by Steven Hatch
FIGURE 1.1. Spectrum of certainty.
At the left end of the spectrum, we encounter the idealized form of medical knowledge, where we have a high level of confidence that we really do know something, that this something indicates clear-cut benefits, and that our knowledge will not be subject to massive revision.* Most people, and many doctors, believe this is the state of much current medical knowledge, although I am a bit less sanguine that this is so. To be clear, I harbor no doubts whatsoever that red blood cells transport oxygen, for instance, or that HIV causes AIDS or that antibiotics improve a patient’s chances of surviving bacterial pneumonia. But there are a good many other aspects of medicine that remain in murkier territory.
Readers shouldn’t infer any political implications from this left-right-center scheme; it’s totally arbitrary, whatever my political views.
The center-left side of the spectrum is what most would consider reasonable but not absolute confidence. Do drugs for diabetes save lives? Depending on the drug, the answer is yes—but several diabetes medications come with some pretty serious side effects such that we can’t assure every patient that taking them will be beneficial. Many diagnostic technologies occupy this part of the spectrum, as we’ll discuss in the first few chapters. This part of the spectrum is still a pretty good place to find oneself, but there is room for improvement.
As we approach the middle of the spectrum, we enter the realm of pure speculation, where evidence is either completely contradictory or lacking altogether. For instance, at present there is much research devoted to the impact of the gut microbiome—that is, the many billions of bacteria that live inside our intestines and the DNA that they possess—on human behavior and mental state. Researchers have a sense that something is going on, though exactly what it is and how this may translate into drugs that might alter our perception of the world and how we interact with it is anyone’s guess. (That hasn’t stopped rampant speculation on the Internet about “mood altering” food regimens, however, an example of the profit to be made in creating the illusion of certainty. This is a problem that extends beyond the hawkers of fad diets: in Chapters 6 and 7, we’ll look at what happens when multinational conglomerates do essentially the same thing.)
As we start to move toward the right side of the spectrum, we begin to have greater confidence in our knowledge, but this time our increasing certainty is of the harms of some drug or innovation or diagnostic approach. Perhaps the most provocative argument I will make in this book is that the practice of using mammograms to screen otherwise healthy women under the age of fifty is on the center-right spectrum of certainty and that there is a minimal to moderate amount of evidence that, as currently performed, mammography in this population carries overall net harm.
Finally, the right side of the spectrum is where we’re quite confident that some practice is harmful. For example, avoiding antibiotics is a bad idea when one has clear signs of a probable bacterial infection. But it’s also a bad idea to take antibiotics, especially for prolonged periods, when there is no evidence of bacterial infection. Occasionally, some groups have a vested interest in sowing seeds of confusion and having people believe that some medical knowledge is in the middle of the spectrum when in fact it is out toward the right end, as it is in this case. Chapter 5, which looks at the treatment of Lyme disease, will explore one such group in depth.
The spectrum of certainty is a crucial tool to help make sense out of the sometimes overwhelming information with which a patient or family member can be bombarded when trying to understand a health issue. My argument is that uncertainty is the great unspoken secret of medicine and that by ignoring this fundamental uncertainty we are doing real harm to ourselves. However, I don’t make this argument in a linear fashion. False certainty can lead us as doctors and patients to misinterpret data and thus make bad choices; each chapter, in some way, adds evidence to this argument. But I cover a lot of ground and investigate many different disciplines. That’s by design so that you can see just how pervasive uncertainty really can be. By utilizing the idea of the spectrum of certainty, readers can envision the broader claims of the book without having to “reinvent the wheel,” as it were, as they read each new chapter, struggling to connect each divergent story, hearing only the static and not the signal.
A working notion of the spectrum of certainty also allows one to move past the binary construction of doctors either knowing everything or doctors knowing nothing at all (more on this latter view anon). It also provides readers with some perspective on how best one can probe a health-care provider, allowing one to ask the deceivingly simple question about what is or isn’t known about a subject. Finally, it provides a framework by which we can apply some mathematical precision to a topic. For example, in Chapter 1 we’ll look at the prostate specific antigen (PSA) test, a screening tool for prostate cancer; when one attempts to quantify the exact benefits of the PSA test, the often fierce debates over the past ten to twenty years about its value seem fairly ridiculous.
Consider the Donald Rumsfeld quote that began the introduction. “There are known knowns, there are known unknowns, and there are unknown unknowns”—his quip came as part of a tart reply to a reporter who had the temerity to question whether the Bush administration should have anticipated the chaos that engulfed Iraq after the US armed forces deposed Saddam Hussein in 2002. His point, a rather dressed-up version of the observation “shit happens,” was meant to convey the impossibility of knowing with certainty how a post-Hussein world would work. Whatever else one may think of Rumsfeld, the administration he served, or the planning and prosecution of the Iraq War, he left us with one of the more crisp and useful observations of the nature of epistemology: sometimes one is certain of the state of the world, sometimes one can have a clue about it, and sometimes one is utterly flummoxed by what’s really out there.
