How Doctors Think, by Jerome Groopman
(Houghton Mifflin, 320 pp., $26)
Better: A Surgeon's Notes on Performance, by Atul Gawande
(Metropolitan Books, 273 pp., $24)
Medicine today is both a wonder and a disaster. The roster of marvels is extensive: technological advances in imaging, from the heart to the uterus, that maximize diagnostic information and minimize side effects; pharmacological agents, whether for cancer or for AIDS, that may not always cure patients but dramatically increase their longevity and improve their quality of life; surgical interventions that replace fractured hips or diseased kidneys and restore patients' capacities. But no less formidable is the catalog of contemporary medicine's problems: pressures from managed care companies, practice plans, and insurers that force physicians to see too many patients and to curtail tests and referrals; a plague of medical errors, which kill thousands of patients annually; financial incentives by drug and device makers that often lead physicians to prescribe too much, too little, or incorrectly.
The discrepancy, however pointed, does not affect everyone equally. Would one rather be a patient today than a patient fifty years ago? Of course. But would one rather be a doctor today than fifty years ago? That question is more difficult, and the answer might well be no. Jerome Groopman's engaging and insightful book reckons with all this and a good deal more. In an intriguing way, Groopman himself straddles the past and the present. Those of us who have had an old-fashioned, caring, and devoted doctor--and I am lucky to be in that company--will recognize him as a kindred spirit. At the same time, there is nothing quaint or antiquarian about him--he is a specialist (hematologyoncology) working in a very high-powered medical environment (Harvard, Beth Israel Deaconess). And so he is particularly well-positioned to provide a primer on the best and the worst in contemporary medicine.
Groopman's book shines in the examining room, brilliantly analyzing the performances of doctors and patients and presenting them in compelling narratives. His overarching goal is to educate both groups so as to improve the likelihood of accurate diagnosis and effective treatment. Yet it must also be said that he is more adept at identifying particular problems than at offering systemwide solutions. In fact, he is highly suspicious of such solutions, persuaded that they will interfere with physicians' best judgments and patients' welfare. His opposition, as we shall see, is not always persuasive, but his arguments make the book the more important to read and grapple with.
Groopman begins by parsing "how doctors think," emphasizing the critical need to counterbalance the predictable and perhaps unavoidable tendency on their part (a tendency we all have in our own fields) to fall into a fixed pattern of decision-making, and thus to make a diagnosis on the basis of initial impressions of the most prominent clinical symptoms. "Physicians," Groopman advises, "should caution themselves to be not so ready to match a patient's symptoms and clinical findings against their mental template or clinical prototypes." The tendency is enshrined in the maxim that senior physicians pass on to medical students and residents: when you hear hoofbeats, think horses, not zebras. But this is precisely the wrong message to deliver, Groopman insists, and he recounts several case histories to make his point. The young man found sleeping on the street, disheveled and uncommunicative, may not be an alcoholic or a junkie--he may be a diabetic in a coma.
To think differently about the tenth case when you have been on target the first nine times, Groopman proposes two strategies. First, stop putting down the zebra hunter. Value the effort to go outside the boundaries. Second, Groopman wants to turn patients into a countervailing force--and his book represents an attempt to do just that. He sets forth a series of probing and genuinely unsettling questions that patients should ask their doctors. Ask your doctor what else it might be--regardless of how confident he is in his diagnosis. Ask what is the worst it could be--again to shake him loose from his first impressions. Whether in an emergency room or a physician's office, do your best to "slow down the doctor's pace and help him think more broadly." Search for the "gaps in his analysis." Insist on giving him a complete case history, because hidden in your details might be clues that were missed in the initial take. Ask: "Is there anything that doesn't fit?" "Is it possible that I have more than one problem?" In sum, Groopman wants both doctors and patients to do all they can to "break out from the box."
But more than psychological and pedagogical dynamics prompt physicians to think horses, not zebras. The workaday environment, the circumstances under which medicine is now practiced, is equally responsible. Its most prominent feature, as Groopman fully recognizes, is the effort by managed care and practice organizations to compel doctors to set strict time limits and resource priorities with their patients. Thirty-minute slots are a luxury: a recent survey reported that 50 percent of patients seeing an internal medicine physician were in the office for fifteen minutes or less. The organizations send physicians a monthly printout informing them of the number of MRIs each ordered, how many times they referred patients to a specialist, and how their rates compared with their colleagues'. The message is unmistakable: do not be outliers. And to make practicing medicine even thornier, some groups gather information on patient satisfaction and use the results in determining physicians' reimbursement. How long did you have to wait before the doctor saw you? Were all your questioned answered? From the doctor's perspective, it seems a "damned if you do, damned if you don't" situation--which helps explain why so many of them do not want their children to follow in their professional footsteps.
