It is not possible to live life in a way that every choice and decision awaits a definitive, double–blind study with a statistically significant population of subjects to determine the best way to proceed. As a matter of fact, very few of our problems are resolved that way. We often reach conclusions emotionally and then invoke logic and reasoning to justify them. Instead of approaching problems by stating a hypothesis and then testing it against evidence, we often unconsciously begin with a conclusion or belief and then seek evidence to confirm it. In medicine, as elsewhere, this is known as confirmation bias. This is a dangerous intellectual trap for a physician to fall into.
Headaches are one of the most prevalent symptoms presented in clinical practice. The migraine headache is common, and is often the doctor’s first consideration before she analyzes other potential causes of the headache.
This is called pattern recognition. The doctor has seen this symptom countless times and is quick to recognize it. If this diagnosis is correct, it saves time, eliminates a lot of unnecessary testing and additional patient anxiety, and measures can be quickly taken to safely alleviate the pain.
What if it’s not a migraine? What else could it be?
A middle–aged woman we’ll call Patty who lives in a small town came to the emergency room of the hospital complaining of a headache. She was well known, and allegedly was quite a drinker. The medical providers made the assumption that she was simply hung over, but they didn’t bother to check the alcohol level in her blood. She did not present with any of the usual medical stigmata of an alcoholic, such as liver or blood disease. They didn’t do computed tomography, otherwise know as a CT scan, to look for other possibilities. They used the heuristic called attribution error: they labeled the patient socially and then used that as a substitute for a scientific diagnosis. Therefore, there was no need to doubt themselves or perhaps ask, “What else could this be?”
It is possible the doctor also used the availability heuristic: he went with the first diagnosis that came to mind. If he has treated a number of patients recently that were under the influence of alcohol, this is the first thing that would come to his mind.
In this case, they sent the woman home. She came back the next day with the same complaint of headache, but more intense. Finally, they gave her a CT scan and discovered that she had a ruptured cerebral aneurysm.
A competent woman in her early sixties who we will call Wendy, gainfully employed as a manager at an airport, presented at a small suburban hospital with a severe headache and flu symptoms. She mentioned that during the previous forty–eight hours she had experienced diarrhea and stomach ache, both of which she attributed to a restaurant where she had eaten. Significantly, her blood pressure was 220/130, which meant she was a walking time bomb. She told the attending physician that she had had high blood pressure for a long time, but that fifteen years previously, she had decided to stop taking medication for it. The only doctor she had been seeing was an allergist, and she had told him about her high blood pressure, but he had not referred her to another physician (as he should have).
Despite her extremely elevated blood pressure and headache, two symptoms often related to each other, Wendy did not have symptoms of a brain hemorrhage such as a stiff neck or sensitivity to light. What could this be other than a case of high blood pressure?
Incredibly, the attending physician gave her two milligrams of morphine for her pain and a brief lecture about the dangers of her high blood pressure, and sent her home. Did the doctor commit the heuristic of anchoring, which means he made an initial judgment that this was an extreme but uncomplicated case of high blood pressure? If so, then he achieved his immediate goal of getting the patient rotated out of his emergency room and referred to a primary care physician who would doubtless try to persuade Wendy to get back on hypertension medication. He went with the obvious, and stuck with his gut reaction, which he threw down like an anchor.
He obviously did not ask, “What is the worst thing this could possibly be?”
Wendy dutifully made an appointment with her primary care physician. The first opening he had was more than a week later. In the meantime, over the next five days, Wendy got sicker and sicker and started becoming mentally confused. She went back to the emergency room where they did a CT scan, which revealed a brain hemorrhage from a ruptured aneurysm.
A week after her first visit to the emergency room, a condition known as vasospasm, meaning a spasm of a blood vessel in her brain, had set in, leading to oxygen deprivation in that part of her brain. Her aneurysm, which a week before could have been surgically clipped, thereby removing the threatening health issue, was now impossible because the vasospasm was so severe that the aneurysm was untreatable.
Today, Wendy lies on her back at home in a vegetative state, unable to walk, talk, or otherwise communicate, with a feeding tube in her stomach. She can’t follow directions, and she suffers occasional seizures. She requires constant care.
The medical practitioners in Wendy’s case forgot Rule Number 1 of emergency room diagnostics: rule out worst–case scenarios. Don’t send a patient home with a condition that may kill them before they get there. For every major symptom that a patient may present with, most doctors can think of about half a dozen of the most dangerous, possibly life–threatening causes. Their first task is to eliminate them or treat them, but never to miss them. If someone comes in the door with chest pain, make sure they’re not experiencing a heart attack right in front of your eyes. If they present with a headache, make sure they’re not experiencing a cerebral hemorrhage or a stroke, and so on.
Lawrence Schlachter is a neurosurgeon and author of Malpractice: A Neurosurgeon Reveals How Our Health-Care System Puts Patients at Risk.
Image credit: Shutterstock.com