This book is primarily concerned about the middle region of the spectrum of certainty: the known unknowns in medicine. Moreover, this book asserts something that may be a surprise to both patients and doctors alike: most of medicine functions in the world of known unknowns—as well as the unknown unknowns! That is, doctors often may know the general outlines of a problem but may not know, or even be able to know, with total certainty the specific problem in a given patient at a given time. This book is an attempt to describe that aspect of medicine where the light of knowledge is dim and the mind can play tricks on itself, diagnosing things that turn out not to be there or creating mass hysteria over relatively trivial and distant threats such as exotic, tropical viruses while simultaneously ignoring the public menace of the double cheeseburger, a significantly more lethal object in the Western world and especially so in the United States.
The writer Michael Pollan wryly informed the readers of his book In Defense of Food that his opening seven words—“eat food, not too much, mostly plants”—were the boiled-down advice of his entire tome, and the chapters themselves were merely clarification and elaboration of that advice. In medicine, too, we have an answer to the seemingly ferociously daunting question, how do I stay healthy? We’ve known the answer with increasing scientific certainty for several decades now, and it mirrors the straightforwardness of Pollan’s advice in the same number of words:
Exercise more,
Eat less, and
Do not smoke.
This is the far left of the spectrum of certainty, and not much else can be found out there except one or two items that we’ll take up during the course of this book. Unless you have some relatively unusual disease such as lupus or primary biliary cirrhosis that is often genetically determined or requires special medications, this is really all you need to know about maintaining your health. That is because most Americans die from cardiovascular disease and diabetes (which is why you should eat modestly and exercise) or emphysema and lung cancer (which is why you shouldn’t smoke). Everything else is, largely, commentary. Of course, there are important health stories that deserve coverage, but one could make a strong case
that these seven words should begin and end every news item that deals with medicine and health. So, for those reading books on the economy class model, I just gave away the secrets to healthy living at the outset, and you can feel free to read no further.
The Road Map
Broadly speaking, there are three major portions of this book. The early chapters deal with problems relating to uncertainty in diagnosis—that is, when do we know that someone has a disease? Chapter 1, “Primum Non Nocere” (an ancient Latin dictum meaning “first, do no harm” that continues to be regarded as a bedrock value of medicine to this day) looks at the conundrum of overdiagnosis. With increasingly sensitive ways of detecting disease by means of technological advances in radiology and biochemistry, we are able to find diseases earlier in their course and thus have a greater impact on mortality. But it’s become clear over the past generation that there is a price to be paid for this, and it has come in our finding “disease” that turns out not to be disease in the conventional sense of the term—namely, some biologic process that would lead to illness or death if left unattended. For instance, we’ll see how doctors have found more and more cases of cancer, even though finding these cancers earlier hasn’t ended up saving any lives, which must mean that the cancers they’ve found aren’t the kind of cancers that actually kill people.
The problem is that we can only know such nondisease diseases exist at a population level; when confronted with an individual patient, it is impossible to know with much certainty that some treatment will carry the expected benefit. Since all treatments carry risks, this means that we are quite probably harming some patients as a consequence of overdiagnosis.
The second chapter, “The Perils of Predictive Value,” briefly recounts the tale of the physicist and author Leonard Mlodinow when he received a shocking diagnosis of a terminal illness as part of a standard insurance exam. Only the diagnosis was wrong, and we will investigate just how wrong it was. Mlodinow’s story is a cautionary tale in the perils of what are known as “false-positive” tests, which is exactly what it sounds like—tests that appear to indicate disease but do so wrongly because no test is 100 percent perfect. The consequences of false-positive tests—and what treatments doctors might suggest as a consequence of those tests—can range from mild anxiety to outright bodily mutilation.
For reasons that I will discuss, false positives are a frequent problem in screening tests, and as such Mlodinow’s story helps illustrate the core issue in the third chapter, “Snowball in a Blizzard,” which looks at the thorny issue of mammography. Although mammograms continue to be regarded as one of the most important ways in which women can have an enormous positive impact on their health, the data suggest a more nuanced reality. In large part this is because the technology can detect breast cancer before it becomes clinically apparent, but uncertainty creates false positives, and women whose mammograms are falsely positive can suffer serious harm. Thus, ascertaining the true value of mammography involves weighing these two opposing variables. I will demonstrate, by looking at some sample data, the relative size of the benefit, as well as the risk.
The middle chapters of the book are mainly concerned with uncertainty in treatment. Chapter 4, “The Pressures of Managing Pressure,” looks at recent guidelines for treating hypertension and how uncertainty divided expert consensus in a fairly dramatic manner. Chapter 5, “Lyme’s False Prophets,” investigates a different set of expert-driven recommendations, which formed a kind of mirror image of the hypertension guidelines: although the expert consensus about Lyme diagnosis and treatment is absolute, the popular perception is that there is great controversy. “Lyme’s False Prophets” looks at how this public confusion arose through the Internet, various advocacy groups, and at least one powerful politician.
