番茄社区

An estimated 1.5% of diagnoses at four pediatric intensive care units (PICUs) were wrong, a new study found, and that's good news because it could have been quite a bit worse.

In the retrospective study of nearly 900 randomly selected patients treated at the PICUs over a 1-year period, 13 had a diagnostic error within a week of their admission, according to critical care specialist Christina Cifra, MD, MS, of the University of Iowa in Iowa City.

Other PICU studies have estimated error rates of 8% to 25%, said Cifra, who presented the new findings this week at the Society of Critical Care Medicine's virtual Critical Care Congress.

"The diagnostic error prevalence is quite low, which is great! But we need to view this rate as an underestimate," she told 番茄社区, adding that even a 1.5% error rate "can translate into hundreds of critically ill children having diagnostic errors per year."

Diagnostic errors are understudied compared to other kinds of medical mistakes, such as medication errors and hospital-acquired infections, Cifra said. For one, "prior research overemphasized the cognitive roots of misdiagnosis as opposed to systems issues. Only recently have we realized that we can apply a systems-based perspective similar to that used to study other safety events."

Also, physicians are reluctant to discuss errors, she said, and "we lack an epidemiologic gold standard for measurement," and mistakes "may not be identified until long after the error occurred."

The researchers launched their retrospective cohort study in four academic tertiary-referral PICUs and analyzed diagnoses within 7 days of admission among 882 patients (age 18 or younger) who were not electively admitted. Researchers used a diagnostic tool called Safer Dx to pinpoint possible misdiagnoses, and investigator teams decided if there were actual errors.

The group relied primarily on chart documentation, "which can be inaccurate or incomplete, to determine error," Cifra cautioned. "And we used a rigorous consensus process favoring a conservative designation of diagnostic error."

In one of the cases of diagnostic error, a patient required a longer hospital stay as a result. Respiratory and infectious conditions (each 31%) accounted for most missed diagnoses.

In 46% of cases, clinicians missed a diagnosis that should have been obvious due to a "suggestive history." For example, Cifra said, clinicians may fail to re-evaluate diagnoses from previous clinical settings such as the emergency department.

In addition, there was a "failure to broaden diagnostic testing given known clinical information" in 46% of the cases.

"One good example of this, which commonly occurs in the PICU, would be patients that are frequently admitted for respiratory infections or have a history of multiple recurrent respiratory infections," Cifra said. "For these patients, it would be prudent to look beyond their current PICU admission and review their clinical course more broadly -- i.e., across multiple hospital encounters -- to be able to identify potential underlying pathology that is causing the recurrent problem, such as anatomical abnormalities in the airway or a heretofore undiagnosed immunodeficiency."

"If the PICU physician/team do not broaden their view, then important underlying etiologies of disease may be missed," she added.

The analysis suggested that two factors greatly boosted the risk of an error: atypical presentation (OR 4.63, 95% CI 1.03-20.79) and diagnostic uncertainty on admission (OR 8.72, 95% CI 2.25-33.73). The researchers also found that intensivists over the age of 45 made more errors than younger physicians, but Cifra said this connection vanished after adjustment for confounders.

Asumthia S. Jeyapalan, DO, MHA, a pediatric critical care physician at the University of Miami, told 番茄社区 that diagnostic errors in PICUs are especially hard to analyze.

"In pediatric critical care medicine, it has been especially difficult to define diagnostic errors as often our patients are unable to communicate their signs and symptoms to us," said Jeyapalan, who was not involved in the study.

The diagnosis of neurological conditions can be especially challenging, she said.

"There are different presentations for seizures, meningitis, strokes at different neurodevelopmental ages. Autism spectrum and other diseases can make diagnosis in children with neurological disease difficult," said Jeyapalan.

In an unadjusted analysis, the study found that neurologic chief complaints were more common in patients who were deemed to be misdiagnosed than in those who were believed to be diagnosed correctly (46.2% vs 18.8%, P=0.024).

Going forward, Cifra said, PICU clinicians should "be aware of the characteristics that place their patients at high risk for diagnostic error," such as unusual presentations. "In addition, if clinicians have significant diagnostic uncertainty, they should not hesitate to ask a colleague or two to evaluate the data independently and provide their own opinion early on."

However, she added, "individual solutions like this will only take us so far. We need to establish systems that can help prevent diagnostic error. Examples include automatic screening for patients at high risk of diagnostic error. Perhaps these patients need an early care conference, or an automatic 'second look' trigger, or standard quantification and communication of diagnostic uncertainty."

Comment