-physician relationship. If the connection isn't ideal, then everything can fall apart, from compliance to even hearing what is being said. We've been trying to study that, and it represents just the kind of thing that has huge effects, but isn't necessarily sexy for technology.

A Million Strong, but Not Unified

Verghese: I want to turn back to you, Eric, because part of the problem has been the loss of autonomy. We have not been able to dictate our work style; rather, it's been dictated for us with how many people we should see, and so on. Eric wrote a wonderful article in the New Yorker not too long ago, talking about how we're a million strong and yet we don't speak as one voice. Would you tell us a little more about that?

Topol: [The article was published in] August and is called "Why Doctors Should Organize." Basically, it's about the fact that we have administrator overlords in medicine, and their growth over the decades from the 1980s on has been 3200% relative to 150% in the physician workforce. They don't take care of patients, they are not involved in any patient care, but they are ruling the roost with respect to more productivity, more efficiency. And all the things that have happened over these years that include relative value units; health maintenance organizations; and, of course, the EHRs have been supported by the administrative side, not the physician side. Frankly, there has not been a voice for physicians because the largest organization, American Medical Association, has barely over 200,000 members from among the million-doctor workforce in the United States.

The other problem is that all the other professional societies are much smaller, with the exception of the American College of Physicians, and they are balkanized so they are all subspecialties or disciplines. And for the most part, they don't really stand up for patients, but rather they are trying to ensure reimbursement for physicians and the business aspects of medicine. The term "trade guilds" comes to mind for many of these societies.

We need a way to stand up to our overlords, who are essentially making this problem very serious. We are facing the greatest opportunity in healthcare and medicine—perhaps not just for this generation but for many—and that is the gift of time, because we can rely much more on machines to process data, liberate us from keyboards with natural language processing, offer machine learning, and get patients to take more charge. This could give this gift of time, but it will just be used by the administrators as a way to increase productivity and make even worse the squeeze that is occurring right now. If we don't have an organization and stand up to take this on, we will miss this opportunity.

That organization is going to take hold in the weeks ahead, and it's called Osler's Alliance. I didn't know in writing the essay that there would be a strong support and robust enthusiasm to build this. Hopefully, we will get as many of the million physicians involved as possible and, as a by-product, that gravitas of having the voice of physicians to stand up for such things as the patient-doctor relationship.

Importance of Meaningful Patient Time

Verghese: The most important discovery we've made on this podcast is how much that time with the patient translates into meaningful results and efficacy. We had the pleasure of talking with Alia Crum, who does a lot of work on placebo and nocebo; it was a very eye opening session. She has a lab where she does nutritional surveys and research—at least, that is what students think at Stanford when they walk into the lab. As part of that survey, she injects some histamine into the skin, and it produces a measurable circle of red erythema. After that, a nurse-type comes in and rubs inert cream on this. That nurse-type can either be someone with a very warm affect, who may ask, "How was your day? How is everything going? Oh, I love that blouse," or may be all science and business. And this [type of interaction is associated with] changes in the diameter of the skin reaction.

This is not entirely new. We've understood for some time that when you give someone a placebo, you are not tricking them. When it works, which is about 30% of the time, you are producing a discrete neurobiological change. We have not done enough to tap into that. Alia and other researchers, such as Benedetti, have shown us that you can have a placebo without a placebo—meaning that your tone of voice, your affect, the setting, the degree of your perceived interest, and the time you spend all have significant therapeutic effects. I think the time will come when we'll have to do placebo-controlled trials with high and low placebo arms, because we have not really controlled for the human element. We just give pills, and sometimes the differences are masked by the degree of the researchers' personality.

Topol: A week or two after that podcast, a remarkable study was published in Nature Human Behaviour, where they took doctor-patient dyads—they actually were students when they were playing the doctor and patient—and they showed that subtle facial gestures transmitted an amazing amount of power to placebo or lack thereof. There are so many aspects of the idea about presence.

We had a rich discussion with Danielle Ofri, who is at Bellevue in New York City. In fact, we had to split it into two episodes because there was so much there. She talked about the physical exam and how people want to have an exam and be touched, and how that is a critical part of the present.

Verghese: Yes, people want their disease localized on their body. It might be a biopsy report, it might be an image, but ultimately, there is something deeply symbolic about the exam and touching where it hurts, so to speak. From my colleagues in anthropology, I've learned that when we do the exam, it has all the trappings of ritual. One individual is coming to a stranger, and they meet in a room with furniture that does not look like the furniture in your house or mine. One is wearing a white coat with shamanistic tools in the pocket, and the other is wearing a paper gown that no one knows how to tie or untie. Then one person proceeds to tell the other things that they would not tell their rabbi or preacher, and in my specialty of infectious disease, things they would not tell their spouse. Then amazingly, at some point, one member of this dyad disrobes and allows touch. It's a profound moment. In any other part of society, that would be assault. There has been some confusion about this at the very highest levels, but it is assault. But in the context of the fiduciary relationship we have with our patients in that moment, it is a profoundly important ritual that we are allowed to do in that setting.