Chapter 6, “The Origins of Knowledge and the Seeds of Uncertainty,” considers how uncertainty forms a structural component of drug trials. I will explore two of the biggest blockbuster classes of drugs of the modern age: lipid-lowering statin drugs such as Lipitor and the antidepressant class of drugs known as SSRIs, such as Prozac. In both cases, I’ll put them under a microscope to see what we do and don’t know about what these drugs can offer to patients and consider the impact that uncertainty has on the term “effectiveness” in relation to drugs. Chapter 7, “The Correlation/Causation Problem,” evaluates ways other than drug trials that we learn about (or fail to know about) a drug’s usefulness. That is, although drug trials produce as a rule the most ironclad data about how good a drug can be, there are other methods for assessing a drug’s effectiveness, and these methods are subject to their own kinds of uncertainty. I’ll consider some of the major challenges involved in interpreting “retrospective” data.
Finally, I will briefly look at the role media plays in shaping our attitudes about medicine either by emphasizing or disregarding uncertainty. We live in an age of unfettered access to all sorts of media, and yet whether one is watching a local television newscast or reading the latest online health report, a good number of stories follow broadly similar patterns, frequently leaving consumers overestimating medicine’s miraculousness on the one hand or overscared by the system on the other. But I’ll also examine one crusader for health media and his organization’s vision for how the media can provide a more balanced picture of what modern medicine has to offer without too much fuss, if they would only listen.
After we’ve gone on this tour, I’ll consider ways in which the average person might benefit from an increased understanding about these concepts because the topics driving health care today will surely be different not long after the publication of this book. Lastly, in the Appendix, I will explain in a very nontechnical way some of the mathematical concepts that underpin the discipline of biostatistics, using some of the studies we have looked at as models for understanding such concepts without using equations.
Uncertainty pervades medicine: surgeons as well as psychiatrists must cope with its presence, whether they are aware of it or not. Problems that arise from uncertainty can be found in the hospital corridors, the pathology lab, the nursing home, and in urgent telephone calls from sick and worried patients. Nearly all exercises in clinical judgment involve incorporating uncertainty into equations of medical reasoning—a variable that, like Einstein’s cosmological constant, cannot be stamped out no matter how much brainpower is brought to bear. By developing an appreciation for uncertainty, we can get at the heart of many of today’s medical mysteries. By bringing uncertainty into open discussion, we can assess the real value of mammograms, recognize the hype of so many medical reports, sense when to push a physician for more testing, or resist a physician’s enthusiasm when other tests or treatments are being offered.
Ultimately, appreciating the subtleties and parameters of uncertainty allows patients and family members to be empowered. I am writing this book to help people understand uncertainty to help them navigate the swift currents and roiling waters of modern medicine. I cannot promise to translate the often inscrutable language of physicians and the medical research that is their touchstone, but I can attempt to give people a tool by which real communication can take place.
Nobody Knows Anything
It may be unsettling to a reader thus far unaccustomed to these concepts to be told that uncertainty is central to modern medicine. A sense of despair can set in when discussions of probability and statistics take center stage in the doctor-patient interaction. Frank admissions of uncertainty can often be met with irritation, because the idea that a test doesn’t provide an unassailable answer that describes a crystal-clear reality is so foreign to many people. Some may have the emotional urge to conclude, after reading thus far, that these tests are pretty much worthless and that, in the immortal words of screenwriter William Goldman, “nobody knows anything.”
But this book is not a jeremiad. The nihilism of “nobody knows anything,” although emotionally satisfying on a certain level, is just that: an emotional response, a spasm of frustration with a healt
h-care system that is mightily complicated enough, to say nothing of expensive, bureaucratic, and frequently impersonal. Only by stripping away the layers of misunderstanding about what medicine is and how it works can patients and families begin to be their own best advocates. Uncertainty is far from the only area in which misconceptions exist, but I would argue it is a critical area, and grasping it might just help people avoid some of the more unpleasant shocks that medicine is capable of delivering.
Indeed, the point of highlighting all these various instances of the limits of our medical knowledge is to demonstrate that these can be teaching moments—occasions where we can illustrate what’s at stake in a medical decision and how we think about a problem. Are the stakes high or low? Are the repercussions of a decision significant or trivial? And is the evidence supporting a given decision overwhelming, minimal, or somewhere in between? By opening up about uncertainty, we are championing patient autonomy, rather than arrogantly flicking it away as an irritating feel-good ideal.
This book is hopeful in its outlook, which I ask readers to keep in mind if they find themselves thinking in the early chapters how deeply flawed our medical practices truly are, and how foolish our certainties. My goal is to offer you a vision: read this book and you will learn something to improve your life and deepen your understanding of the process of medicine. I want readers to see how embracing uncertainty allows for more humane treatment, less anxiety, and better care. But to do that we will need to confront some sobering realities of our modern medical system. It may require the periodic deep breath and the awareness that acquainting yourself with this medical machine can occasionally make for bleak reading. Have faith, for there are rewards in knowing and understanding. There is a tangible and powerful light at the end of the tunnel.