A middle-aged man develops chest pain at home. Minutes after calling 911, he’s in an ambulance, whizzing through traffic to the nearest emergency room. The paramedics radio ahead, and by the time the patient arrives in the ER, the hospital’s heart attack team has been activated. A stat electrocardiogram shows an ST-elevation myocardial infarction (STEMI), and suddenly It’s Showtime: the patient enters the “chest pain pathway,” automatically triggering a pre-prescribed set of actions: medications, blood tests, oxygen, and more. The cardiologists, nurses and technicians congregate in the cath lab, waiting for the handoff. Nothing is left to chance: the STEMI team even has a special key that allows them to commandeer the elevator.
This is life-and-death choreography – George Balanchine meets Dr. John Carter – and it works. In the old days, it might have taken 3 or 4 hours to get a catheter into the patient’s coronary artery. By following this cookbook approach, the patient’s blocked vessel is stented open within 90 minutes. And this scenario (and similar ones for patients with strokes and sepsis) plays out tens of thousands of times each day in the U.S. In a crowded ER, with its waiting room full of patients and where triage is the name of the game, nobody questions whether this is the right approach. It saves lives.
But, as the saying goes, when you’re a hammer, everything looks like a nail. In their new book, “When Doctors Don’t Listen: How to Avoid Misdiagnoses and Unnecessary Tests,” Drs. Leana Wen and Joshua Kosowsky, emergency medicine physicians from Brigham & Woman’s Hospital (Wen, impressively, is still a resident; Kosowsky is an attending), argue that this cookbook approach has infected not only emergency medicine but all of medicine, rendering healthcare not only impersonal but unsafe. They write:
“Algorithms” and “pathways” have proliferated in ways that have reduced each person’s unique story to simplistic recipes…. Doctors end up adopting the role of automaton, following recipes and doing as directed, but no longer empowered to listen, to think, to diagnose, and to heal.
Wen and Kosowsky argue that we have entered an “era of depersonalized diagnosis,” in which we have…
come full circle, back to the time of the black box, when a visit to the shaman turns up a solution in the form of a random incantation or ritual. Today, doctors order a CT or MRI and expect the patient to emerge from the (literal and figurative) black box with an answer… There’s no thinking necessary. But at what costs to patients?
How does this actually play out? Take the patient who presents to the ER with a headache, some abdominal discomfort, and a little chest pain. Within minutes of arrival, the patient is forced, sometimes nearly bullied, onto a clinical pathway, as in “would you agree that your belly pain is your biggest problem?” The authors describe one poor college kid who came to the ER hung over after a night of partying. After she made the mistake of saying she had the “worst headache of her life” (which reflexively triggers the “Rule-out Subarachnoid Hemorrhage” pathway), she ended up with a CT scan, which was predictably negative. She would have ended up with a lumbar puncture too, if she hadn’t escaped out a side door.
I found this indictment of cookbook medicine-run-amok both interesting and, for the most part, persuasive. Where the authors began to lose me was with their other central argument: that the only way to improve this situation is for patients to serve as the agents of change. Accordingly, the book’s audience is not physicians – whom the authors believe are hopelessly stuck – but rather patients, whom they believe can force doctors to slow down and listen better… if they are suitably empowered and coached. “Make your doctor into the ideal doctor that you always wanted by showing them the way!” they write, with characteristic (over)exuberance.
Given the audience, I liked the book’s conversational tone; I was even okay with the potentially awkward device of having the two authors speak independently at times (Dr. Wen: “Supposedly Yolanda answered ‘yes’ to everything, and the one time she said ‘no,’ it turned out not even to be accurate!” Dr. Kosowsky: “That’s exactly the point. Close-ended questions have limited value…”). Complex concepts, including iterative hypothesis testing, Bayesian reasoning, and anchoring bias, are described simply and effectively, usually without their jargony labels (probably a good choice on the authors’ part). The reader is regaled with lots of clinical vignettes, many of them versions of this tale: A patient comes to the ER with a complaint, and ultimately receives the wrong diagnosis, or no diagnosis at all – a victim of cookbook medicine. In the end, the answer was right there all along, if only the doctor had taken the time to listen to the patient’s story.