My anthropology colleagues have taught me that rituals are all about transformation. We engage in rituals like marriage ceremonies, baptisms, and installations to signal the crossing of a threshold. Patients from every culture know about ritual, with all its trappings. And if we have the trappings of the ritual, but just do a half-assed prod of the belly and stick the stethoscope on the paper gown, they are onto us right away. They are onto us just as you are onto the sloppy barista or the hairdresser who does not know what they are doing. We may not know how to do what they do, but we understand when it's done well. I believe there is a timeless aspect of medicine, that no amount of technology can completely take away, which is the need for this human being as the face of care to do it with some skill.

Using AI to Bring Humanity Back to Medicine

Topol: I could not agree more. AI has already been seen through deep learning to improve image interpretation at much higher accuracy at high speed, whether it's radiology, pathology slides, electrocardiograms, or soon even ultrasounds, and the list goes on and on. In fact, a study applying AI to colonoscopy video found that AI not only could pick out polyps, but it could tell whether they were [precancerous] or not in real time by machine vision.

But putting all that aside, the greatest gift that AI can give us is to go "back to the future"—to get us to the humanity in medicine, which is presence, a physical exam, listening, and building trust, communication, and deep empathy. If we don't seize this opportunity and just focus on what AI can do with respect to technical aspects, we're missing out.

Verghese: That was the epiphany for me when I read your book, Deep Medicine, a beautiful book that catalogues all these mind-blowing advances. You had a very simple graph showing the machines progressing and their capability. You said basically that machines are better than us; they can process more images than we can ever do, they don't get fatigued, they work night and day, they can assimilate more data than a human physician can hold in their head. But then you came up with this extraordinary comment at the end of the paragraph that since machines have now eclipsed us, so to speak, this is the moment where we can learn to be more human. For that aspect of medicine that involves us, this is our moment to take that further. There is a whole new realm of exploration that we're just beginning to tap into.

Topol: That says it well, because we're so interested in technology yet seem to focus much more on the short-term benefits. Whereas here we're looking at something that could be transformational to get medicine to the way it used to be.

When I finished medical school around 1980, it was very different. The time we had with patients, the fundamentals and importance of the physical exam, the bedside teaching in the inpatient setting, and the communication with patients were very different. It wasn't down to 7 minutes. The essentiality of that human connection was there as a routine, the reliance on this precious relationship. The question is, can we get back there? And are we seeing the tools emerging that will help us get there?

Verghese: Let's be clear: I don't think we want to go back to science in the 1980s, because many of the patients we cared for died for lack of what we now take routinely. But we do want to go back to some elements of caring that we have lost along the way. One of our guests was Arthur Kleinman, a very distinguished professor of medicine and anthropology at Harvard who trained Paul Farmer and many other wonderful physicians who have their hands in both worlds of anthropology and medicine. He wrote a very poignant book called The Soul of Care, about his wife getting premature Alzheimer disease. The most damning thing he said that really shook us was that 99% of what she needed was not going to be supplied by healthcare, so that word "healthcare" becomes almost ironic, because the care provided in that prolonged illness was by him day or night. It was by him, his children, and caregivers and had nothing to do with us and all the science.

We know that by 2020, we will have a big percentage of older Americans. By 2050, the proportion of people over 85 will be huge; a quarter of them will need help walking, one quarter of them will need help dressing themselves, half of them will need other sorts of minor help in activities of daily living. That is where a lot of our energy might go—in preventing that and coming up with technologies that allow us to be autonomous at that stage as best we can.

Topol: That is another important direction where this technology can help. Arthur Kleinman is the father of medical anthropology. He's written 40 books, but the one that hit me the most was the essay he had in the Lancet several weeks ago, "The Soul in Medicine," and his outpouring about how medicine has lost its soul for the reasons we've been discussing. We need to get it back, and we can. The most exciting thing about where we are right now is that we are looking at real, exciting potential.

Verghese: One of our challenges is to make this sexy. How do you make the human aspect of medicine of interest to entrepreneurs? I'm reminded of the term "evidence-based medicine," which was a science that became popular because of Alvan Feinstein, an extraordinary clinician at Yale. When he labeled evidence-based medicine, he was not talking about taking data and mining it; he was really talking about taking the evidence from the patient, using a skilled physician, and making that the basis of some of the decisions. We've gotten very far away from that, where we think the only good data are the data printed out—the other data we simply ignore.

We have this wonderful ontology for disease. You can have a disease, and I can give it a number. I can tell you whether it was the left side or the right side. Acute or not. But we really know nothing about the patient beyond that this is a 47-year-old white man who came in with chest pain. Yet there are reams of facts about their life that they possess, perhaps on social media, that we can do a much better job of bringing to the people caring for them. Great avenues for research, in my mind,

Topol: No question. To sum up, a lot of the progress that has been made in AI in the medical sphere has been reported as the algorithm, the neural network, outperforming physicians on tasks, whether it's in pathology or radiology or whatever the field. What happens when you put these two together? There has not been anything yet studying what we're interested in, which is how to make the communication or the bond between patients and doctors better. That is something that we hope to see in the research going forward.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: