Joe: Most people think the rise of AI in medicine means being treated by a cold, unfeeling robot. But what if the opposite is true? What if AI is the only thing that can save the human soul of healthcare? Lewis: That’s a bold claim. It sounds like a paradox. You’re saying more machines will lead to more humanity? That feels like it goes against everything we assume about technology. Joe: It’s the provocative argument at the heart of Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again by Eric Topol. Lewis: And Topol isn't just some tech guru. He's a world-class cardiologist and researcher, but what makes this book so powerful is that he also writes from his own harrowing experiences as a patient. It's that blend of expertise and vulnerability that makes this so compelling. Joe: Exactly. And before we get to the AI solution, Topol argues we first have to understand the problem he calls 'Shallow Medicine.' It's his diagnosis of our current system, and it's pretty grim.

Lewis: ‘Shallow Medicine.’ That’s a harsh term. What does he mean by that? Is it just about doctors being rushed? Joe: It’s deeper than that. It’s a crisis of insufficient data, insufficient time, and most importantly, insufficient presence. It leads to mistakes. He tells this chilling story about a patient he calls Robert, a 56-year-old store manager. Lewis: Okay, let's hear it. Joe: Robert has a ministroke. One moment he's fine, the next his vision is blurring and his face goes numb. It’s terrifying. He’s rushed to the emergency room, put through a whole battery of tests—head CT, blood work, EKGs, the works. The symptoms eventually fade on their own, but nobody knows why it happened. Lewis: A classic medical mystery. So what happens next? Joe: He sees a neurologist, then a cardiologist. The cardiologist does an echocardiogram—an ultrasound of the heart—and finds something called a PFO, a patent foramen ovale. It’s basically a tiny hole between the heart's chambers. It’s actually pretty common, about one in four people have one. Lewis: One in four? So it’s probably not the culprit, right? Joe: Well, the cardiologist confidently declares, "That's it! That's the cause of your stroke. We need to do a procedure to plug the hole." Case closed. They schedule him for this invasive heart procedure. Lewis: Wait, just like that? Based on a common anatomical feature? That's terrifying. It feels like they're just jumping to the first plausible answer, treating the most obvious clue instead of the whole person. Joe: That's the essence of shallow medicine. It’s reflexive. It’s pattern-matching without context. But thankfully, Robert gets a second opinion. This new doctor, representing the kind of medicine Topol advocates for, decides to dig a little deeper. Instead of another expensive scan or procedure, he prescribes a simple, non-invasive device called a Zio patch. Lewis: A Zio patch? What’s that? Joe: It’s a small adhesive monitor you wear on your chest for a couple of weeks. It just continuously records your heart rhythm. It’s simple, cheap, and gathers a ton of data over time. Lewis: And what did it find? Joe: It found the real culprit. Robert was having multiple, asymptomatic bouts of atrial fibrillation—an irregular heartbeat that is a major cause of strokes. The PFO was a red herring. The Zio patch proved it. Lewis: Wow. So the big, scary heart procedure was completely unnecessary. Joe: Completely. It was cancelled. Robert was put on a simple blood thinner, the right treatment for atrial fibrillation, and he was fine. The right diagnosis was found not with a million-dollar machine or an invasive surgery, but with a simple patch and a doctor who took the time to think deeper. Lewis: That story is infuriating and hopeful at the same time. It perfectly illustrates this idea of 'shallow medicine.' The system is so geared towards speed and intervention that it misses the bigger picture. Doctors are so burned out and buried in data entry that they don't have the mental space to just... think. Joe: That's exactly Topol's point. He argues that doctors have become data entry technicians, spending more time with keyboards than with patients. The system is broken, and it's leading to an estimated 12 million significant misdiagnoses in the U.S. every year. Lewis: Twelve million. That’s a staggering number. It’s a public health crisis hiding in plain sight. Okay, so that's 'Shallow Medicine'—rushed, incomplete, and potentially dangerous. It's easy to see the problem. But how on earth does AI, a computer program, fix this very human problem? It seems like it would just make it worse.

