Emperor's Mukherjee on the Three Laws of Medicine


October 12, 2015

This feature requires the newest version of Flash. You can download it here.

Editor's Note: Last month, Medscape filmed a special One-on-One interview with Siddhartha Mukherjee, MD, in front of a live audience at the Rainbow Room in New York City. During the interview, Dr Mukherjee discussed his new book, The Laws of Medicine: Field Notes From an Uncertain Science (Simon & Schuster/TED Books, 2015) with Medscape Editor-in-Chief Eric J. Topol, MD.

Dr Mukherjee offers an insightful glimpse into the uncertainty, imperfections, priors, outliers, and biases that form his new book, The Laws of Medicine, and how these laws affect the practice of medicine. Below is a transcript of this discussion, which has been edited for clarity.

If Medicine Is a Science, It Must Have Laws

Dr Topol: I have been fortunate to interview some of the most interesting people in the world of medicine, but only one person has been interviewed by me twice because he is the most interesting and the most talented writer in medicine of our era. Tonight, we are going to talk about his second book, The Laws of Medicine, and we will also touch on his third book coming out next spring. Let's get into this book. It's a very interesting, short read, not like the opus, The Emperor of All Maladies. What was the stimulus for you to do The Laws of Medicine?

Dr Mukherjee: Thank you for having me. I enjoy these conversations. They are just like fireside chats. The Laws of Medicine began as a very simple project. The best books don't begin as books. When I was a young resident, I read a book that was simultaneously very funny and very acerbic—the famous The House of God. How many people have read The House of God? It's a strikingly important book. The author, Samuel Shem, wrote these very acerbic laws in The House of God. Law number 10 is: If the radiology resident and the intern both see a lesion on the x-ray, that means that the lesion doesn't exist.

Dr Topol: Another was GOMER (Get Out of My Emergency Room).

Dr Mukherjee: Another important law, of which I remind myself often, is law number 3 or 4, which says (and I might be misquoting here), "The patient is the one with the problem." That is a very important thing to remember in medicine because we often forget. We think that it's our problem, when, in fact, it's the patient who has the problem. I don't intend these remarks to be humorous. We take them very seriously.

The laws were very acerbic, but they were also very important because they framed a way of thinking around medicine. I wanted to find a frame. I didn't want to make capital "L" laws or commandments. I wanted some direction for myself as I navigated a field that was fundamentally uncertain and full of things that I didn't understand. And I wanted to ask the question, "How do we imagine what principles hold for the future?" The framework that I used (and that is important here) was that colloquially, we hear about the "science of medicine." If there is a science of medicine, then science has laws. Physics has laws. Chemistry has laws. Biology has laws.

The simple question was: If that's the case, then what are the laws of medicine? These were not meant to be universal commandments. These were meant to be explorations about principles that might hold true about medicine today and about medicine in the future. That was the framework for this book.

The "Inspiration" of Lewis Thomas

Dr Topol: We're going to talk about three laws. But first, it seems that the book was inspired by Lewis Thomas. Can you tell us more about Lewis Thomas? Had you met him?

Dr Mukherjee: I never met Lewis Thomas, but I was inspired by his book.

Dr Topol: He wrote a book called, The Youngest Science: Notes of a Medicine-Watcher. That seemed to set the tone.

Dr Mukherjee: Even the title is echoed in this new book. Lewis Thomas worked at New York University and subsequently at Memorial Sloan Kettering, where he was head of Memorial Sloan Kettering. Lewis Thomas was very influential for me and for many young physicians. He wrote some incredibly important books, among them, The Lives of a Cell. But the book that I thought was very interesting, and which inspired me, was The Youngest Science, in which Thomas made the claim that medicine is the youngest science.

Lewis Thomas trained in the 1930s and 1940s and began to practice medicine in the 1950s. He experienced an absolute transformation in medicine. He came from a time when medicine was observational. If you had a heart attack in the 1930s and you had congestive heart failure following that, doctors would admit you to the hospital. They had four or five interventions. They could put you in an oxygen tent. If you had severe volume overload, they could insert a needle and remove several mLs of blood and throw it away. That was the way they would cure volume overload. In the worst-case scenario, they used all sorts of things to try to diurese you. There were rumors that they would make patients drink coffee to try to diurese them.

