It has been well over a year since the controversial publication of the Agency for Healthcare Research and Quality (AHRQ) report on diagnostic errors in the emergency department (ED).¹ The striking conclusions of this report included estimates that 0.1 to 0.4 percent of ED visits resulted in preventable death as result of diagnostic error. While these percentages appear small, when multiplied by the 130 million ED visits annually, these authors arrive at an estimate of 250,000 preventable deaths. Adding in errors and preventable deaths in those occurring following admission to the hospital, these same authors now propose up to 371,000 deaths annually in the United States as a direct consequence of diagnostic error.² This accounts for nearly 15 percent of all deaths annually in the United States alone.

Putting aside, for a moment, the methods leading to concerns regarding the magnitude of these estimates, the fact remains diagnostic errors occur with under-appreciated frequency. The better questions remain: Where in the process of care do the harmful errors occur? If these can be characterized, what preventive interventions might be considered?

Taking a different approach than the authors of the AHRQ report, Auerbach et al used a “look back” approach to perform both qualitative and quantitative evaluations of the types and frequencies of errors occurring in hospital settings.³ In their study, patients were identified retrospectively as having suffered deterioration following admission resulting in transfer to the intensive care unit, death, or both. From their initial search, the authors found 2,428 patients meeting criteria, for each of whom they performed manual chart review. Of these, the authors identified 550 (23.0 percent) patients had experienced a diagnostic error. As a “look back method,” these are not estimates for the overall prevalence of error but do provide substantial sample size for their analyses of subtypes of diagnostic error.

By far, the most common subtypes of diagnostic error were those associated with patient assessment and obtaining appropriate testing. The bulk of patient assessment errors fell into the category of failure or delay to consider the diagnosis. The included example describes a critically ill patient admitted with hepatitis and hematochezia whose acidosis failed to improve with initial inpatient management. After developing encephalopathy and hypoxemic respiratory failure, the patient was transferred to the ICU. Further diagnostic testing in the ICU identified salicylate toxicity. In this case, the harm resulted directly from the initial failure to consider the correct diagnosis. It ought further to be noted, as with many cases of diagnostic error, several other types of error were identified as contributing to the primary error causing harm.

The second major category involves failure or delays in obtaining the appropriate testing. An included example of a failure to order proper testing describes a case of a chronically anticoagulated patient admitted for a psoas hematoma following bone marrow biopsy. After a period of inpatient observation, anticoagulation was restarted. Subsequently, the patient developed increasing extremity pain and tachycardia. A CT angiogram was not ordered until the following morning, nor performed for several additional hours. Active extravasation of blood into the hematoma was identified, and the patient was referred to interventional radiology. The diagnostic error was therefore classified as relating to the delays associated with testing and its effect on subsequent definitive management.

The remaining scope of diagnostic errors includes a veritable cornucopia of illustrative morsels. An example of inadequate physical examination interpretation is provided by failing to elicit reflexes in a patient with weakness, leading to a delayed diagnosis of Guillain-Barré. An example of an error in history-taking is illustrated by a patient with fecal impaction and kidney injury whose excessive ingestion of magnesium citrate was missed, delaying identification of serum magnesium level of 10.2 mg/dL. A patient was admitted with abdominal pain following hysteroscopy, and the free air seen on CT was thought to be related to the procedure, rather than the ultimate diagnosis of small bowel perforation. Other examples included patients transferred between services whose failure to include complete handover information led to other delays in care.

Each of these examples and their classifications into Diagnostic Error Evaluation and Research frameworks helps safety researchers develop strategies to improve processes systematically contributing to diagnostic errors. Understanding of the domains of diagnostic errors allows for further exploration of the foundational causes of subtypes of error, with the ultimate hope of identifying acceptable interventions to mitigate such deficiencies.

Circling back to the estimates of diagnostic errors causing severe harm to hundreds of thousands in the United States annually, it remains reasonable to recognize these estimates are built on precarious scaffolding and extrapolation from the authors’ own prior work. The issues and flaws in their methods have been competently dissected elsewhere.^4,5 Rather than rehash the accuracy of these estimates, however, an alternative thought experiment involves taking these numbers at face value. This includes such estimates as the rate of diagnostic error for diseases such as “aortic aneurysm and dissection” between 21.0 and 51.7 percent. These rates of error are, at the least, consistent with the colloquial “standard of care” for dissection which has been to “miss the diagnosis” initially. Similarly elevated rates of error, and related harms, are associated with other serious vascular, infectious disease, and cancer-related diagnoses, as well.

Therefore, it may be reasonable to consider the profuse rate of diagnostic error comprising the current state of clinical practice directly informs the legal definition of the “standard of care.” It is regularly noted the “standard of care” is not “perfect” care, but a much lower standard.⁶ Three cases, Hall v. Hilburn, Mc-Court v. Abernathy, and Johnston v. St. Francis Medical Center are cited as forming the general basis for the modern definition of “standard of care.” Relevant portions include:

“Our law says that a physician is not an insurer of health, and a physician is not required to guarantee results. He undertakes only to meet the standard of skill possessed generally by others practicing in his field under similar circumstances.” (Abernathy)

and

“When a physician undertakes to treat a patient, he takes on an obligation enforceable at law to use minimally sound medical judgment and render minimally competent care in the course of the services he provides.” (Hilbun)

Taken together, these opinions reinforce a standard associated with the skill of a “minimally competent” clinician. Relying upon these published estimates of the rate of diagnostic error, it would follow that error-prone practice is an unfortunate reality in modern medicine. A typical, reasonable, clinician providing “minimally competent care” will routinely make mistakes, and it is solely by virtue of good fortune and context by which the errors do not result in serious harms.

Patient safety and diagnostic errors research is critical to the design of a health system in which errors are minimized. With the advent of further decision-support and artificial intelligence-augmented medical care, potent tools exist to address the systemic barriers and cognitive biases that result in error. However, striving for ideal care is not incompatible with recognizing the challenges associated with diagnosis in our present complex medical ecosystem. These differing approaches to the problem of diagnostic error ought to both help us move forward, while also aiding a conversation regarding legally protecting physicians from the expectation of perfection.

Dr. Radecki (@emlitofnote) i s an emergency physician and informatician with Christchurch Hospital in Christchurch, New Zealand. He is the Annals of Emergency Medicine podcast co-host and Journal Club editor.

References

1. Newman-Toker DE, Peterson SM, Badihian S, et al. Diagnostic errors in the emergency department: a systematic review [internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2022 Dec. (Comparative Effectiveness Review, No. 258.) Available from: https://www.ncbi.nlm.nih.gov/books/NBK588118/
2. Newman-Toker DE, Nassery N, Schaffer AC, et al. Burden of serious harms from diagnostic error in the USA. BMJ Qual Saf. 2024;33(2):109-120. Published 2024 Jan 19.
3. Auerbach AD, Lee TM, Hubbard CC, et al. Diagnostic errors in hospitalized adults who died or were transferred to intensive care. JAMA Intern Med. 2024;184(2):164-173.
4. Kelen GD, Kaji AH; Consortium of societies of emergency medicine. the AHRQ report on diagnostic errors in the emergency department: the wrong answer to the wrong question. Ann Emerg Med. 2023;82(3):336-340.
5. Carpenter CR, Jotte R, Griffey RT, Schwarz E. A critical appraisal of AHRQ’s “diagnostic errors” report. Mo Med. 2023 Mar-Apr;120(2):114-120.
6. Moffett P, Moore G. The standard of care: legal history and definitions: the bad and good news. West J Emerg Med. 2011 Feb;12(1):109-12. PMID: 21691483.