Joe: That’s the leap of faith the book asks us to take. Topol argues that AI’s greatest strength is its ability to see things we can't. It's a master of pattern recognition. Think of it this way, Lewis. The human genome has millions of genetic variants. For a doctor, finding the one causing a rare disease is like finding a single typo in a library of thousands of books. Lewis: An impossible task. Joe: For a human, yes. But for an AI, it can read that entire library in seconds. And this isn't theoretical. Topol shares a story from Rady Children's Hospital that is just breathtaking. Lewis: Okay, give me the mind-blowing example. Joe: A newborn baby boy is sent home healthy. A few days later, he's back in the ER with constant, unrelenting seizures. Doctors are throwing everything at him, but nothing works. They have no idea what's wrong. His prognosis is bleak. Lewis: Oh, that’s every parent’s worst nightmare. Joe: Absolutely. In a last-ditch effort, they send a blood sample for rapid whole-genome sequencing. But instead of having a team of geneticists spend weeks or months analyzing it, they feed the data into an AI. The AI also ingests the baby’s entire electronic medical record, identifying 88 distinct features of his condition. Lewis: So it’s cross-referencing the genetic code with the actual symptoms. Joe: Instantly. The machine-learning algorithms sift through millions of genetic variants and in just a few hours, it flags one. A single, incredibly rare variant in a gene called ALDH7A1. It causes a metabolic defect that leads to seizures. Lewis: And the cure? Is it some complex gene therapy? Joe: This is the most beautiful part. The cure was simple dietary supplements. Vitamin B6 and arginine. The AI identified the problem, and the solution was readily available and simple. They gave the baby the supplements. Lewis: And what happened? Joe: The seizures stopped. Abruptly. Thirty-six hours later, he went home. On follow-up, he is perfectly healthy, with no brain damage, no developmental delay. That baby is alive and thriving today because an AI saw a pattern that no human doctor could have found in time. Lewis: That's unbelievable. That’s not just an efficiency tool; that’s a life-saving miracle. So the AI isn't just a calculator; it's a detective with superhuman vision. It's seeing clues that are invisible to the human eye. Joe: It's a superhuman detective. And that brings us to the book's biggest, most beautiful paradox. Topol argues that the greatest gift of this technology isn't its superhuman intelligence. It's the gift of time.

Lewis: The gift of time. What do you mean? Joe: Think about all the 'shallow' work that crushes doctors. Reading scans, interpreting lab results, writing notes, navigating electronic health records, handling billing codes. Topol says this is what AI is built for. Let the machine do the superhuman pattern recognition. Let it analyze the thousands of mammograms, pathology slides, and retinal scans. Lewis: So it automates the grunt work. Joe: Exactly. It automates the drudgery that leads to burnout and errors. And what happens when you free the doctor from the keyboard? What happens when they don't have to spend hours deciphering data because an AI has already synthesized it for them? Lewis: They have time. Time to actually talk to the patient. Joe: Time to be present. Time to listen. Time to build trust. Time to show empathy. This is the core of 'Deep Medicine.' The AI handles the science so the doctor can practice the art of medicine. Lewis: That’s a powerful idea. But is there any evidence that technology can actually foster that kind of connection? It still feels like a machine is getting in the middle of a human relationship. Joe: There is, and it's fascinating. Topol highlights a study from the Institute of Creative Technologies. Researchers created a virtual avatar on a screen, named Ellie, to interview people. They had one group believe Ellie was just a sophisticated computer program, and another group believe a human was controlling her remotely. Lewis: A kind of digital Turing test. Joe: Right. And Ellie would ask progressively more personal questions. What they found was stunning. People, especially military veterans dealing with PTSD, were far more willing to open up and share their deepest, most painful secrets with the avatar they believed was a machine. Lewis: Really? Why? Joe: Because they knew the machine wouldn't judge them. There was no fear of social consequence, no worry about what this "person" thought of them. The technology, in its non-judgmental nature, created a space of profound psychological safety. Lewis: Whoa. So the AI isn't meant to be the empathetic one. It creates a safe space and handles the grunt work, which allows the human doctor to be more present and empathetic. The machine makes the human more human. That's the 'Deep Medicine' idea. Joe: That's it exactly. The goal is to combine deep learning with deep empathy. The AI provides the data, the insights, the patterns. But the human clinician provides the context, the wisdom, and the care. It's a partnership.

Lewis: So when we strip it all down, this isn't really a book about technology. It's a book about attention. Shallow medicine is a crisis of attention. Doctors are forced to give their attention to keyboards and billing codes instead of patients. Joe: Precisely. And Topol's vision, which is both optimistic and, as some critics point out, faces huge hurdles in our current profit-driven healthcare system, is that AI can redirect that attention back to where it belongs. The ultimate goal isn't a smarter diagnosis; it's a more caring conversation. Lewis: It’s about using the machine to free the human. To let doctors be doctors again. Joe: Yes. There's a famous quote from a physician from a century ago, Francis Peabody, that Topol brings back. He said, "The secret of the care of the patient is in caring for the patient." It’s so simple, yet we’ve lost it. Topol believes AI, paradoxically, is our best hope of finding it again. Lewis: It makes you wonder, the next time you're at the doctor's, what would that visit feel like if the doctor never once had to look at a screen? If their full attention was on you? Joe: A powerful thought. We'd love to hear your own stories about this. Have you ever felt the effects of 'shallow medicine,' or maybe even seen a glimpse of this deeper, more human-centered care? Share your thoughts with the Aibrary community. Lewis: This is Aibrary, signing off.