One morning last June, my husband, Joel, sat down with the newspaper and discovered he couldn't drink his coffee without it dribbling out the corner of his mouth. He looked in the mirror and the right side of his face was drooping, like a one-sided frowning clown, and his right eye was frozen into an unblinking squint. This was especially alarming, because he's been legally blind in his left eye since birth. We drove to the doctor. In the parking lot, Joel, clearly disoriented, tried to climb out of the car before it had stopped.
As a reporter, I have been listening to Canadians' medical stories for years, from the heroic to the horrible. But this was the way I learned first-hand how faulty doctors' diagnoses can be.
Joel had seen his family doctor the week before, having woken up one day covered in round, bull's-eye-like splotches and with a shivering case of the flu. He could barely stay awake for three hours at the time. As patients do these days, we went online: He was just home from renovating our cottage in Nova Scotia, so maybe it was a cold-water rash from swimming in the ocean (a preposterous case of what doctors disparagingly call “cyberchondria”).
However, a physician friend, after hearing about an unusual “spider” bite on Joel's stomach, suggested Lyme disease – a potentially serious, relatively rare illness caused by deer-tick bites. He had seen cases before.
But Joel's family doctor had not, and he had not been convinced: He had reluctantly ordered the blood test for Lyme disease, given him some oral antibiotics and sent him home. The morning of the facial droop, his reaction was no more urgent: He said Joel had Bell's palsy, a fairly common and benign condition that usually appears for no reason and goes away on its own. We could go to the emergency room, but his attitude was that we were overreacting.
At emergency, we saw a polite, professional young doctor. We asked him about Lyme disease. No, he said. It was definitely Bell's palsy. Joel should not worry, and return to work – even though he could no longer speak clearly and could barely see.
The doctor never asked what Joel did for a living: He's a church minister.
But why had he tried to climb out of the car? That was just stress. When we pressed, the doctor reluctantly referred us to a specialist, but the appointment would be more than six weeks later.
This is where doubt sets in, as patients often find – the fear that you are making a fuss over nothing. But finally, our doctor friend told a colleague, and the infectious-disease clinic called the next day.
In the exam room, the doctor patiently listened, interrupting only to clarify the timeline. In less than 10 minutes, Joel was diagnosed with Lyme disease, and by the time we left the hospital, he'd had a spinal tap to ensure that the bacteria had not entered his nervous system, a round of intravenous antibiotics pumped directly into his heart, a prescription for injections from a home-care nurse for four weeks and a referral to a cardiologist to check his heart for infection.
Joel was lucky to be diagnosed so quickly, he was told later. (We knew we were lucky to know a doctor to advocate for us.) The longer the delay, the harder the infection is to treat, raising the risk of neurological side effects, arthritis or heart complications.
An overloaded system
How could this misdiagnosis have happened – and, what's more, persisted even as counter-evidence piled up? It not as unusual as patients – and doctors – would hope: Statistics show that 10 to 15 per cent of patients are misdiagnosed; the number is likely higher, since many medical errors are never discovered.
A 2001 study estimated that one in five mistakes occur because the system fails – a report is lost or a test is inaccurate, as in the case of nine Newfoundland women who launched a suit against the province this week, alleging that sloppy testing misdiagnosed them with severe breast cancer and that they unnecessarily underwent double mastectomies.
But that means roughly 80 per cent of mistakes are physician error. Sometimes the flu is meningitis; what presents as a kidney stone is really acute appendicitis. Even the best doctors blunder, as the recent confusion around Sidney Crosby's injury demonstrates: Was it a concussion or a broken neck or soft tissue damage, a combination of all three? It's still not clear.
Some mistaken or delayed diagnoses may be impossible to prevent when a disease is just too rare, the symptoms too generic. And luckily, the harm caused by a mistake is most often stress and worry, plus further clogging of the system as patients return for appointments.
But other mistakes can be deadly. A Canadian study in 2004 estimated that 9,000 to 23,000 Canadian patients die yearly after a preventable “adverse event,” such as surgical errors and improperly prescribed medications.