Groopman does not have a lot to say on how to reconcile the tension between best practices and the structure of medical care delivery. Indeed, to the extent that patients follow his advice and start asking unconventional questions, time demands will certainly increase. He declares that "doctors and patients should push back." You cannot do medicine "like a race being run." But precisely how they should do this is unexplored. This is not to fault Groopman for the health policy book that he did not write. Since his advice runs so counter to current examining-room realities, one desperately needs counsel on how to reconcile the personal demands of the sophisticated patient with the organizational demands facing the practicing physician.
Groopman does provide vignettes of a few physicians who have resisted the demands of the system, but their career paths are idiosyncratic and cannot serve as a model for colleagues. Victoria McEvoy, a pediatrician, worked in a busy group practice outside of Boston. The combination of seeing two dozen or more children and their parents each day, along with incessant night calls, produced "relentless work and sleep deprivation." McEvoy believed she was on the way to burnout, losing the crucial ability to focus her mind and to avoid complacency: "It is almost always a virus or a strep throat. They can all blur. But then there is that one time it's meningitis." So she left the group practice, took a half-time administrative post with Boston's Partners HealthCare, and devotes half- time to clinical practice, strictly limiting the number of patients she sees. Groopman comments that the arrangement has not decreased her income--but how many slots are there at Partners HealthCare for administrators, and how many can cut practice time in half and still make the same living?
Or take Groopman's example of another Boston physician, Judy Bigby (an exceptional advocate whom I have met). She, too, divided her time between administration (two-thirds) and practice (one-third). This way she set her own clinic schedule. Although her colleagues are supposed to see patients every fifteen minutes, she gave them more time. Do the bean counters object? "Not anymore. I think if I were a full-time clinician someone might. But I have reached the point in my career where this is simply the way I want to doctor." Yet clearly not everyone has Bigby's options. Most have to bow to the bean counters.
What else might a doctor do to limit the number of patients? Groopman mentions, without comment, setting up an exclusive "concierge medicine" practice. The physicians charge a premium over insurance coverage and usually insist on cash or credit card payments; they leave it to the patients to negotiate reimbursement from insurance companies and other third parties. The concierge doctors may also charge patients an annual retainer fee, taking their cues from corporate law firms. Patients who can afford the costs may think concierge medicine is a good thing: their phone calls will be returned within two hours, and their office visits will be longer. But data on whether concierge medicine actually produces better outcomes is not available, and the number of such practices is quite small--the very best physicians are not likely to practice medicine in this style. Whatever the findings, as a model for American health care or as a model for medical professionalism, concierge medicine falls far short. If this is what it will take to slow the pace of medical examinations, we are in even worse shape than we thought.
As concerned as Groopman is about the power of habit and the constraints of time to distort the diagnostic judgments of physicians, he is even more alarmed by the growing commitment among medical experts to the idea of evidence-based medicine. At first glance, his resistance might seem odd. After all, the principles--as developed over the past decade and defined by David Sackett, one of the most avid proponents--seem incontrovertible. Evidence-based medicine calls for "the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients." The gold standard is findings on efficacy based on randomized, double-blind clinical trials; but when such data is unavailable, clinical decisions should reflect scientific evidence demonstrating that benefits outweigh risks. The best evidence then serves as the basis for clinical guidelines or algorithms, usually formulated by medical specialty societies. The American College of Cardiology has issued guidelines on when to intervene with coronary artery bypass graft surgery; the American Academy of Pediatrics, on treating attention- deficit/hyperactivity disorder; the American College of Obstetrics and Gynecology, on the use of ultrasound in pregnant women. Which of us, you might ask, would want to be diagnosed or treated by any other criteria?
Knowing all this, Groopman is still disturbed that evidence-based medicine has become a "canon in many hospitals." "Clinical algorithms," he contends, "discourage physicians from thinking independently and creatively." Of course doctors should reckon with research findings, "but today's rigid reliance on evidence-based medicine risks having the doctor choose care passively, solely by the numbers." When doctors turn into computers, personal experience that draws on the clinical encounter is lost. Numbers cannot and should not substitute for the art of medicine.