Dr Topol: You allude to the three Ps: placebo, palliative, and plumbing. And there was a therapeutic nihilism.

Dr Mukherjee: That's right. Thomas described this as therapeutic nihilism because he said that doctors had nothing to do. But as I point out in my book, therapeutic nihilism in medicine was extremely important because it was like taking an eraser and saying, "The past is the past. It didn't work. All that stuff was nonsense. Let's just observe our patient and do nothing." So, I say that the dictum (the Hippocratic oath) was transforming from "First, do no harm," to "First, do nothing" and just look. In Thomas's time in the 1930s, for a time, people like William Osler at Johns Hopkins decided that they would erase the past and do nothing. They would just observe.

What happens? What is the natural history of disease? That idea was contaminated with nonsensical interventions, half of which didn't work. There were no trials. Therefore, doing nothing was like a cleansing. They did nothing, and they observed. What happens when the heart fails? What are the sequelae? What are the sequelae of a heart attack? What happens when someone gets bacterial pneumonia? What is the natural history of bacterial pneumonia?

Thomas watched therapeutic nihilism first cleanse the discipline of medicine, and then he watched the first great reasoned scientific medical interventions enter the world. He gives a beautiful example. In his time, if a patient had a streptococcal infection, the first thing that would happen is that the patient would be admitted to the hospital. The intern's job was to type the streptococcal infection based on the agglutinin reaction. Blood would be drawn and sent to the pathology lab. Your job that night was to identify the antigenic subtype of the Streptococcus. If you didn't do it correctly, you would have a dead human in the morning who might be a 30-year-old man who had developed an incidental streptococcal infection. You would spend the entire night figuring out the agglutinin reaction and which antibody it would cross-react with, so that by the next morning, you could transfuse that patient with the convalescent serum (antibodies) from another patient who had the same subtype of streptococcal infection.

Imagine this: Your job as the intern was to save the patient, to find that reaction so that the next morning, you could inject the patient. If you got the answer right, the patient would be cured and rise up the next morning from a literal deathbed. Streptococcal pneumonia with empyema was a lethal infection. After being injected, the patient would have a brisk immunologic reaction, clear the bacteria, and rise from the dead. Thomas watched this happened, first with infectious diseases and then with cardiac diseases. All of a sudden, you weren't just putting the patient in an oxygen tent. You were starting to give real diuretics and antiarrhythmics—all of these new medicines that were being developed.

Finally, he started watching that happen in cancer. Patients started getting chemotherapy. They would go into remission. He said, "There is the beginning of a science here because what started from reason (the first principles), then identified pathophysiology, and finally went from pathophysiology to therapeutics."

The Power of Prior Probability

Dr Topol: This is what I love about your writing; you take lessons from history. They are rich lessons that are often either neglected or were just not known. Then you integrate that with your own clinical experience with patients, which you do in this book quite a bit.

Let's talk about the laws. The first one is about priors. "A strong intuition is much more powerful than a weak test." Tell us about that.

Dr Mukherjee: The book is actually dedicated to Thomas Bayes. I was mesmerized by the idea that an 18th century cleric working by himself in a church could figure out the idea that prior probability dictates posterior probability. How applicable has that become to medicine today, and to horse racing, and banking, and everything that we do? I'm fascinated by the idea. Lewis Thomas had reached a point at which he said, "We have all these tests. The tests will be matched with medicines, and then we are going to solve the medical problem. This is going to be the science of medicine." We talk about precision medicine all the time. Say you come in, and you are going to have a test. The test will match, and that will help us understand the disease. We will understand the pathophysiology that will be matched to a medicine. You will get the medicine. You will go home happy. Everyone is happy; that's the end of the story.

