The study, led by researchers at Brigham and Women’s Hospital in Boston, analyzed cases of adult patients who either transferred to an intensive care unit, died within 90 days of admission, had complex clinical events or did not meet those criteria.
Among 675 randomly sampled cases, 7.2% of the diagnostic errors were harmful, 6.1% were preventable and 1.1% were severely harmful. Based on the sample, the researchers estimated 1 in 14 patients received a harmful diagnostic error.
About 62% of harmful errors were characterized by delays.
The researchers recommended interventions for process failures often associated with harmful diagnostic errors, such as uncertainty in initial assessments, complex testing and interpretation, suboptimal subspecialty consultation, patient-reported concerns and history-taking. They also suggested the rapid adoption of AI to detect complex patterns of risk factors.
