In a new study, researchers from Intermountain Health in Salt Lake City have shown that by using a computerized clinical decision support tool to guide diagnosis and treatment in emergency department patients with pneumonia, they were able to reduce mortality rates, especially in patients experiencing socioeconomic hardships.
Pneumonia, a serious infection in the lungs that causes inflammation and difficulty breathing, is one of the most common causes of death worldwide. An estimated 150,000 people are hospitalized annually for pneumococcal pneumonia, and about 1 in 20 dies as a result.
While anyone can get pneumonia, some people are at risk, including young children, older adults, and people with certain chronic health conditions.
Pneumonia has also been found to disproportionately affect individuals and communities experiencing economic hardship. This is due to a combination of factors that increase susceptibility to the disease and reduce access to treatment and prevention.
ePneumonia was created in 2011 to predict and guide diagnosis and treatment of patients with pneumonia and is integrated into Intermountain's Health normal workflow. It uses prior medical history, emergency department vital signs, findings on chest imaging lab results, and chest imaging findings.
"By using ePneumonia to predict who will be affected by pneumonia, we flattened the differences in mortality in patients across all social groups," said Jason R. Carr, MD, an investigator in the study and pulmonary medicine specialist and researcher at Intermountain Health. "With this technology, we're saving more lives, no matter where you come from."
The Intermountain Health study findings were presented at the ATS 2025, the American Thoracic Society's annual conference, which is being held this year in San Francisco.
The study examined whether use of ePneumonia improved outcomes in patients with higher Area Deprivation Indexes (ADI), a calculation that scores socioeconomic disadvantage based on poverty, education, housing and employment. The higher the score, the more socially disadvantaged the individual.
Patients with high ADI are of particular concern in Utah, which has rural areas with areas that may experience economic challenges.
In the study, Intermountain researchers looked at the clinical impact that implementation of ePneumonia had on patients at 16 Intermountain Health community hospitals across Utah from June 2016 to June 2019.
They did so by identifying adults with emergency department and hospital discharge diagnosis codes from pneumonia or sepsis/respiratory failure associated with pneumonia that was confirmed by X-rays by emergency departments.
Researchers found that overall, ePneumonia was associated with significant reduction in mortality among patients with pneumonia who also had high ADI scores, meaning that fewer people at higher risk died of pneumonia.
The team believes that ePneumonia reduces unconscious bias in physicians' calculations on who has pneumonia, and treatment decisions, said Nathan C. Dean, MD, a study investigator and at Intermountain Medical Center in Murray.
Artificial intelligence is built into ePneumonia and it reduces inappropriate deviation from best practices of care, said Dr. Dean.
"Physicians may not realize they're doing this, but it happens. These findings show that having help from a tool like ePneumonia helps our patients by helping physicians overcome those unconscious biases they might not see themselves to ensure all patients are getting the best care possible," he added.
These preliminary results are encouraging but need further study for confirmation, said Dr. Carr.