But in reality, medicine has not worked out so simply. One thing that has not worked out simply is that no test, as we found out, is a pure test. Tests have false negatives. They have false positives. That is an intrinsic feature of testing; whether you measure a prostate-specific antigen (PSA) level or a BRCA1 mutation, we are not able to predict with 100% certainty whether a woman with BRCA1 mutation will develop breast cancer. We can't predict, obviously, if you have a high PSA level, whether you have prostate cancer or whether your prostate cancer is the aggressive or the nonaggressive indolent type. The first idea that I wanted to convey in the book is that the only way that we learn how to use tests is by using prior knowledge. The example I use is very simple and one that even Bayes understood.

You go to a street fair, and someone is tossing coins. The coin lands heads the first time, heads again the second time, third time, fourth time, and fifth time. If you ask a pure statistician, "What is the chance that the coin will land heads the next time?" the pure statistician says "50/50." But if you ask a child, the child will say "Stupid, the coin is rigged." Right? That's why it's landing heads all the time. The child knows more than the pure statistician because the child understands that prior probability dictates posterior probabilities.

That idea has been difficult for even doctors to understand in medicine, that with the things we do that seem so complicated—genomic sequencing, epigenomic sequencing, complex family history mapping, etc—all we are really trying to do is take tests and switch their prior probability. That's the real message that we are trying to understand.

We cannot interpret the new data without the context. That's why medicine exists. One thing I conclude is that medicine exists because you need that prior context. When you come as a patient to a physician's office, one of the first things they do is get a history. That's why the medical data sheet begins with history, not tests. It begins with history and physical, not test and physical or conclusion and then physical.

That's the first law. It's really a Bayesian idea, but Bayesian analysis toward, in the future, how we integrate genomic information, family information, information about a person's particular risk factors, and rich data about their exposures, epidemiologic history, genetic history, and racial history, and use that in a sensible way to ask the right questions about the future.

Inliers vs Outliers

Dr Topol: That's great. Those who are not into Bayesian principles will get this very well from The Laws of Medicine. The second law is about outliers, but I also want you to tell us about inliers because I like that term. "Normals" teach us rules. Outliers teach us laws.

Dr Mukherjee: This second idea grew out of my own experience. I'll begin with a simple analogy. A lot of what is happening in medicine today is that we have solved quite effectively what I call the "inlier problem." Inliers is a word I made up. The inlier problem is that we have a relatively well-demarcated idea of the normal range of physiology. There is a normal range of blood pressure, height, weight, etc, based on large population studies.

Once in a while, however, we have people who lie outside of that. There are people, for instance, who have very high blood pressure but who have never had a stroke or heart attack. In the past, that was dismissed as coincidence, and surely, some of that is coincidence. Some of that is because of random effects. In time, this is like the person who comes and says, "I smoked all my life and didn't get lung cancer. Therefore, cigarettes cannot possibly cause lung cancer." And you can say, "Well, that is obviously not true because we know from large amounts of data [that they do]." But the converse is very interesting, which is to ask the question, "Why is it that some people who have had long histories of smoking don't have lung cancer?"

Some fraction of that is clearly because of chance, but some fraction is not. And when you identify those people, you are going to get new insights into the pathophysiology of cancer. You could say the same thing about heart disease or strokes. The point about the second law is that we have spent a lot of time creating this understanding of the inlier problem. But what is really interesting is to find the outliers and figure out what they tell us about the deeper structure of the pathophysiology of a disease.

Dr Topol: Like the exceptional responder in cancer.

Dr Mukherjee: I use the exceptional responder as an example. Cancer is a highly genetic disease. Obviously, environmental carcinogens trigger genetic events. For the longest time, we spent a lot of time in cancer biology and clinical trials, saying, "Who's the typical responder?" That was very helpful because it tells you who responds—it's precision medicine. Who responds to tamoxifen? Who responds to some combination of drugs?

But now, with all of these new tools that we have, we can begin to answer the questions, "Who does not respond to the standard therapies, and who responds to some exceptional therapies?" Those are very interesting because they point to deeper principles of pathophysiology and deeper principles of medicine, which are brought out by these exceptional responders.

The Bias Hunters

Dr Topol: The third and final law is about bias. For every perfect medical experiment, there is a perfect human bias. What do you mean by that?

