Tuesday, July 31, 2012

Book Review: How Doctors Think by Jerome Groopman

How Doctors Think
By Jerome Groopman

“Different doctors have different styles of practice, different approaches to problems.  But all of us are susceptible to the same mistakes in thinking.” 
              -  From How Doctors Think  by Dr. Jerome Groopman.

Dr. Jerome Groopman is an oncologist/hematologist and chief of experimental medicine at Beth Israel Deaconess Medical Center in Boston.  He has written several previous books and many magazine articles on medicine and biology.  These have appeared in “The Boston Globe Sunday Magazine”, “The New Yorker” and “The New Republic”, among others.  My opinion is that How Doctors Think, his latest book, should be required reading for every medical student, intern, resident and practicing physician.  It wouldn't be a bad idea for patients who wish to better understand their medical care and insure that they are being treated for the proper diagnosis to read this as well. 

The author notes the growing public demand for “quality of care” and reduction of medical error, citing recent reports from The Institute of Medicine and the National Patient Safety Foundation.  He notes, however, that medical error differs from misdiagnosis.  “It (misdiagnosis) is a window into the medical mind.  It reveals why doctors fail to question their assumptions, why their thinking is sometimes closed or skewed, why they overlook the gaps in their knowledge.  Experts studying misguided care have recently concluded that the majority of errors are due to flaws in physician thinking, not technical mistakes.  As many as 15 percent of all diagnoses are inaccurate, according to a 1995 report which matches classical research, based on autopsies, which shows that 10 to 15 percent of all diagnoses are wrong.”

What are the flaws in thinking that lead to these misdiagnoses?  Dr. Groopman, referencing experts in cognitive theory and psychology, discusses three main categories of flawed thinking.  The first is termed “anchoring”.  This is when the physician grabs on to the first symptoms discussed by the patient and comes to a snap judgment.  The classic example of this is the overweight patient who complains of heartburn and is diagnosed with esophageal reflux when he actually is having angina.  The second flaw derives from the first and is termed “attribution error” which is a sequence of diagnosis and treatment begun with anchoring and continuing because of the doctor’s stereotyping of the patient.  Again, the physician has formed a quick first impression and doesn’t deviate from that assumption.  The third and final flaw is called “availability error,” that is, when thinking is clouded by recent experience.  In this example, the last thirty patients with this symptom have had reflux disease, so that must be what this patient has as well. He illustrates each of these cognitive errors with very compelling true life case histories, many from his area of expertise, oncology. 

            Dr. Groopman also has unkind things to say about how physicians are taught to diagnose and treat.  He dislikes diagnosis decision trees and treatment algorithms, feeling they limit thinking and promote another form of error which is termed confirmation bias.  In this setting the physician makes the data he obtains “fit” the algorithm which often leads to misdiagnosis and failed treatment.   The author is also not a fan of evidence based medicine, feeling that it relies too much on numbers which may be generated by flawed or limited technology.  He also feels that strict use of evidence based medicine leaves no room for intuition, “hunches” or thinking outside of the algorithm.

            The author also warns against what he terms “diagnosis momentum”.  This situation occurs when a patient is given a diagnosis by a specialist or academician and that diagnosis is never called into question by other physicians.  The correctness of this diagnosis is never challenged, even if subsequent data may contradict it.

            In the most important chapter in this book Dr. Groopman decries the lack of history taking skills in modern physicians.  He quotes William Osler: “If you listen to the patient, he is telling you the diagnosis.”  One study is cited where the average time from the first open ended question from the doctor to the time of the first interruption by the doctor was 18 seconds.  “Technology has taken us away from the patient’s story” decries Dr. Groopman.  Further: “Competency is not separable from communication skills.  The art of medicine is the sensitivity to language and emotion that makes for a superior clinician.”  Profound words, those.

            Dr. Groopman suggests three questions which patients should ask their doctors when they feel they are not responding to treatment.  These are:

   1- “What else can it be?”   This question should prompt the doctor to pause, think again, and extricate himself from a cognitive trap
   2- “Could two things be going on to explain my problem?” This addresses the pitfall of “satisfaction of search”, so if the physician stopped looking after finding an initial abnormality (Ockham’s Razor), he or she will now consider multiple causes for persistent symptoms.
   3- “Is there anything in my history, physical examination or lab tests that seems to be at odds with the working diagnosis?” This question is an important safeguard against confirmation bias (making the data fit the diagnosis).

All of these seem to be reasonable questions a patient can use to refocus their diagnostic and treatment process when things are not going well.

