Clinician Uncertainty in Pneumonia Diagnoses Leads to Misdiagnoses

Barbara Jones, MD, MSCI

Credit: University of Utah Health

Recent AI-based research found pneumonia diagnoses often come with uncertainty and can be prone to misdiagnosis.¹ Patients may be admitted to the hospital for pneumonia only to be diagnosed with something else upon discharge, or they might seek treatment for a different condition and discover they have pneumonia instead.

“Pneumonia can seem like a clear-cut diagnosis,” said lead investigator Barbara Jones, MD, MSCI, pulmonary and critical care physician at University of Utah Health, in a press release, “but there is actually quite a bit of overlap with other diagnoses that can mimic pneumonia.”²

A third of patients who were diagnosed with pneumonia did not receive a pneumonia diagnosis when they were first admitted to the hospital.¹ Conversely, nearly 40% of initial pneumonia diagnoses were discovered to not be pneumonia but something else entirely.

Treating community-acquired pneumonia relies heavily on having an accurate initial diagnosis of the condition. An incorrect initial diagnosis can affect the effectiveness of the treatment strategies based on evidence and guidelines. Thus, investigators sought to examine the evolution of pneumonia diagnoses among patients hospitalized in the emergency department.

Investigators conducted a retrospective nationwide study, leveraging data from 118 United States Veterans Affairs Medical Centers. Participants, aged ≥ 18 years, were hospitalized in the emergency department between January 1, 2015, to January 31, 2022. Investigators analyzed 2,383,899 hospitalizations.

The team collected data on discordances between initial pneumonia diagnosis, discharge diagnosis, and radiographic diagnosis identified by natural language processing of clinician text, diagnostic coding, and antimicrobial treatment. Barbara and colleagues also examined how uncertainty was expressed in clinical notes, as well as its correlation with patient illness severity, treatments, and outcomes,

Among the hospitalizations, 13.3% received an initial or discharge diagnosis and treatment for pneumonia, 9.1% received an initial diagnosis, and 10% received a discharge diagnosis. More than half (57%) of the participants had discordance between admission and discharge.

“These findings highlight substantial diagnostic uncertainty and treatment ambiguity in pneumonia that warrants recognition by systems, providers, and patients,” investigators wrote.

Of those discharged with a pneumonia diagnosis and had a positive initial chest image, 33% lacked an initial diagnosis. Among patients initially diagnosed with pneumonia, 36% lacked a discharge diagnosis and 21% lacked positive initial chest imaging.

Investigators observed uncertainty was expressed in clinical notes for 58% of patients in the emergency department and 48% at discharge. Despite uncertainty, 27% received diuretics, 36% received corticosteroids, and 10% received antibiotics, corticosteroids, and diuretics within 24 hours.

Patients with discordant diagnoses had greater uncertainty and received more treatments than patients with nonchanging diagnoses. However, patients lacking an initial pneumonia diagnosis had a greater 30-day mortality compared to concordant patients (14.4%; 95% confidence interval [CI], 14.1 – 14.7% vs 10.6% 95% CI, 10.4 – 10.7). Furthermore, patients with inconsistent diagnoses were more likely to seek care at complex medical facilities. These facilities tended to have a high number of emergency department visits and a high patient count in their inpatient wards.

Investigators wrote the findings were limited by the retrospective design, which does now allow for the examination of casual relationships.

Jones advised doctors and patients to keep in mind the uncertainty after an initial diagnosis and be willing to change treatment courses.²

“Both patients and clinicians need to pay attention to their recovery and question the diagnosis if they don’t get better with treatment,” Jones said.

References

1. ^{Jones, B, Chapman, A, Ying, J. Diagnostic Discordance, Uncertainty, and Treatment Ambiguity in Community-Acquired Pneumonia: A National Cohort Study of 115 U.S. Veterans Affairs Hospitals. Annuals of Internal Medicine. 2024. https://dx.doi.org/10.7326/M23-2505}
2. ^{Hospital Pneumonia Diagnoses Are Uncertain, Revised More Than Half the Time, Study Finds. EurekAlert! August 5, 2024. https://www.eurekalert.org/news-releases/1052966?. Accessed August 5, 2024.}