Again putting names to ideas, Groopman tells us about those who get it right. One is Donald Schon of MIT, a thoroughly unconventional and wide-ranging thinker who contended that large databases inevitably break down when experts confront individual cases. Another is Stephan Nimer, a hematologist at Memorial Sloan-Kettering, who takes as his mission teaching young physicians not to allow algorithms to think for them. The algorithms, he says, give too little attention to patients' values and lifestyle choices: "It's a static way of looking at people. Strictly speaking, it's correct. But clinically speaking, it's wrong."
In one sense, the debate over evidence-based medicine is a century old. It first appeared in the 1920s, when the laboratory began to achieve its ascendancy in medical education and practice. To the researchers, clinicians put too much faith in outcomes from a few cases: if an intervention worked once or twice, they believed it would always work. To the practitioners, lab scientists belonged in the lab: they might know how to treat mice, but they knew nothing about the care of sick people, and they denigrated the wisdom achieved at the bedside. The controversy was essentially about epistemology and authority--and it continues to this day. What source of knowledge is to be most respected? What is the relative worth of clinical experience against carefully controlled trials? Then, as now, for every Groopman there was a Sackett.
But some things have changed, and Groopman's objections notwithstanding, the likelihood is that, for a variety of reasons, evidence-based medicine will continue to expand its influence. For one, medicine confronts an extraordinary information overload, making it impossible for general practitioners--indeed, even specialists--to keep up to date. As Stefan Timmermans and Marc Berg note in The Gold Standard, their perceptive book on the history of evidence-based medicine, when there are 20,000 medical journals publishing two million articles a year and reporting on more than 250,000 controlled trials, and when in a five-year period 14,000 articles are published on the use of one type of medication for hypertension, some mechanism must to be developed to track, to evaluate, and to communicate the findings.
For another, medical decision-making is marked by extraordinary variations. As John Wennberg has demonstrated, where you live all too often determines what treatment you get. We may think of medicine and medical information as a national enterprise, but an astonishing amount of practice follows local mores. One example will suffice: rates of radical mastectomy per 1,000 Medicare enrollees range from 1.4 in San Francisco to 2.0 in New York and on to 3.1 in Green Bay. And the variation in this procedure can be matched by almost any other you can name. The term "standardization" can connote a mechanistic rigidity, but if it means that the care delivered in Wisconsin is not significantly different from the care delivered in New York, then standardization should be and will be welcomed.
Moreover, medical errors are far too common. The 2000 report of the Institute of Medicine (IOM) sent tremors through lay and professional audiences. Its precise numbers and methods of calculation have been debated and disputed, but the IOM found that medical error caused between 40,000 and 90,000 deaths annually, making it a greater killer than automobile accidents. While the IOM focused most of its attention on system errors (for example, putting very different medications in almost identical bottles), there is no minimizing the amount of outright medical incompetence. A recent RAND study concluded that less than half of patients presenting with myocardial infarction (the classic heart attack) received the right medication (beta blockers), which reduces by almost one-quarter the number of subsequent deaths. Outcomes were no better when the RAND investigators looked at treatments for diabetes, asthma, elevated cholesterol, or urinary tract infections.
In all, Groopman's emphasis on physician autonomy and making independent and unexpected diagnoses is crucial, but his concern--and it is not comforting to say this--may be more relevant among elite practitioners than among average ones. (Or as the old joke has it: What do you call the person who graduates last in his class? The answer: Doctor.) When a majority of patients are not receiving standard care for standard conditions, evidence-based medicine has much to commend it. To be sure, the floor should not become a ceiling. No one wants algorithms to stifle clinical discretion, intuition, and acumen. But at the moment the need to set the floor seems far more urgent.
Groopman considers one other major impediment to best medical practices: the marketing strategies of pharmaceutical companies. Even those familiar with the machinations of drug company representatives will be taken aback by his opening vignette. A rep from a company (not identified) that makes a testosterone hormone product accosted Karen Delgado, an endocrinologist, outside her office to complain that she was not prescribing his product to her elderly male patients. "I want you to write three prescriptions for the next month," he demanded. She was dumbstruck, but one of her senior colleagues, who received research support from the same drug company, apparently was not. When he and Delgado met the next day, he asked her outright to see the rep. Again, she refused. Delgado did not believe that her colleague was pressuring her because of the research money he received; she preferred to think that he was a true believer in testosterone as an anti-aging medicine. But she was being overly kind. True believers in the efficacy of testosterone to prevent aging are scarce. More, there are a number of such products on the market, and he was not urging her to prescribe more of the drug but to spend time with a particular rep.