Dr Mukherjee: The third law was inspired by an idea that I had while I was practicing medicine—really while I was learning medicine as a resident. I was inspired by a very important book called Microbe Hunters (Harcourt, Inc., 1926) by Paul de Kruif. If you ask the major scientific and medical minds of today which books inspired them, they will probably answer Lewis Thomas's The Youngest Science or Lives of a Cell or Paul de Kruif's book, Microbe Hunters, in which he describes the birth of microbiology and infectious diseases.

I thought to myself, if the doctors of the 1930s or 1950s were microbe hunters or cause hunters, what are we doing today? What are we doing today when we look at all of these clinical trials? I realized that one thing we are doing (not the only thing) is hunting bias. So much information comes to us from clinical trials, and the popular press is full of information.

As physicians, we have to figure out how to think about these trials in a critical way and be skeptical about them and say, "Look, that's an important piece of information, but let me tell you what the bias in that information is. I can now take the larger body of information and apply it to a single human being, to a single patient." I will give you one example. A large randomized trial is run on patients and clearly shows that tamoxifen works extremely well for preventing breast cancer in high-risk populations.

Ten days later, an African-American woman comes to your clinic, meets the profile criteria, and says, "Doc, should I take tamoxifen? I took it for 2 months and had a terrible reaction to it. I had terrible side effects. I really don't want to do it. It made my life miserable. But if you think it's useful, I'll do it." A stupid thing to say is, "Here's the trial. It says that women who take this drug have a benefit. You're a woman. Let me give you the drug."

The much more interesting and much more important way is to integrate all of this information and say, "Well, wait a second. That trial was run on white women in Kansas. What are the chances that the findings of this randomized trial—although it's a powerful, focused trial—are going to apply to you, a 36-year-old who was not in the initial group and who has a very different racial and genetic background?"

Our job right now is to interpret very complex pieces of data, some randomized, some nonrandomized, and other pieces of information and incorporate all of them into the treatment of an individual who is sitting in your office and needs help. That's basically the third law.

Teaching an Imperfect Science

Dr Topol: You have put these together—priors, outliers, biases, and other laws of imperfection in medicine. There are many implications of your thoughts here. One that I want to get into is related to medical education and selecting doctors because, as you point out, the profusion of data that we have is knowledge but not necessarily wisdom. Should we be selecting different types of people to be the doctors of the future based on some of these points?

Dr. Mukherjee: I'm not sure that who we select or don't select is interesting. What is interesting to me is how we train the people that we do select. Once you have selected the people, how do you allow individuals—physicians in training—to be able to deal with fundamentally imperfect information? That is the science of medicine. The science of medicine tells us that uncertainty, imperfections, priors, outliers, and biases are endemic. They are not going to go away today, tomorrow, or 50 or 100 years from now. The sooner we instill these ideas in ourselves, our students, and our patients, the more likely it is that we will have a much more collaborative effort to solve the major problems of human disease. We need to do that.

We have to approach this with the kind of humility with which Thomas Bayes approached his central problem. Thomas Bayes was a profoundly religious person, but he didn't see a conflict between religion and science. He was a scientist, a mathematician, and a religious person as well. He said to himself (and I find this very moving), "There is no way that I can a priori know the mind of God. But I can use the current knowledge of the universe and the world around me to guess. I can use prior knowledge to guess at absolute knowledge. And it will always be the case that I will never find the absolute knowledge. This is a given. But it will also always be the case that I will have enough prior knowledge with me that I can try to abstract and guess what the real knowledge is going to be." What a beautiful idea.

It's amazing that even in 2015, we are struggling with that idea. How do you use real information to get at the most important and beautiful information that exists—truths?

Asking the Right Questions

Dr Topol: We are facing some very important issues today that suggest how bad things are still in 2015. One is that we make 12 million serious diagnostic medical errors a year, and that is unchanged, as best we know, since To Err is Human was published in 1999. It seems that without any changes, this will continue, and it has become inhumane to have all of these errors. To the top 20 drugs that are prescribed, by sales, at least, 80% of people are nonresponders. We give these drugs. We have hope, but the fact is that the plurality of patients don't respond.

Then we have the issues of false-positive results in screening (mammography, PSA levels) at rates that are greater than 60%. Yet these tests are done widely in millions of people every year. The sense I got from your book was that you don't think we can change this too much. We have to live with this.

