Using a streamlined alert system that draws information from the hospital’s electronic health record, Yale researchers have developed a software tool to more quickly detect deteriorations in patients’ health.
The software successfully identified patients with high risk of deterioration — particularly those who have developed sepsis, a common but life-threatening condition that arises when the body’s response to infection injures its own tissues and organs. The team tested the tool with 15,000 patients over the course of a year, and the results were published on Jan. 23 in the Journal of Patient Safety.
“Rather than waiting for a nurse to open a chart or a doctor to walk into the room, the tool automatically examines for these conditions whenever the data is immediately available,” said School of Medicine professor Robert Fogerty, the researcher who led the project.
Hospital-acquired infections, in which patients develop an illness while in the hospital, remain highly underrecognized, Fogerty said. When treatment is delayed, the infections can trigger sepsis, and outcomes deteriorate quickly.
Christopher Sankey, a professor at the School of Medicine and member of the team, added that internal data at Yale New Haven Hospital suggested a delay in identification of sepsis-spectrum disorders that develop during hospitalization.
“What we know is that, if you have sepsis, your risk of death goes up for every 60 minutes that your antibiotics are delayed,” Fogerty said.
To address these delays, the team came together to design a tool that would allow clinicians to identify sepsis more quickly. According to Fogerty, they received funding through a pilot grant program from the hospital’s malpractice insurance company.
Sepsis can be detected using systemic inflammatory response syndrome criteria, which include fever, heart rate, respiratory rate and white blood cell count. Unfortunately, all these data points enter the electronic health record from different departments, Fogerty said, so the data are only collected together when a clinician manually opens and reads through a patient’s medical record.
Recognizing the tedious and time-consuming nature of the process, the team developed a software program to automatically synthesize the data points when they are each entered. After the data entry, the program generates an alert to clinicians if four of six criteria — which include blood pressure and change in kidney function in addition to the systemic inflammatory response syndrome criteria — are satisfied.
Although the team originally focused on sepsis, they soon realized that these criteria could also detect patients with conditions such as uncontrolled alcohol withdrawal syndrome or respiratory failure, who would also benefit significantly from early intervention and therapy, Fogerty said.
The software was then tested at the York Street campus of Yale New Haven Hospital. The researchers found that their “Sepsis Alert” system was able to identify patients with increased risks of death or intensive care unit admission, according to Sankey.
Additionally, Fogerty said, the software program identified patients as the same risk levels of deterioration as the clinicians and nurses who manually checked the health record.
The researchers noted that one major benefit of leveraging the electronic health record to improve care for patients is the low cost of the new software.
“Rather than spending a lot of money on something new, we took something off the shelf and refocused it,” Fogerty said. “For a really miniscule amount of money, we created something that’s easily deployable, widely expandable and highly sustainable.”
Further studies are required to understand the cause of increased risks of ICU admission and mortality and to investigate how to intervene to improve clinical outcomes after the identification of high-risk patients, said Co-Director of Analytics at the Yale Center for Analytical Sciences Fangyong Li, the leading statistician on the project.
About 250,000 people die from sepsis each year in the U.S., according to the Centers for Disease Control and Prevention.
Amy Xiong | email@example.com