The care given in emergency departments came under fire yesterday with the release of a government study saying that 250,000 Americans die every year due to misdiagnoses.
The findings spurred an immediate response from the president of the American College of Emergency Physicians (ACEP), who questions the study’s veracity and methodology.
Christopher S. Kang, M.D., president of the ACEP, said in a statement that “in addition to making misleading, incomplete and erroneous conclusions from the literature reviewed, the report conveys a tone that inaccurately characterizes and unnecessarily disparages the practice of emergency medicine in the United States.”
One of the authors of the study—David Newman-Toker, M.D., Ph.D., a neurologist at Johns Hopkins University—vehemently defends the methodology and told Fierce Healthcare that “high levels of variation in care (across conditions, across hospitals, across demographic groups) tells us that these errors do not need to be thought of as ‘the price of doing business.’ It tells us that there are already probably ways to get it right.”
The study, conducted by Johns Hopkins for the Agency for Healthcare Research and Quality, states that among 130 million ED visits in the U.S. per year, 7.4 million patients are misdiagnosed. In addition, 2.6 million suffer an adverse event, and about 370,000 suffer serious harm from diagnostic errors.
The study states that “put in terms of an average ED with 25,000 visits annually and average diagnostic performance, each year this would be over 1,400 diagnostic errors, 500 diagnostic adverse events, and 75 serious harms, including 50 deaths per ED.”
The top five conditions associated with misdiagnosis are stroke, myocardial infarction, aneurysm and dissection, spinal cord compression and injury and venous thromboembolism, according to the study. Three conditions are tied for sixth and seventh place: venous thromboembolism, meningitis and encephalitis, and sepsis. Then comes lung cancer, traumatic brain injury and traumatic intracranial hemorrhage, arterial thromboembolism, spinal and intracranial abscess, cardiac arrhythmia, pneumonia, gastrointestinal perforation and rupture, and intestinal obstruction.
Researchers looked at studies of ED performance over two decades, but much of the data were collected from EDs in Canada and Europe. Kang said that “while most medical specialties have similar training in Western nations, emergency medicine does not.”
Newman-Toker counters that the researchers obtained input from “key informants,”—that is public health officials, healthcare providers, business leaders, social service providers, officials with faith-based organizations and other community leaders as well as from an expert panel that included ED physicians.
“They told us what countries were similar enough to the U.S. ED system of care and what ones were not,” says Newman-Toker. “We did exactly what they recommended that we do—we restricted ourselves to those countries (United States, Canada, United Kingdom, Western Europe, Australia, New Zealand). In the report, we describe at some length why we believe the results (some of which derive from non-U.S. studies) are applicable to the U.S."
Robert Wachter, M.D., chairman of medicine at the University of California, San Francisco, tells The New York Times that “diagnostic errors are a huge part of the problem.” He adds: “This is a really complicated calibration problem. The answer can’t be let’s test everybody for all this stuff all the time.”
Newman-Toker agrees saying that “in many cases, I believe we are both over-testing and undertesting at the same time. The goal is to get better at bedside diagnosis so that we do a better job of selecting the right tests for the right patients at the right time—if we do that, I believe we will improve diagnostic accuracy and simultaneously decrease unnecessary and inappropriate testing.”
Kang said that “as with all medical specialties, there is room for improvement in the diagnostic accuracy of emergency care. All of us who practice emergency medicine are committed to improving care and reducing diagnostic error.”
Q&A With David E. Newman-Toker, M.D.
Fierce Healthcare: It appears as if the data were collected before the COVID-19 pandemic. Do you think the pandemic changed the dynamic in EDs in some way? What are your concerns about the tripledemic and diagnoses in EDs?
David Newman-Toker: Yes, most of the studies were published before the pandemic and there were relatively few data on the impact of the pandemic on diagnostic accuracy. I think there is no question that the pandemic adversely impacted diagnosis in the ED—first and foremost by causing patients to stay away from the ED, out of fear and secondarily by making care in the ED impossibly difficult in many instances. When the ED is overwhelmed by any epidemic (pan-, triple-, etc.) or chronic, endemic problems with overcrowding and boarding, misdiagnosis of subtler cases is bound to increase.
FH: How do you think EDs can do better with diagnoses without a major influx of new funds? In other words, if someone in charge of an ED read the study and decided today that they want to make improvements, how would they start?
DNT: Pick one problem that you know is harming patients at your institution (from malpractice claims, incident reports, analysis of unplanned revisits, etc.), measure the approximate frequency of adverse events and implement a new pathway of care for the most commonly misdiagnosed clinical presentation of that disease. If you don’t have your own data on errors/harms or solutions, pick the one that our team at Hopkins has identified as a major problem most places—No. 1 harmful misdiagnosis is stroke; No. 1 risk factor for missed stroke is presenting with dizziness (as opposed to more “classic” stroke symptoms).
FH: Is there anything you’d like providers who work in EDs to know about the study and what you hope might result from the findings?
DNT: Just five conditions (No. 1 stroke, No. 2 myocardial infarction, No. 3 aortic aneurysm/dissection, No. 4 spinal cord compression/injury, No. 5 venous thromboembolism) account for 39% of serious misdiagnosis-related harms, and the top 15 conditions account for 68%. This suggests the problem is more tractable than previously imagined. If we focus our efforts on fixing problems that are most likely to cause serious harms when missed and the most common situations for us to miss them, we might be able to cut the problem in half fairly quickly.
This interview has been edited for clarity and length.