Drug companies are masterful at using a variety of financial incentives to get doctors to prescribe more of their product. These include giving doctors gifts (trivial and otherwise), food (from sandwiches to banquets), honoraria for speeches, junkets to meetings, and stipends for service as a consultant or a member of an advisory board. As Groopman astutely adds, surgical-device manufacturers are no less adept, and, worse yet, surgeons have their own financial self-interest: they earn a great deal more by operating than by not operating. Drawing on his own experience as a patient, Groopman takes his examples from back pain and spinal surgery. He is certain that many diagnostic scans are prescribed because neurologists are very well reimbursed for performing them, and that unnecessary spinal surgeries are done because surgeons are even more handsomely reimbursed for them. From the perspective of device-manufacturing companies, the greater the number of surgeries, the more rods, screws, and plates they sell. And the hospitals and medical centers are no less eager. More surgery translates into higher bed-occupancy rates and thus higher reimbursements. (The profit center for hospitals is surgery, which is why they recruit surgeons the way sports teams recruit superstars.) "The current culture of medicine," Groopman concludes, "fosters lucrative networks of referrals and procedures but discourages critical examination of their value. "
What might be done to curtail, if not to eliminate, such conflicts of interest? Groopman summarizes the recommendations of a task force that I happen to have co-chaired on the financial relationships between drug companies and physicians. Our report appeared in the Journal of the American Medical Association (JAMA) in January 2006 and received substantial media and academic attention. Composed of physicians, policy analysts, and social scientists, the committee started out as moderates--gradual abolitionists, if you will--on the question of physicians taking drugcompany largesse. We ended up, after extensive discussions and literature review, as immediate Garrison-type abolitionists. Our bottom line was that there is no such thing as a free gift. We urged academic medical centers to ban all gifts, food, and travel stipends to doctors, fellows, and residents. We would not prohibit drug-company research support or consultation agreements--sharing and creating scientific knowledge is essential-- but we called for transparency, so that colleagues and patients would know who was receiving such funding and how much. Medical schools often do ask for "disclosure" of such relationships, but not of the exact sums, and they keep even this limited information private. For their part, drug companies cloak their various efforts as educational. But when we asked some companies if they would give their funds to a central office, rather than to a prescribing physician, they responded candidly that they would not. You want us to give charity, they said, but we reserve that for the third world: in the United States we do marketing.
Groopman observes that his own hospital does have conflict-of-interest guidelines, but they occupy a gray zone, not nearly as strict as our JAMA recommendations. He is not convinced that medical centers and hospitals will adopt tougher rules, or that physicians will become more scrupulous. Although he is as offended by many of these practices as we are, he does not issue a clarion call for medicine to clean its own house or, failing that, for outside regulators to come in and do the job. Instead, he counts on the educated patient finding the scrupulous doctor. Making an informed decision now requires patients to learn "how science, tradition, financial incentives, and personal bias" shape clinical decisions. Groopman would have patients not only master the questions to ask their doctors about their medical condition but also understand the intricate financial incentives affecting medicine and the biases they produce. Put another way, they have to undergo crash courses in both medicine and business.
The chapters of Atul Gawande's uneven book are drawn from his previously published articles, most of them from The New Yorker. Many of them are well crafted, but they do not add up to a whole, and Gawande's efforts to join them are perfunctory. In the section labeled "Diligence," the chapters recount the complex military evacuation systems that increase the likelihood that wounded soldiers will get quicker treatment, and a field trip to India to observe polio vaccination efforts. In another section called "Doing Right," Gawande presents a dramatic and revealing interview with a physician who violates professional medical ethics by participating in executions. It is an exceptional piece of reporting--but as Gawande makes clear, the physician is not doing right.
