As many as 40,500 American adults may die in hospital intensive care units each year because their critical care teams didn't accurately diagnose their illnesses, according to a Johns Hopkins University School of Medicine review of 30 international papers that examined autopsy results.
That's more people than die each year of breast cancer in the U.S. or from bloodstream infections acquired in the ICU, the researchers say. And many more patients suffer harm from care provided for the wrong condition.
"The bottom line is that these were misdiagnoses made by the ICU staff," says Bradford Winters, MD, associate professor of anesthesiology and critical care medicine at Johns Hopkins University School of Medicine and lead author of the paper.
"We hope this article is a wake-up call so people realize the impact this has," Winters adds.
He notes that most of the thrust of the articles the researchers reviewed "lament the lost art of the autopsy," because autopsies in hospitals or in coronor's offices have been steeply declining over the years, in part because of their expense.
"The papers seem to be complaining about that, rather than recognizing that we're missing all these diagnoses, and we need to fix the system so we don't miss them. Just increasing the number of autopsies isn't going to fix the problem because these patients are already dead."
The JHU researchers say that while performing autopsies helps identify the manner and cause of death, and point to common mistakes, the real need is to develop better ways ICU providers can "measure the prevalence and impact of diagnostic errors and potential strategies to reduce them."
The paper was published online last month by the BMJ Quality & Safety.
Winters gave two examples of the most serious types of errors frequently identified:
"A patient comes in with certain symptoms and is treated for a pulmonary embolism (a clot in the lung), but an autopsy determines the patient didn't have a PE, but a heart attack," he says.
Another common error is the misdiagnosis of the fungal infection aspergillosis, which can affect patient with damaged immune systems. "The patient may come in with acute respiratory distress syndrome and is treated for a respiratory infection because no one recognizes that the patient has aspergillosis, which requires a very different treatment."
The researchers examined 31 studies that described 5,863 autopsies in about 12 countries including Brazil, France, Germany, and Slovenia. While these autopsy reports found that 28% of patients had at least one missed diagnosis, in 8% of the cases, the error was serious enough to be a cause of death.
For those studies of autopsies performed on patients who died in U.S. ICUs, between 22,600 and 40,500 adults may have died due to a missed diagnosis.
Medical conditions most commonly missed also included heart attack and pneumonia. Combined with aspergillosis and pulmonary embolism, these four conditions accounted for one-third of missed illnesses.
Winter says that correcting the problem of inaccurate diagnosis isn't easy, and may require technological improvements or upgrades with equipment that isn't yet available or approved.
Better algorithms to square conflicting test results that may point to different diagnoses – for example a troponin enzyme level could mean a heart attack or it could mean renal failure or a PE.
Likewise, better blood and imaging tests that aren't yet ready for acute care use could reduce false positives and false negatives. But there are many types of improvements that hospitals can make now that will improve diagnostic accuracy that may reduce mortality.
For starters, Winters says, hospitals should consider employing intensivists to monitor their ICUs, but only 7,000 in the U.S. are board-certified, far fewer than are needed to meet the demand.
Second is the nurse-to-patient ratio in the ICU, which more often than not should be 1 to 1 but may frequently slip to 1 to 2.
Checklists or "goal sheets," something Johns Hopkins is well known for, are extremely helpful in the ICU, but not all hospitals use them, Winters says.
"These are a list of considerations that every single patient in the ICU should have made for them every day, asking such questions as "Is the patient on a ventilator, is the patient appropriate for a sedation hold to take them off the ventilator, is the patient appropriate for deep vein thrombosis prophylaxis.
"Checklist tools help democratize knowledge and close the information gap in the ICU," Winters says.
Another source of error comes with ICU alarm fatigue, and the problem of distractions and noise.
Winters summed up saying that those who care for patients in the ICU need diagnostic help perhaps the most. Their patients are those who are most vulnerable, and are usually the least capable of participating in the diagnostic process because they are often intubated, sedated or unconscious.
"The fact that the patient can't tell you their symptoms contributes to this problem," he says.