An Ontario study published last year, which looked at 22 million patients visits from 2003 to 2007, found that the longer patients waited in a crowded emergency room – only to be discharged and sent home – the more likely they were to die or be admitted to hospital within seven days. That's disconcerting when we know that more Canadians, unable to find family doctors, are using the ER as an alternative. The problem may get only worse.
Certainly, system failure is a factor: Faced with bed shortages and lengthy test delays, doctors may lower their risk threshold, sending home patients who, at slower times, would be kept for observation, speculates the study's co-author, Michael Schull, an emergency physician at the Sunnybrook Health Sciences Centre and the director of emergency medicine at the University of Toronto. Stressed-out, overworked doctors, seeing too many patients in too short a time, may be less likely to collect full patient histories or check drug interactions, or spend as much time explaining the follow-up steps to a patient. Their decision-making processes are essentially overloaded, increasing the chance that cognitive bias may cause a rash misdiagnosis.
Some Canadian ER physicians, such as Brian Goldman of Mount Sinai Hospital in Toronto, who hosts the program White Coat, Black Art on CBC Radio, have publicly challenged doctors to more openly share their errors with colleagues. (See his TED talk here.)
“The culture of medicine has an almost immunological response to error: The first instinct is to send out the antibodies and try to contain the contagion and get rid of it,” Dr. Goldman says, citing a survey of 1,800 American physicians released this month that found 20 per cent had kept mum about a mistake. Clearly they fear lawsuits, but that's not all. “Underneath is this terrible insecurity that many physicians feel, that they'll be caught out on a mistake and people will find out that they are not perfect, and somehow admitting you make mistakes is the first step in being asked to leave the profession. … We're ashamed to talk about it.”
Canadian physicians are among the best trained in the world; their error rates are on par with other industrialized countries. But, like all of us, their fallible human brains mess them up. They stumble unwittingly into false assumptions and snap judgments. “When you hear hoof beats, think horses, not zebras,” the old saying goes. But every year, thousands of patients, such as my husband, are zebras. And too often, doctors don't look closely enough to see the stripes.
The persistent rate of misdiagnosis has spurred a new conversation about how to teach medical students better, how to open up discussion about mistakes in hospitals and how patients themselves might play a role in preventing false conclusions. For instance, Dalhousie University is developing a new critical-thinking program for pre-med students. In December, Quebec became the first province to publish a standardized registry of errors, although already some hospitals have been criticized for filing incomplete reports.
“We need to train physicians to understand how decisions are made and how to think more critically,” says Pat Croskerry, a Halifax emergency-room physician and the director of medical education at Dalhousie, where he is developing the school's new curriculum. “A number of people think this is one of the next major hurdles in medicine.”
In May, 2008, Carolyn Thomas, a 58-year-old communications co-ordinator, walked into an emergency department in Victoria. She was feeling a pain down her left arm, nausea and a crushing tightness in her chest. She didn't have to wait long – it was morning and not crowded and she had worrisome symptoms – but the doctor she saw barely made eye contact, and his manner was curt and rushed. He ordered a treadmill test and an electrocardiogram. The tests came back normal.
“You fit the demographic for gastric reflux,” he told her, and she went home feeling embarrassed.
The pain returned off and on for the next two weeks. “I consider myself a smart cookie,” Ms. Thomas says. “But a guy with the letters MD after his name had told me quite clearly it was not my heart. I had no intention of going back because I didn't want to make a fuss.”
Finally, the pain was too great – she could hardly walk 10 steps. This time, in the ER, the staff reacted quickly and a problem was flagged on her test results. When the cardiologist on call examined her, he performed a quick 30-second abdominojugular test, an old-school way of checking a vein for signs of a compromised heart . His diagnosis was conclusive: She had severe heart disease that required emergency surgery.
“How come I didn't have it two weeks ago?” she asked.
The cardiologist shrugged. “Sometimes, it is hard to pick up in women.”
In fact, research shows that women similar to Ms. Thomas are up to seven times more likely to be misdiagnosed. The symptoms of a heart attack in women often appear differently than in men, for whom much of the research has been done: As she points out, they're called widowmakers, not widower-makers. But that's too simple an answer. Likely, by the time the doctor had her normal test results (not uncommon in the early stages of a heart attack), he had already sized her up and committed several cognitive errors.
