Diagnostic errors — that is, overlooking a disease, or diagnosing it erroneously or late — are a known occurrence in health care, and one that can have tragic effects. Yet so far there have been few attempts to quantify misdiagnoses in terms of deaths and permanent disability, and the full scale of their impact has been underestimated by the medical community, according to a 2015 report by the National Academy of Medicine.
A new study, published this week in BMJ, finally addresses that gap. An estimated 371,000 people die every year following a misdiagnosis, and 424,000 are permanently disabled — a total of 800,000 people suffering “serious harm,” said David Newman-Toker, the lead author of the paper and a professor of neurology at Johns Hopkins School of Medicine and director of its Center for Diagnostic Excellence. Settling on an exact number is hard because many cases of diagnoses go undetected, he said. It could be fewer than his study identified, or more — between half a million and a million — though in any event it would be the most common cause of death or disability due to medical malpractice.
He likens the issue of misdiagnosis to an iceberg, saying cases leading to death and disability are but a small fraction of the problem. “We focused here on the serious harms, but the number of diagnostic errors that happen out there in the U.S. each year is probably somewhere on the order of magnitude of 50 to 100 million,” he said. “If you actually look, you see it’s happening all the time.”
But misdiagnoses typically don’t lead to severe consequences, because most times people aren’t visiting the doctor with a serious condition. “The risk level just walking through the door in the doctor’s office that something horrible is going to happen to you because of a diagnostic error is actually quite low,” said Newman-Toker.
That’s different, though, for the minority of people who walk into the doctor’s office with serious, potentially fatal, conditions — say, someone who feels dizzy because of the onset of a stroke, but is diagnosed with vertigo instead, or a young woman whose breast cancer is missed during a routine check-up. For them, the study found, the risk of death associated with misdiagnosis is 4%, and the risk of severe disability 11%.
Solutions within reach
Analyzing the nature of misdiagnoses also provides significant opportunities for solutions: the errors are many, but they are quite concentrated. According to the study, 15 diseases account for about half the misdiagnoses, and five diseases alone — stroke, sepsis, pneumonia, venous thromboembolism, and lung cancer — caused 300,000 serious harms, or almost 40% of the total, because clinicians failed to identify them in patients. “That’s a lot that you could accomplish if you cut those harms by 50% for just those five diseases — that would be 150,000 prevented serious permanent disabilities or death,” said Newman-Toker.
The overwhelming reason for misdiagnosis is cognitive error on part of the doctor, he said. Doctors may make poor diagnostic decisions because the signs of the disease aren’t typical, or because the symptoms are too common and not frequently associated with a severe disease: Dizziness, for instance, occurs in a majority of cases of stroke, but only very few cases of dizziness result in stroke.
Or, it may be an issue of bias, either circumstantial — for instance, overlooking the possibility that a young patient may be having a stroke, because it’s exceedingly rare — or systemic, resulting for instance in dismissal of symptoms reported by Black patients. (Newman-Toker says his research estimates that being a woman or a person of color increases chances of misdiagnoses by 20% to 30%.)
This means that finding a way to avoid such cognitive errors, at least in the most commonly misdiagnosed severe diseases, would save a large number of lives. A quick way to do it? Routinely incorporating a second opinion in visits, suggests Newman-Toker. “I believe that the quickest way to solve the diagnostic error problem in the real world would be to construct approaches that basically rely on the ‘phone a friend’ model,” he said. This doesn’t mean that the patient should have to seek a second opinion, but rather that providers should make it standard practice to consult with a colleague before providing a diagnosis or dismissing a patient.
The economic burden
The study findings also provide insight into the economic burden of misdiagnoses, something that has so far been hard to quantify due to the lack of a reliable estimate of occurrences.
To explain the huge cost of misdiagnoses, William Padula, an assistant professor of pharmaceutical and health economics at the University of Southern California, who did not participate in the study, also uses a missed stroke example — it is the most commonly misdiagnosed severe condition, so much so that AI is being tested as a solution.
“A patient comes into the ED with a headache or dizziness and they get told it’ll go away, and then they go home. And then a week later, you find out that they have a stroke,” he said. ”By then, the stroke has compounded so much that what could have been addressed in the moment […] for $10,000 now becomes a $100,000 issue. […] So there’s a margin of $90,000 that has been added to the U.S. health system burden because of the misdiagnosis.”
Multiply that for hundreds of thousands of misdiagnoses and the total costs are exorbitant, on the order of $100 billion, and likely untold billions of dollars more, said Padula, especially considering societal costs such as years of lost productivity due to premature death or disability. “On the one hand we could be saving a hundred or more billion dollars by accurately diagnosing. And then on the back end, if we get this problem right, then instead of a $22 trillion GDP, we might have a $24 trillion GDP just because everybody’s getting the care that they need,” he said.
This helps put into perspective the costs that would be incurred to get a second opinion, or to add extra tests to rule out certain conditions. “It’s a cost-effective solution upfront to spend the extra money to more accurately diagnose patients,” said Padula.
Yet, the funding for research to improve diagnoses is extremely limited — proportional to, Newman-Toker said, the insufficient political (and medical) importance given to the issue.
“The amount of federal research dollars support that’s going to address this problem is on the order of magnitude of $20 to $30 million a year,” he said. Before 2016 the federal funding was even lower, at $7 million a year, he said, but it was raised as more data came out on the impact of misdiagnoses. That amounts to about $50 per death — a pittance, he said, considering the federal investment made in researching treatment for some serious diseases (such as certain types of cancer) can run as high as $400,000 per every death recorded from the disease.
By way of comparison, Newman-Toker noted the $39 million going toward smallpox research. “I’m sure that that’s important research, but smallpox was eradicated 50 years ago and causes zero deaths,” he said.