            I have only two complaints about this book.  The first is that the author only spends one paragraph describing “defensive medicine” and the impact of litigation on decision making.  There is no denying the reality that the shadow of malpractice litigation falls on every medical decision made and I think the author underplays it.  The second complaint that I have is that he glosses over the business side of medicine.  He does spend some time describing economic pressures which have dramatically increased the number of patients physicians must see daily just to remain economically viable.  This obviously decreases the amount of time spent with each patient and increases the likelihood of cognitive as well as other errors.  He never really delves into the cost of following “hunches” or thinking outside of the algorithm.  The cost of searching for obscure or unlikely diagnoses is never really brought into the argument. 

            This book has already changed the way I practice.  I find myself reviewing lab and radiology data looking for inconsistencies with my diagnoses.  I am wary of “diagnosis momentum” and find myself asking a long-time patient the question “Let’s start over and tell me how all of this started again?”  I try not to stereotype patients and try to look past “non-compliance” as a diagnosis and try to find out why a patient is non-compliant with their treatment regimen.  I try to let a patient talk for more than 18 seconds.  Read this book, you’ll be glad you did.


  1. May I send you a digital copy of my novel, Reluctant Intern, to read and possibly review on your blog?


    Addison Wolfe never wanted to be a physician. He wants to be an astronaut. NASA turned down his application, forcing him to seek employment as a doctor. The problem with obtaining a physician's license is the need to complete an internship to acquire one. Wolfe finds himself in an undesirable rotating internship in a very busy public hospital. Inexplicably, the Director of Medical Education seems to have developed an instantaneous dislike of him and the remainder of the internship class. Another mystery is why an attractive female physician expresses a romantic interest in him on the first day of internship.

    “The absolute worst time to go to a teaching hospital as a patient is the month of July. Recent medical school graduates, known as doctors, start their real training on July first. They don’t know anything. They don’t get any sleep. They are underpaid and overworked. Their stress is at catastrophic levels. Is it any wonder they make mistakes?” – Anonymous

    “In local news today,” the reporter said, “state and federal authorities are in the process of taking into custody the entire intern class at University Hospital in Jacksonville. Officials cited the number of deaths attributed to this class as the reason. It seems that wrong doses of medications, inappropriate surgeries, failure to diagnose lethal conditions, and other mistakes have led to hundreds of deaths….”

    “The overdose?” Wolfe asked.
    “Yes,” Dr. Rubel replied, “that will be her legal cause of death, of course. The real cause of death was the autopsy. Barbiturate overdose, followed by refrigeration outside and then here in pathology, slowed her metabolism down. She was actually alive when they started the autopsy. The flexing of her limbs when the saw touched her brain happened because of nerve conduction, brain to extremities. But it was too late; we cannot put her back together. A hard lesson for those poor boys to learn. You, too, gentlemen. It is also true for those who are clinically dead from exposure or drowning. Remember this: a patient is never dead until he is warm and dead. Don’t forget that!”

    The senior resident started his description, “EMS responded to a report of a cardiac arrest at 1:07 a.m. in Junior’s Topless Bar, on East Bay Street….”
    Figueroa again jumped to his feet. “What is this, a bad joke?” he asked. “Two EMTs walk into a bar…. Let’s be reasonable, guys. The most likely reason for needing a paramedic in a bar at 1 a.m. is a knifing or a gun shot wound, not a heart attack.”
    The autopsy and x-rays were condemning. The thirty-nine year old, black male had no history of heart disease. No medical history of any kind. He did have a bullet entrance wound to the back of his head with no exit, bullet still in his brain.

    The patient was a massively obese woman who complained of a headache. The intern knew only that she was complaining of a headache and had requested aspirin. Extremely busy, and assuming the nurse would let him know if it were not a good idea to give the patient aspirin, he quickly flipped to the order page and signed the order that had been written by the nurse. Figueroa asked the intern if he had talked with the patient. No. Had he examined the patient? No. Had he even skimmed the chart? He had not. He asked if he knew what allergies the patient had. The intern did not know. At the time he approved the order for aspirin, did he realize the patient was on warfarin, another clotting inhibitor? No. Did he know that aspirin also inhibited platelets and clot formation? Yes. Did he know the patient had a history of blood clots? No. Did he suppose that a blood clot in someone's brain, or a ruptured berry aneurysm in the same area might cause headaches? Yes, he knew that. The autopsy pictures revealed stenosed carotid arteries, two small clots in the patient's brain, and massive bleeding from a ruptured berry aneurysm.

    1. Sure, send it on - can't guarantee when I will get to it, but I'll try! Thanks for reading my blog!