Most significantly, as in my husband's case, he quickly “anchored,” as psychologists call it, on the simplest, most common diagnosis. He had fallen into the trap of confirmation bias, mentally discarding the symptoms that didn't fit and piling up the bricks to support a false conclusion: In the exam room, Ms. Thomas would have presented as a trim and healthy professional with a large, Type-A personality and deadlines on her mind – the “demographic” of patient who might have stress-induced heartburn. Her test scores bolstered that first impression.
The doctor also may have considered, even unconsciously, the statistical likelihood of a heart attack, and his experience in treating previous cases. Perhaps he was tired after a long shift, or the hospital was full.
Psychologists have identified about 100 different cognitive traps people generally fall into – shortcuts to quick decisions that, once upon a time, probably saved us from being eaten. In a medical context, for example, doctors are susceptible to overconfidence bias (relying too much on intuition), diagnosis momentum (accepting a previous doctor's findings without enough skepticism) and availability bias (concluding that a new patient has the same problem as a recent patient with similar symptoms). Those mental miscues are even stronger when the correct diagnosis is a rare one.
Doctors also aren't immune to stereotypes. They may judge a patient on appearance, gender or race. They may be misled because someone has come in smelling like alcohol or is snappish in the examining room.
Dr. Croskerry poses this question: If doctors understood the psychology behind those errors, would they more on guard to watch for them? And if mistakes were put within the framework of cognitive science, as opposed to competence, would physicians be more willing to discuss them?
“You interact with people and those interactions colour your judgment,” he says. “What we would like doctors to do is detach themselves from the immediate pull of the situation.”
He gives this example: A sprightly 65-year-old woman came into a Halifax emergency department last summer. She was complaining about a hurt shoulder, saying she had strained it pushing the lawnmower. An X-ray was ordered, and she was sent home to rest. Days later, she returned, this time having a heart attack. “People had focused on her interpretations,” Dr. Croskerry says. They failed to consider the alternatives.
That woman lived, but some mistakes are discovered only in the morgue: A few years ago, a 28-year-old woman arrived in a Halifax ER from an addiction centre complaining of chest pain. The ER staff were instantly on guard – they had plenty of experience with addicts feigning injuries, hoping to be prescribed drugs. While she was there, the women stepped outside for a smoke. The doctor on duty felt irritated that she was wasting his time: He diagnosed anxiety and discharged her. That night, she died from massive blood clots in her lungs.
Without a full medical history – and an open mind – the doctor never considered the risk of clotting caused by her smoking habit and birth-control pills.
Jerome Groopman, the author of How Doctors Think, explains that 80 per cent of most diagnoses can be made strictly from a patient's medical history and description of symptoms. And yet, he says, doctors typically interrupt a patient within 18 seconds – they are already floating several possible diagnoses within a few minutes of the encounter.
They have become increasingly reliant on technology to make the call (which assumes, of course, that patients get the right tests) and are less likely than in the past to physically examine patients.
In ERs especially, where doctors don't know patients personally and are often sleep-deprived and incredibly rushed, there is little time to take long patient histories. Patients themselves, after stewing in the waiting room for hours, also may be impatient about giving them.
It can be even harder for doctors to overcome confirmation bias when the accurate diagnosis is more unusual. Last February, Tara Jacklin went to her family doctor in St. Catharines, Ont., complaining of bloating, constipation and a lack of appetite.
An emergency nurse and a long-distance runner, she was, at 24, the picture of health. She was worried, however, and asked for an abdominal ultrasound to check for ovarian cancer. Her doctor told her that it wasn't necessary – she speculates he might have seen her as a rookie nurse with worst-case scenarios on her mind. She had no family history of ovarian cancer. The doctor diagnosed irritable bowel syndrome and told her to eat more fibre.
Ms. Jacklin convinced herself that he was right, though her symptoms persisted. When she went back a second and third time, the doctor maintained his diagnosis. “I felt stupid, like I was overreacting,” she says.
Finally, she visited a walk-in clinic for a second opinion: Was it unreasonable, she wanted to know, to insist on an ultrasound? That doctor directed her to emergency, where tests revealed a tumour in an ovary. She was eventually diagnosed with aggressive but early-stage ovarian cancer.
