Use of EHR integrated sepsis model did not impact time to antibiotics for sepsis in ED
Click Here to Manage Email Alerts
NASHVILLE, Tenn. — Use of an electronic health record integrated sepsis model did not impact time to antibiotics for patients with sepsis in the emergency department, according to a study presented at the CHEST Annual Meeting.
Researchers conducted a retrospective analysis to determine how the EHR-integrated EPIC Sepsis Model impacts time to antibiotics among septic patients who presented to the ED from February to June 2019.
“We wanted to try to understand better the EPIC Sepsis Model as it is functioning in our hospital and how it was impacting antibiotic use,” Daniel M. Burgin, MD, a third-year internal medicine resident at Louisiana State University Health Sciences Center, Baton Rouge, Louisiana, told Healio. “We previously reviewed the performance characteristics of the model, which we found to be poor, and we wanted to understand more about its overall impact on delivering care.”
The retrospective analysis included 226 patients who had at least two of the Systemic Inflammatory Response Syndrome criteria. All patients were considered infected according to the physician chart review and were classified as meeting Sepsis-2. A physician panel further adjudicated patients for sepsis using the Sepsis-3 definition. If patients reached a threshold of 6 on the EPIC Sepsis Model, the EHR prompted an order set for providers on management options, which included antibiotics. The EPIC Sepsis Model was calculated every 15 minutes.
“What we found was that the EPIC Sepsis Model has a very poor correlation with antibiotic orders. Many of the EPIC Sepsis Model triggers came well after antibiotic orders. Several of the patients who were adjudicated as septic did not generate an alert at all,” Burgin told Healio. “We did not find the model to improve antibiotic orders or care delivered to patients with sepsis.”
According to the results, 48.2% of patients received antibiotics. Of those, most (74.5%) had an EPIC Sepsis Model score below 6. The researchers reported a decrease in time from triage to antibiotics ordered for patients with an EPIC Sepsis Model score of 6 or greater compared with a score below 6 (90.5 minutes vs. 131.5 minutes; P = .011). There was also a significant decrease in time from triage to antibiotics administered for patients with a score of 6 or greater compared with a score below 6 (136 minutes vs. 186 minutes; P = .011), according to the abstract.
Overall, 188 of the 226 patients in this study were evaluated for infection. Of those, 45.7% met Sepsis-2 criteria and were more likely to receive antibiotics in the ED compared with those who did not meet the criteria (76.7% vs. 22.8%; P < .001), according to the results. Twenty-one patients were adjudicated as septic using the Sepsis-3 definition; all received antibiotics, but five did not receive antibiotics within 6 hours of triage. Median time from the EPIC Sepsis Model score of 6 or greater to antibiotics ordered by the ED provider in Sepsis-3 patients was –5 and median time from score to antibiotics administered was 38.5 minutes.
Looking ahead, the researchers plan to evaluate if there is an alternative way or workflow in which this tool could be utilized to help improve delivery of care for patients with sepsis in the ED, Burgin said.
“One of the key take-home points is that this software is going to be looked at as a medical device and it is important to balance its role with potential for harm so these alerts are not causing an overload of alerts generated that could potentially lead to harm,” Burgin told Healio.