Dr Mukherjee: The sense that I have, from writing the book, is not that we are going to have to live with it but that we have to refine it. If a person has a PSA test, and it's positive, to ask whether that person has prostate cancer is a stupid question. The question to ask at that stage is: What are the risk factors? What do we know? Who are the outliers? In other words, how do we modify knowledge so that, going back to the Bayesian analysis example, you go from concrete test knowledge to truth?

So the book is a plea for all of these changes in context, which is to say that we have a rich opportunity today to integrate extraordinarily deep medical context into the information system of a human being. I still think that we will have perceptual biases. That's not going to change.

Dr Topol: It's hard to get over that.

Dr Mukherjee: There will still be outlier and inlier problems. Tests will still be imperfect. But we have the opportunity to integrate all of this unbelievable information and make a real difference in medical care.

Not Big on Big Data

Dr Topol: I don't know whether I have trapped you, but we are getting to the point that you are not big on big data. In the final two sentence of the book, you quote Malcolm Gladwell saying that the political revolution will not be tweeted. Your final sentence is, "Well, the medical revolution will not be algorithmized." The problem that I see is that we have an enormous amount of data about each person, allowing us to have this context that we never had before. (That's "precision medicine," if you will, or individualized medicine.) No human being can process that much data, right? You would agree?

Dr Mukherjee: Yes, that's correct.

Dr Topol: So, don't you think that algorithms and computing power—machines, machine learning, and artificial intelligence—will have a place in getting us over some of this imperfection?

Dr Mukherjee: I agree with that. I have stated this before. I don't have a problem with big data. I don't believe or disbelieve in big data. Big data exists. Whether I believe or disbelieve in big data is not relevant. What is relevant is: Are we using intelligence to interpret big data? There is a seduction to big data where a lack of intelligence can show. I'll give you the classic, historical example—the great argument between Galton and Mendel. This brings me to the new book.

Galton was a classic big data man. He measured things. He went around the world and said, "What do the shapes or sizes of noses look like?" And he plotted a smooth curve for the size of noses, and it looked like a big bell-shaped curve. He said, "What does intelligence look like? What does IQ look like?" And sure enough, it's a bell-shaped curve. What does the structure or color of skin look like? Bell-shaped curve. If you take all this information, you would think that human heredity is moved from human beings to human beings through bell-shaped curves.

It took a monk, who had no experience in statistics, for the most part, who crossed two peas, the smallest of small data—and found that, in fact, things don't behave that way. You cross a short plant and a tall plant; you don't get a bell-shaped curve. You get a short plant. It's very discrete. You cross a yellow pea and a green pea, and you don't get an intermediate color pea. You get only one pea. The point is that there is a seduction to big data without smart interpretations of data.

I have no problem with big data. My problem is that if you don't use intelligence to interpret big data, then that big data produces nonsense. It produces more big data. And the job of big data is not to produce more big data. It is to produce wisdom, knowledge, and truths.

Discussion of Another New Book

Dr Topol: Your next book is going to be out next May or June? And it is called The Gene? Can you tell us a little about it?

Dr Mukherjee: I'm happy now that the book is complete. It's finished. It's a much deeper, bigger book. You could throw it at someone and kill them. I hope that it reads with the same kind of intensity and people fall into it. It opens a universe, and the book is about what happens once we learn to read and write the human genome. In your laboratory, as in our own practice and our laboratory, we are now reading genomes for cancer and other diseases in a way that we couldn't do before. And we are writing genomes. We are editing and changing genomes with a facility that we didn't have before.

The question is: Now that those tools have become available and we know how deeply the genome is linked to human identity, what happens once we begin to change it? That is the question of the book. It starts from Mendel and goes into the future.


Dr Topol: Let me just say that this has been really stimulating, and the book, The Laws of Medicine, is great. It's a short read, but it's an impactful read. And also we are looking forward to the next biggie, The Gene, because that's also going to be a highly impactful book. The way you're able to juggle being a researcher, still be a medical oncologist, and an author—I don't know how you do it.

Thanks so much for joining us.

Dr Mukherjee: Thank you.


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.