“Statistically, it didn't make sense to look into anything serious,” says Ms. Jacklin, who is now receiving chemotherapy. “I know that every time someone has an ache and pain, you can't just take 12 test tubes of blood.”
What she regrets, though, is not pressing her doctor harder to consider alternatives.
She will take that lesson back to work: “I will really give credit to the patient when they come in,” she says. “I will advocate for them.”
In Joel's case, once he was finally diagnosed, everything else fell into place and his care was excellent. He was a subject of study for residents, whose minds may some day pull up his unusual case should they ever see a patient with a drooping mouth and a strange insect bite.
But the young ER doctor who told Joel to go back to work, or the older physician who sent Ms. Thomas home to take Tums may never have discovered their errors. That's the truly tragic medical blunder, as Dr. Goldman and Dr. Croskerry would say – the failure to learn from a mistake, so it doesn't happen again.
HOW TO IMPROVE YOUR OWN DIAGNOSIS
The good news is that most of the time doctors get their diagnosis right: Your sniffles really are just a nasal infection. But about one-fifth of the time, something is missed or conclusions are reached too quickly. Here's how it can happen, and how a patient can help prevent it.
Your doctor steps into the room, and already she has made a preliminary scan: your gender, your weight, your demeanour. She may have glanced at the chart with the nurse's note and be considering two or three diagnoses, suggests Harvard medical professor Jerome Groopman, the author of How Doctors Think. It takes doctors, on average, 18 seconds before they interrupt a patient, which means facts in a patient's history are often missed.
Be as specific as you can about your symptoms. Write the details down if you need to keep them straight. Make sure you know your medications and the amounts you are taking, so the doctor doesn't have to spend time figuring it out based on your “little white pill” description. And definitely don't complain about how long you have been waiting – this just wastes your appointment time.
Doctors' minds, full of medical facts, can fall prey to several common cognitive traps, assuming, for instance, that the fit guy in running clothes is too healthy to have a heart attack. They may glance at a previous doctor's finding and accept it without really questioning it – diagnosis momentum. Or they may be waylaid by “premature closing” – stopping searching for another answer too early, or missing details that would dispute their conclusion.
Dr. Groopman says to ask questions that prompt your doctor to consider alternatives: What else could it be? Could two things be happening at once? Is there anything in my exam or test results that may contradict your working diagnosis?
Sendoff and follow-up
Your doctor has given you his diagnosis; her mind is already drifting to the lineup of other waiting patients. Make sure you understand clearly what warning signs to watch for after leaving the doctor's care, and what you should do if your symptoms don't improve.
EXPERIMENT: THE SUCCESS OF FAILURE
The argument that discussing medical mistakes more openly would make doctors better at saving patients found some scientific support in an interesting Toronto study published last fall.
In a computer-simulated experiment, 35 doctors were presented with a series of 64 patients having a “heart attack” in an emergency room. The doctors were given a short, six-point history for each patient, and two fictional drug treatments. To know which drug worked best for which patients, they would have to experiment; they had to decide which drug to use within 10 seconds. Based on their choice, the computer would flash “success” or “failure” to say which patients survived.
As a group, they didn't do very well. Only nine of the 35 doctors figured out the proper treatment pattern: Drug A worked better for patients with diabetes, and less well for those without; Drug B was the opposite.
Using fMRI scans, the researchers found that the physicians who failed the test showed more activity in their frontal lobe when the treatment worked; the small group of physicians who passed, however, had busier frontal lobes when their treatment failed.
In other words, says Jonathan Downar, a psychiatrist at the University of Toronto and lead author of the study, the doctors who learned more quickly were the ones who paid the most attention to the times when they messed up.
By focusing on their failures, the doctors had avoided a confirmation bias: They had not assumed that because they danced and then it rained, that their dance had caused the rain. Through a process of “disconfirmation,” Dr. Downar explains, “the only rules you are left with are the ones that stand the test of time.”
Being right may be just a distraction. For diligent fact-finding, in medical settings and everywhere else, the question to ask should be, “How can I prove myself wrong?”
Erin Anderssen is a Globe and Mail feature writer.