Today is World Sepsis Day—a day for people worldwide to unite in the fight against sepsis. One of the primary goals is to increase awareness about sepsis, the No. 1 preventable cause of death worldwide and arguably one of the most challenging conditions for hospitals to treat.
According to the CDC, each year in the U.S. more than 1.7 million people get sepsis, a potentially life-threatening condition that occurs when chemicals released to fight infection trigger inflammatory responses that can damage organ systems. A staggering one in three patients who die in a hospital have sepsis. Hospitals spend more than $24 billion annually on sepsis, making it the most expensive condition to treat in the entire U.S. healthcare system.
Statistics like these prompted the University of Utah Health to take a closer look at how they detect and treat sepsis in their facilities. Like many other hospitals, University of Utah Health discovered sepsis was the leading cause of death at its flagship University Hospital (42 percent of all mortality), prompting the academic medical center to embark on a journey to improve sepsis response.
Devin Horton, MD, hospitalist and assistant professor in the division of internal medicine at University of Utah Health, recently presented their approach and successes in a webinar with Spok and Becker’s Hospital Review. The webinar attracted more than 1,000 registrants, and for good reason—Dr. Horton and his colleagues were able to reduce mortality rate for patients with MEWS scores 7-11 by 20 percent and length of stay and total cost for all sepsis patients by 10 percent in just a few years.
From Awareness to Action: The PSDA Approach
It didn’t take much to convince Dr. Horton that there was room for improvement in how they responded to sepsis. They were seeing decompensation throughout the hospital and anecdotally seeing delays in providers recognizing and treating sepsis. Most eye-opening, they discovered that residents and sometimes even seasoned providers couldn’t accurately recognize the signs of sepsis. In a 21-question quiz about sepsis, the average score for residents was 56 percent, and 17 percent for nurses. It frequently took hours to complete the gold standard of treatment for sepsis—check lactate, then administer antibiotics and IV fluids—and the step-intensive process had multiple opportunities for human error.
Dr. Horton and his team methodically tackled this challenge beginning in 2014, with the plan-study-do-act process from the Institute for Healthcare Improvement. They formed a cross-functional team to fully understand the current process to identify and treat sepsis throughout the facility, and gather input on how to design a better system with forced functions and constraints, automation and computerization, and standardization and protocols.
A key learning? The signs of sepsis were there—they just weren’t always noticed. “When I was consulting for a fever, I’d go see a patient, get into their chart and find they had abnormal vital signs that had been there for several days,” Dr. Horton explains. “Our EHR imports those notes every day, but there were no discussions about those vital signs.” Nursing assistants often recorded vital signs at the bedside for several patients, then entered them into the computer all at once, making the first vital signs up to an hour old by the time they were entered into the computer.
“If this was an emergent case, we’d already lost an hour,” Dr. Horton said. “There may not be communication about those vital signs, they may just sit in the computer waiting for the nurse to see them and a provider may not get back to them quickly.”
Introducing Automation to Speed Sepsis Response
Dr. Horton said they quickly realized that University of Utah Health could leverage its existing systems to bridge this gap and proactively deliver notifications to the right providers when patient vital signs or test results indicate the possibility of sepsis. They harnessed the surveillance and clinical decision support capabilities of its EHR, Epic, to create a best practice alert when MEWS scores were elevated or when a critical test result was logged. When those values exceed the set thresholds, the health system’s clinical communication platform, Spok Care Connect®, sends that sepsis alert to the right clinicians’ mobile devices automatically.
Not only is the alert delivered automatically, cutting down on wait times, but the alert also includes all of the information the clinicians need to act right away. By including the name of the patient, their room number, and their MEWS score in a secure, HIPAA-compliant message, the clinician receiving the alert already has a good understanding of the situation and the urgency required.
“What was helpful for us was having all our sepsis data in one place — we can look at the data and take it back to our providers to tell them what we’re seeing,” Dr. Horton said. “If you have an EHR-based algorithm, patients’ illness can be detected earlier and resuscitated earlier.”
Results and Lessons Learned
While the sepsis initiative is ongoing, and Dr. Horton says they’re continually learning and refining the process, the initial results of the project are incredibly promising. In just a few years, University of Utah Health reduced the mortality rate for its patients with elevated MEWS scores (7-11) by 20 percent, and across all sepsis patients, decreased length of stay and total direct cost by 10 percent.
One of the biggest lessons Dr. Horton wants to share with other hospitals aiming to improve sepsis response is that it’s an organization-wide, multi-faceted approach that encompasses people, processes, and technology:
- Understand the current state of sepsis detection and recognition at your hospital
- Tailor alerts to meet institutional needs
- Leverage EHR clinical decision support
- Hardwire communication to automate notification and speed response
- Measure impact, adjust, and expand success throughout the hospital
“We can have the best hospital in the world,” Dr. Horton explains, “but if you don’t know what vital signs are, and if the vital signs aren’t entered into the computer in real time, then that patient is losing, and the institution is losing.”
Dive deeper into this incredible case study by watching the full webinar recording.