And the book is chock-full of patient-oriented tools. After the authors describe an ineffective doctor-patient encounter, readers are presented with an alternative script, the things the patient might have said to better get her points across. Wen and Kosowsky also include a number of patient-oriented tools, including the “8 Pillars to Better Diagnosis,” “21 Exercises Toward Better Diagnosis,” and worksheets for patients to write down their stories before seeing their doctor. They do a nice job answering obvious objections to their recommendations, including the “I’m worried about being sued” argument that many ER physicians will raise when told that they should listen more and scan less.
While “When Doctors Don’t Listen” has a lot going for it, there are major problems – this patient needs resuscitation. Much of the writing is pedestrian, repetitive, and breathless (the term “common sense” appears 42 times, and if one got a nickel rebate for every exclamation point, the book would be free). Some of the devices are a bit cutesy for my taste (“+911 Action Tip”). In a book whose subject is preventing errors, it was disheartening to find at least a half-dozen typos (“Acknolwedgments”, “a big burley man”), and a few substantive errors (Dr. Herbert Chase is not “the founder of [IBM’s] Watson,” he’s a Columbia University medical informaticist consulting with IBM to develop the computer’s medical functions). The bibliography is weak; it gives little evidence that the authors have read many prior important works in diagnosis, such as Groopman’s “How Doctors Think,” Kahneman’s “Thinking Fast and Slow,” or most of the seminal academic articles on diagnostic errors and reasoning. Also, the book’s viewpoint is truly ER-centric, although one can see how the proliferation of disease-specific quality measures will generate more and more of this type of thinking (think: sepsis bundle) in other parts of the healthcare system. While granting that poor doctor-patient communication is a problem everywhere, the lens here is relatively narrow, a problem that the authors should have acknowledged and mitigated.
Moreover, the stories begin to wear thin after a while, particularly since, in them, patients always seem to know what is wrong with them, and the cookbook-addled doctor would also get it right if only he junked the pathway and listened more carefully. This heroic patient/clueless doctor narrative may feel good (to some), but it’s not always true. Every now and then, the patient who doesn’t place much stock on his chest and back discomfort dies of an aortic dissection. And sometimes the patient who goes on and on really is a blabbermouth who does need to be gently guided to focus on what really matters.
More fundamentally, while I agree with many of the suggestions to physicians (such as trying very hard to avoid leading questions and giving patients time to tell their stories), the admonition that we should be completely open with patients about all the diagnoses we are considering (“I think you have pneumonia, but I can’t be sure you don’t have lung cancer,” they’d have doctors say) and what we’re looking for as we go through the physical examination is a recipe for unproductive anxiety, for both patient and doctor. And relatively few patients – even after completing all of the book’s exercises (and really, who is going to do that?) – will be able to be as blunt and forward as the authors recommend. (To the ER doc running out of the room after recommending a CT scan, they advocate that the patient say, “Actually, doctor, I’m not OK with that”.)
The book’s emphasis on what patients can do to protect themselves is a clever premise, and – as they say in the PR business – a nice “hook.” It will boost sales, it will get the authors booked on the morning talk show circuit, and it will excite people who mistrust the medical profession or who yearn to regain some control in what can feel like a lopsided power relationship. Yet to me this emphasis seems fundamentally misplaced, and even risks blaming the victim. Patients can help, sure, but they are not the crux of the problem, nor should they be the focus of our search for solutions.
The key test, of course, is whether this approach will actually improve patient safety? For most patients, sadly, I doubt it.
Would I buy this book? If I were a patient who was having lots of medical encounters (particularly ER visits) and was invested in trying to improve things, I might find it useful. And emergency medicine physicians, and those who set ER policy, would do well to reflect on its cautionary notes regarding the slippery slope of cookbook, algorithmic medicine. But for everyone else, probably not.