Diagnosis-specific estimates also suffer from similar issues, particularly with respect to the incidence of stroke in patients presenting to the ED with dizziness. The authors of the AHRQ review, in repeated and prominent calls for future research in their personal area of academic work, cite an “estimated 45,000 to 75,000” missed strokes in dizziness annually. The source for this statistic in the AHRQ review is, in fact, an editorial by these same authors, which subsequently cites their own practice seminar article regarding the HINTS (Head Impulse, Nystagmus, Test of Skew) exam, which ultimately extrapolates data from a study of patients with dizziness.¹²

This retrospective study from Neuces County, Texas, was conducted in 2000-2003, and focuses on 53 patients who were adjudicated to have had a cerebrovascular diagnosis following chart review. The AHRQ authors’ analyses of this study neglect to mention a third of these patients were not diagnosed with stroke, but with transient ischemic attack, and it is the mere 46 ED cases from this cohort forming the foundation for the proposed rate of missed strokes in modern clinical evaluation. It is absolutely the case patients with dizziness can manifest underappreciated etiologies, but studies of over 40,000 patients provide estimates of subsequent stroke diagnoses of 0.18 percent within 30 days, a far cry from tens of thousands of missed strokes.¹³

Finally, the accounting of the frequency of various clinical conditions affected by diagnostic error is derived substantially from a U.S. database of closed malpractice claims. While there is certainly alluring face validity to serious harms percolating to the level of a tort claim, these data cannot realistically inform any sort of reliable estimate of relative disease-specific errors. Likewise, using these tort data to approximate estimates of the frequency of types of diagnostic error is likewise invalid. The authors of the AHRQ admit as much in the text, but do not refrain from heavily utilizing this citation.

Unsupported and Misleading

The field of diagnostic error, patient safety, and cognitive biases in medicine is of profound importance to the specialty of emergency medicine. These issues of diagnostic accuracy must also be considered within the challenges of resource stewardship, overdiagnosis, and unintended consequences. The findings promulgated by this AHRQ review are, bluntly, unsupported by the evidence cited and misleading as to the gaps requiring further study. In light of the comprehensive issues marring this publication, I personally believe it should be retracted for further revision.