番茄社区app

Diagnostic errors among hospitalized adults who died or were transferred to the intensive care unit (ICU) were fairly common, a retrospective cohort study suggested.

In a random sample of over 2,400 patients who died or were transferred to the ICU at 29 academic medical centers, 23% experienced a diagnostic error (95% CI 20.9-25.3), 17.8% of which were judged to have contributed to temporary harm, permanent harm, or death (95% CI 15.9-19.8), reported Andrew Auerbach, MD, MPH, of the University of California San Francisco, and colleagues.

Of the 1,863 patients who died, diagnostic errors were judged to have contributed in 6.6% of cases (95% CI 5.3-8.2), they noted in .

"The main takeaways for us were the incidence of errors, which was higher than we expected, as were the harms," Auerbach told 番茄社区app in an email.

The diagnostic process faults most highly associated with these errors were problems with patient assessment, as well as test ordering and interpretation, the authors explained.

"We expected problems with assessment as being central as an underlying cause," Auerbach said. "But the role of test choice and interpretation as a cause of errors (as opposed to physical examination or history) was intriguing in that it may be an indicator of how complex test ordering can be in seriously ill patients."

Diagnostic errors are thought to contribute to more patient harms than other types of errors, according to "Improving Diagnosis in Health Care," a from the National Academies of Sciences, Engineering, and Medicine, and few studies have examined diagnostic errors in hospital inpatients, Auerbach and colleagues noted.

In an , Grace Zhang, MD, of the University of California San Francisco, and Cary Gross, MD, of the Yale School of Medicine in New Haven, Connecticut, wrote that excessive physician workloads, along with "gaps in house staff training," could contribute to these errors.

"While these findings are striking, it is important to highlight that this was a selected sample of the sickest patients in the hospital," they added.

In an , Robert Wachter, MD, of the University of California San Francisco, compared medical errors to the response to COVID-19. Medical errors have been reframed as "mostly manifestations of system rather than individual failures" through the "Swiss cheese model" of layers of protection, which was also applied to COVID safety measures, he wrote.

Both issues, he noted, "illustrate that motivating ongoing prevention-oriented behaviors and policies requires continued dissemination of both data and stories, strong and consistent leadership, and a combination of carrots and sticks."

Auerbach said that "in my practice, I use all adverse events -- not only ICU transfers and deaths, but things like readmissions or urgent escalations in care -- to take a diagnostic pause or timeout to walk through a short checklist of questions about what we should be (or should have) considered." He is studying this kind of "pause" in further research, he added.

For this study, the researchers used the Vizient Clinical Data Base to identify 2,428 patients with a general medical diagnosis who died or were transferred to the ICU during their hospitalization after the second hospital day from January 1 to December 31, 2019. Mean patient age was 63.9 years, 45.6% were women, and about two-thirds were white.

Patient cases were reviewed by two physicians trained in error adjudications, who agreed entirely on the assessment of error before finalization. A third physician resolved disagreements. The reviewers evaluated medical records for the presence or absence of diagnostic error, and underlying diagnosis process faults. Records with these faults were reviewed for harms because of the error.

They identified errors using the modified , and diagnostic process fault information with the . Harms related to errors were reviewed using the .

The authors acknowledged that their research did not capture diagnostic errors across all hospitalized patients, and that data could have been subject to documentation and detection bias. They also could not distinguish what type of cognitive process was associated with a diagnostic error, and did not capture issues with team communication or dynamics, or non-medical harms associated with the diagnostic errors.

They also did not capture external pressures on teams or clinicians (like workload), and local reviewers may have been influenced by local norms and professional standards, or by the fact that all patients had experienced ICU transfer or death.