Editor's Note
January 8, 2024
Protecting Patients by Reducing Diagnostic Error
Grace Y. Zhang, Cary P. Gross
JAMA Intern Med. Published online January 8, 2024. doi:10.1001/jamainternmed.2023.7334
Almost a decade after the release of “Improving Diagnosis in Health Care,” the National Academies of Sciences, Engineering, and Medicine report that highlighted the imperative to improve the diagnostic process in health care, diagnostic errors continue to be a cause of patient harm and death.1
In JAMA Internal Medicine, Auerbach et al2 investigated diagnostic errors among hospitalized patients who had experienced a major clinical deterioration, defined as either death or requiring an intensive care unit transfer. Their study of 2428 patient records across 29 academic medical centers in the US found that 23% had experienced a diagnostic error, with 17% of patients experiencing harm or death as a result.
While these findings are striking, it is important to highlight that this was a selected sample of the sickest patients in the hospital. Some of these patients may have had poor outcomes regardless of the errors. Subsequent research could incorporate comparison groups of patients who were admitted with similar diagnoses and clinical severity and assess the association between diagnostic error and subsequent outcome.
The study by Auerbach et al2 also noted variation in error rates across centers, emphasizing the need to understand contextual factors that impede or improve the diagnostic process. The most common and deleterious errors included the failure to recognize, or order appropriate testing for, a particular diagnosis or complication. Why might the frequency of diagnostic errors vary across hospitals? Excessive physician workloads, exacerbated by the COVID-19 pandemic, have been previously implicated in adverse patient outcomes and warrant further investigation.3 Gaps in house staff training in both error reporting and clinical reasoning may also contribute to diagnostic errors. Future studies can assess training interventions that might improve diagnostic reasoning among physicians and trainees. Other interventions might include discussion of diagnostic uncertainties during rounds and patient handoffs, expert check-ins for challenging cases, or using natural language model–based chatbots to ensure that critical differential diagnoses are considered.
At the population level, diagnostic errors contribute to more patient harm than other types of errors, including treatment error.1 As the complexity of medical practice grows, we have a responsibility to patients to examine our role in contributing to patient harm through diagnostic error and invest in research and quality improvement initiatives to strengthen the diagnostic process in medical education and clinical care.