Like Groopman, Gawande teaches Harvard medical students, and his last chapter comprises one of his lectures, lessons in how to be a "positive deviant. " Some of his advice is inconsequential, some is potentially mischievous. Gawande counsels the future doctors to "ask an unscripted question" to their patients--not to improve diagnostic outcomes but to cultivate the relationship. But does he really believe (the examples are his) that asking a patient why he moved to Boston or whether he watched the Red Sox game last night is going to bridge the gaps? This is doctor-patient communication brought down to pure trivia. Then Gawande goes on to tell his students, "Don't complain," and illustrates the axiom with a story. One surgeon told his colleagues over lunch that he had to overrule an emergency department physician's decision on whether to operate on a patient, that the ER doctor never bothered to call him back to review the decision and proceeded to give the patient inadequate information. Gawande concludes the tale by recording that "when lunch was over, we all returned to our operating rooms and hospital wards feeling angry and sorry for ourselves." That is all. He says not a word on what it means to encounter, to correct, or to report a medical error or outright rudeness.
Perhaps most troublingly, Gawande instructs his students simply to "change." Conceding that some past medical practices were misguided--performing frontal lobotomy to treat mental illness, or prescribing Vioxx to alleviate simple pain- -Gawande nevertheless urges them to be "an early adopter." He warns that "it often seems safest to do what everyone else is doing--to be just another white coat cog in the machine." Instead, he suggests, "find something new to try, something to change." Count it, write about it, ask people about it, and "see if you can keep the conversation going."
This is curious, and even dangerous, advice. It is nothing like Groopman's suggestion to think unconventionally. Gawande's wisdom here seems to be that one should be the first on the block to try the latest drug or procedure-- not just to stand there but to do something. This is an attitude that could undercut responsible medical practice. Doctors have more to worry about than their non-conformity or their originality. Would that Gawande had emphasized a more rigorous and decisive message: to be suspicious of new claims, to go behind the published paper to the data, to resist drug-company hype, to insist that innovations meet all the criteria of evidence-based medicine. The test of time counts for a lot, certainly when a disease is not rapidly progressing or deadly. The case can be made for early adoption in oncology, but arthritis is a very different matter--witness the Vioxx fiasco. More, new drugs are often very expensive. An older generic version may not represent an escape from the grind, but it will be far cheaper and perhaps no less effective.
In the 1970s and 1980s, it was not uncommon to hear physicians complain about the intrusions of lawyers and bioethicists upon their turf. They were annoyed by the need to establish oversight committees for human experimentation, to obtain informed consent, to tell the truth about a diagnosis at the bedside, to encourage and to respect living wills--in effect, to give up some autonomy and share power with patients. The changes that have occurred over the past ten years make these earlier encroachments seem altogether modest. Neither Groopman nor Gawande may like it, but medicine now confronts an army of regulators, some coming from within the profession, even more from without, and on occasion the insiders and outsiders unite.
One telling example of the new alliance reflects the power of evidence-based medicine. It was physicians who determined that infections from in-dwelling catheters ought not to occur. Medicare officials then took this conclusion and ruled that reimbursements would not be made to hospitals for treating such infections. So, too, many physicians abhor the drug company giveaways, and a number of them, serving as deans in medical schools, have enacted guidelines and procedures that follow many of the recommendations in our JAMA article. The list is already impressive--Stanford, Yale, Pennsylvania, Wisconsin, and Michigan, among others.
In this same spirit, several states, including Vermont and Minnesota, have passed legislation requiring drug companies to report all gifts to doctors. Owing to this reporting requirement, we have learned that psychiatrists are among the physicians most on the take, that recipients of drug-company money often sit on the state committees that determine which drugs to buy, and that drug companies appoint recipients of their funds to be researchers even when they have been disciplined for major infractions by state medical organizations. Meanwhile, Senator Grassley is eager to enact a federal gift disclosure law. He also wants legislation to help ensure that continuing medical education courses for physicians are truly objective and are free of drug-company influence. Even business corporations are intervening. A number of major companies have formed a coalition, called Leapfrog, that refers their employees to hospitals with the best staffing ratios and outcomes in particular fields. Will these efforts by insiders and outsiders raise the level of physician performance, as many of us hope, or will they stifle creativity and insight, as Groopman fears? As a good doctor will sometimes admit, only time will tell.
DAVID J. ROTHMAN is a professor of social medicine at Columbia and president of the Institute on Medicine as a Profession. He is also associate director of the Prescription Project, working to strengthen conflict-of-interest policies at academic medical centers.
By David J. Rothman