Overconfidence, anchoring errors, biases, and other factors have all led to misdiagnoses at points in my internal medicine career.
I had a conversation with a woman who was having major surgery to reconstruct her right ear after a basal cell had been removed. I casually said, “you will look like Vincent van Gogh before they repair your ear.” She said, “that was the other ear.” I replied, “actually not, I’ve seen the picture and am sure it was the right ear.” She calmly said, “no it was the left.” Annoyed now, I opened my external brain, Google, and found a picture of van Gogh and pointed out that the affected ear as the left. She countered, “he painted the portrait from a mirror image.”
Confidence is a fine trait particularly for surgeons, but overconfidence is one of several mental traps that leads doctors to make misdiagnosis and errors. Fortunately, misdiagnosis does not always lead to harm, most illnesses are self-limiting as long as we don’t prescribe a harmful treatment. I’ve often been saved by remembering that when you don’t know what to do, don’t do anything.
It’s fair to say that misdiagnosis is not usually due to a lack of intelligence or caring on the part of most doctors. It’s a combination of a broken medical system and the fact that doctors are folks with cognitive blind spots common to humans. We all have the capacity to not see what we don’t want to see. With diagnosis, that defense mechanism becomes problematic because the stakes of not seeing things clearly are higher.
Modern medicine has been severely affected by a new business model demanding productivity and quality as measured and directed by the electronic medical record and by the increasingly corporate intrusion into medical practice. This had led to severe constrains on our time spent with patients. EHRs force us to ask direct, pointed questions which must be check marked in the correct boxes. The questions are driven by algorithms created by programmers, not clinicians, and do not lend themselves to listening, contemplating, and developing a diagnosis based on experience or intuitiveness.
The first diagnosis we make when overconfidently short-circuiting listening to the patient’s story is often not the right one, as I discovered once I decided to listen more and talk less. I often change my mind during the interview about the diagnosis, as I grow more confident in what I am hearing. As a result, I order far fewer tests, scans, and drugs than I used to.
I had a young man visit me complaining about a headache of moderate intensity, that had started recently, was band-like, not throbbing, but lasted all day. He had enjoyed good health his whole life and had no prior history of headache. He started to tell me about some stresses he had but I interrupted by firing off several quick closed end questions. “Where?” “Does it radiate?” Severity?” I then did a cursory neurological exam and quickly diagnosed tension headache. Sensing his anxiety, and running late, I gave him a prescription for an MRI, scribbled the name of neurologist on a scrap of paper, and a prescription for Fiorinal.
As he was leaving, he turned to say, actually what he really wasn’t worried about the headache as much as he wanted to know if he could have a brain tumor, because his father was just diagnosed with one. The only things he was able to tell me during the visit were the answers to my closed-ended questions, so he didn’t get a chance to tell me what the issue was really. He didn’t need anything but reassurance that brain tumors didn’t usually run in families, and his symptoms were not typical of brain tumors.
My thinking process was short-circuited because it was easier to order tests, and consultations rather than take the time to listen to him. Diagnosis can be elusive but is often obvious if the patients are given enough time to let their story unfold in an unhurried and nonjudgmental manner.Doctors all have patients we bond with and we naturally don’t want them to ever have anything bad. We also think about the discomfort and expense that diagnostic tests cause and may be reluctant to order studies that may actually be indicated. This is probably why celebrities and politicians get bad medical care. The doctor is intimidated or reluctant to recommend things which may be unpleasant for their patient. This has been called “The VIP Syndrome.” It leads to either too much treatment or under treatment because of cognitive errors on the part of doctors who want to please their celebrity patients.
William was a musician who I saw from time to time. He was actually quite famous in Europe where he had several top 10 records. We often spoke about music and concerts we’d been to and had become friends.
On a Friday afternoon he came to see me for a headache. The headache had been persistent for 6 months, was getting worse, and had started to distract him. At times it was severe. I told him it was just stress. A neurologist in New York had checked him 3 months before, said it was migraine, and gave him Imitrex and Percocet, which helped a little bit. I did an exam and looked in his left eye and saw engorged and tortuous retinal veins and flattening and mild hyperemia of his optic nerve. He had no visual symptoms but the eye looked peculiar, and although I had only seen papilledema in a textbook, I considered it a possibility. I looked and looked several times, wondering.
Since it was late Friday afternoon, and I knew he had to be in New York that evening, I started to rationalize that if I ordered an MRI, it would likely cost him a fortune, it would complicate his schedule, it would annoy the radiologists who wanted to go home on a Friday afternoon, and it would probably be normal. I convinced myself he couldn’t have papilledema. After all I had never seen it, so what did I know. The previous doctor had given him Imitrex, so I renewed it.
That night he had several convulsions, was admitted to the hospital and found to have a glioma the size of an orange. He survived long enough to have surgery and radiation but died within 4 months. The delay in diagnosis did not change his course but it taught me that I should learn to trust my own instincts. I should also not let the fact that patients and providers may be inconvenienced deter me from ordering tests. Besides the VIP error, and my personal bias about William, I also fell victim to diagnosis momentum error because I assumed that since a neurologist had diagnosed migraine, that’s what he must have.Mrs. Feldman was a nervous 65-year-old woman with an unpleasant personality who made frequent visits for a multitude of minor ailments. She came to see me because of vague abdominal pain, on and off diarrhea, mucous in her stool and boating. She hadn’t lost weight, didn’t have fever or blood in her stool, but felt unwell. The word bloated stuck in my head, and after a quick exam I diagnosed her to have irritable bowel. I was already convinced that whatever she had was nothing serious based on my prior experience with her. She was advised on diet, given some medicine and a handout on irritable bowel and went home.
Two weeks later she was back because the bloating and mucous in her stool hadn’t improved. (Of course, it hadn’t, it was only 2 weeks!) I told her to stop drinking milk. She returned in a few weeks with the same complaints and I ordered an ultrasound of her gallbladder and a colonoscopy, as well as standard blood tests. They were all normal except for a few diverticula in the colon and the gastroenterologist suggested irritable bowel syndrome. She didn’t get better so I ordered a CT of her abdomen. It was normal. I suggested she try Bentyl and did a pelvic exam which was normal.
In a month she returned with similar complains and abdominal distention. She’d lost 4 pounds, I presumed she was worrying too much. I told her we had done everything, and it was an illness that would come and go. She went to another doctor who repeated the CT scan which showed a little bit of ascites now, and it turned out she had ovarian cancer. The outcome was predictably bad.
I had misdiagnosed her, which is common with this disease. My bias kept me from looking further into her complaints. I ignored the slight weight loss which is not usually seen with irritable bowel. Anchoring errors occur when we quickly fixate on a single diagnosis without consideration of other possibilities. For every 10 patients like Mrs. Feldman, 9 will have irritable bowel, but 1 will have ovarian cancer. It was a terrible lesson to learn, and I feel some shame and sadness to this day about it.Overconfidence, VIP errors, diagnosis momentum, anchoring errors, and personal biases about patients are common causes of diagnostic errors, but by no means the only ones. To those who are interested in other common errors I would urge you to read the book “How Doctors Think” by Jerome Groopman, MD, which will hopefully give you some ideas about the diagnostic errors you will or have made.
To err is human, but awareness is the first step in limiting the number of errors we make. A wise man learns from his mistakes. A wiser man learns from other people’s mistakes. If nothing else, don’t always believe what you believe. It may keep you out of trouble.
Simon Murray, MD, is an internist based in Princeton, NJ. The piece reflects his views, not necessarily those of the publication.Healthcare professionals and researchers interested in responding to this piece or contributing to MD Magazine® can reach the editorial staff here.